Title

Diving into the Depths of Spotting Text in Multi-Domain Noisy Scenes

Masked Local-Global Representation Learning for 3D Point Cloud Domain Adaptation
Continuous Adaptation in Person Re-Identification for Robotic Assistance
Spectral Geometric Verification: Re-Ranking Point Cloud Retrieval for Metric Localization
Incorporating Scene Graphs into Pre-Trained Vision-Language Models for Multimodal Open-Vocabu
LPS-Net: Lightweight Parameter-Shared Network for Point Cloud-Based Place Recognition
Joint Response and Background Learning for UAV Visual Tracking
ZS6D: Zero-Shot 6D Object Pose Estimation Using Vision Transformers
Fluxformer: Flow-Guided Duplex Attention Transformer Via Spatio-Temporal Clustering for Action
Development of Variable Transmission Series Elastic Actuator for Hip Exoskeletons
Optimization of Mono and Bi-Articular Parallel Elastic Elements for a Robotic Arm Performing a P
A Novel Compact Design of a Lever-Cam-Based Variable Stiffness Actuator: LC-VSA
Design and Modeling of a Compact Serial Variable Stiffness Actuator (SVSA-III) with Linear Stiffnes
Optimally Controlling the Timing of Energy Transfer in Elastic Joints: Experimental Validation of th
Experimental Comparison of Pinwheel and Non-Pinwheel Designs of 3D-Printed Cycloidal Gearing
Design and Optimization of an Origami-Inspired Foldable Pneumatic Actuator
A Non-Magnetic Dual-Mode Linear Pneumatic Actuator: Initial Design and Assessment
Variable Stiffness Floating Spring Leg: Performing Net-Zero Energy Cost Tasks Not Achievable Usi
Path and Trajectory Planning of a Tethered UAV-UGV Marsupial Robotic System
Cooperative Exploration of Heterogeneous UAVs in Mountainous Environments by Constructing
Perception-And-Energy-Aware Motion Planning for UAV Using Learning-Based Model under Hetero
Representing On-Orbit Rendezvous and Proximity Operations with Fully-Actuated Multirotor Aerial
On-Device Self-Supervised Learning of Visual Perception Tasks Aboard Hardware-Limited Nano-Q
Aerobatic Trajectory Generation for a VTOL Fixed-Wing Aircraft Using Differential Flatness
EVOLVER: Online Learning and Prediction of Disturbances for Robot Control
Time-Optimal Path Planning in a Constant Wind for Uncrewed Aerial Vehicles Using Dubins Set Cla
Robust and Efficient Depth-Based Obstacle Avoidance for Autonomous Miniaturized UAVs
Sim-To-Real Robotic Sketching Using Behavior Cloning and Reinforcement Learning
Safe Table Tennis Swing Stroke with Low-Cost Hardware
Pluck and Play: Self-Supervised Exploration of Chordophones for Robotic Playing
MBot: A Modular Ecosystem for Scalable Robotics Education
SO(2)-Equivariant Downwash Models for Close Proximity Flight
Hierarchical Meta-Learning-Based Adaptive Controller
A Novel Wide-Area Multiobject Detection System with High-Probability Region Searching
Real-Time Robot Navigation and Manipulation with Distilled Vision-Language Models
Vision-Language Interpreter for Robot Task Planning
Trust-Region Neural Moving Horizon Estimation for Robots
Multi-Category Decomposition Editing Network for the Accurate Visual Inspection of Texture Defec
Kinematic-Aware Prompting for Generalizable Articulated Object Manipulation with LLMs
ISR-LLM: Iterative Self-Refined Large Language Model for Long-Horizon Sequential Task Planning
Efficient Hybrid Neuromorphic-Bayesian Model for Olfaction Sensing: Detection and Classification
Disentangled Neural Relational Inference for Interpretable Motion Prediction
DeFlow: Decoder of Scene Flow Network in Autonomous Driving
Subequivariant Reinforcement Learning Framework for Coordinated Motion Control
Ethically Compliant Autonomous Systems under Partial Observability
Prompt, Plan, Perform: LLM-Based Humanoid Control Via Quantized Imitation Learning
Infer and Adapt: Bipedal Locomotion Reward Learning from Demonstrations Via Inverse Reinforce
Online Distribution Shift Detection Via Recency Prediction
Simplified Continuous High Dimensional Belief Space Planning with Adaptive Probabilistic Belief-
Conformal Policy Learning for Sensorimotor Control under Distribution Shifts
Resampling-Free Particle Filters in High-Dimensions
A New Perspective of DL Testing Framework: Human-Computer Interaction Based Neural Network
Toward AI-Enabled Commercial Telepresence Robots to Combine Home Care Needs and Affordabi
SliceIt! - a Dual Simulator Framework for Learning Robot Food Slicing
SG-Bot: Object Rearrangement Via Coarse-To-Fine Robotic Imagination on Scene Graphs
Scalable Multi-Robot Collaboration with Large Language Models: Centralized or Decentralized Sy
Object-Centric Instruction Augmentation for Robotic Manipulation
Learning to Play Foosball: System and Baselines
Language-Conditioned Robotic Manipulation with Fast and Slow Thinking
How to Prompt Your Robot: A Prompt Book for Manipulation Skills with Code As Policies
A Multifidelity Sim-To-Real Pipeline for Verifiable and Compositional Reinforcement Learning
When to Replan? an Adaptive Replanning Strategy for Autonomous Navigation Using Deep Reinfo
Resolving Loop Closure Confusion in Repetitive Environments for Visual SLAM through AI Founda
Prepare the Chair for the Bear! Robot Imagination of Sitting Affordance to Reorient Previously Un
Gen2Sim: Scaling up Robot Learning in Simulation with Generative Models
FLTRNN: Faithful Long-Horizon Task Planning for Robotics with Large Language Models
Drive Anywhere: Generalizable End-To-End Autonomous Driving with Multi-Modal Foundation Mod
AutoTAMP: Autoregressive Task and Motion Planning with LLMs As Translators and Checkers
ASC: Adaptive Skill Coordination for Robotic Mobile Manipulation
Forgetting in Robotic Episodic Long-Term Memory
Spatial Assisted Human-Drone Collaborative Navigation and Interaction through Immersive Mixed R
A Compiler Framework for Proactive UAV Regulation Enforcement
Real-Time Dynamic-Consistent Motion Planning for Over-Actuated UAVs
Trajectory Optimization for Cooperatively Localizing Quadrotor UAVs
Extending Guiding Vector Field to Track Unbounded UAV Paths
Tethered Lifting-Wing Multicopter Landing Like Kite
AirFisheye Dataset: A Multi-Model Fisheye Dataset for UAV Applications
Bio-Inspired Visual Relative Localization for Large Swarms of UAVs
Heuristic-Based Incremental Probabilistic Roadmap for Efficient UAV Exploration in Dynamic Envi
Under Pressure: Learning-Based Analog Gauge Reading in the Wild
MORALS: Analysis of High-Dimensional Robot Controllers Via Topological Tools in a Latent Space
TinyMPC: Model-Predictive Control on Resource-Constrained Microcontrollers
A Movable Microfluidic Chip with Gap Effect for Manipulation of Oocytes
Efficient Composite Learning Robot Control under Partial Interval Excitation
Sim-On-Wheels: Physical World in the Loop Simulation for Self-Driving
Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing
Robust Autonomous Vehicle Pursuit without Expert Steering Labels
Risk-Aware Trajectory Prediction by Incorporating Spatio-Temporal Traffic Interaction Analysis
Reinforcement Learning with Human Feedback for Realistic Traffic Simulation
Plug in the Safety Chip: Enforcing Constraints for LLM-Driven Robot Agents
InterCoop: Spatio-Temporal Interaction Aware Cooperative Perception for Networked Vehicles
Improving Autonomous Driving Safety with POP: A Framework for Accurate Partially Observed Traj
FIMP: Future Interaction Modeling for Multi-Agent Motion Prediction
Adaptive Pedestrian Agent Modeling for Scenario-Based Testing of Autonomous Vehicles through
KT-BT: A Framework for Knowledge Transfer through Behavior Trees in Multi-Robot Systems
Distributed Matching-By-Clone Hungarian-Based Algorithm for Task Allocation of Multi-Agent Sys
Convolutional Vision Transformer As a Path Following Controller for Omnidirectional Robots
Can an Embodied Agent Find Your “Cat-Shaped Mug”? LLM-Based Zero-Shot Object Navig
AutoExplorers: Autoencoder-Based Strategies for High-Entropy Exploration in Unknown Environme
LLM-BT: Performing Robotic Adaptive Tasks Based on Large Language Models and Behavior Tre
DyHGDAT: Dynamic Hypergraph Dual Attention Network for Multi-Agent Trajectory Prediction
Driving with LLMs: Fusing Object-Level Vector Modality for Explainable Autonomous Driving
HHGNN: Heterogeneous Hypergraph Neural Network for Traffic Agents Trajectory Prediction in Gr
Odometry Estimation by Fusing Multiple Radar Sensors and an Inertial Measurement Unit
Thermal Voyager: A Comparative Study of RGB and Thermal Cameras for Night-Time Autonomous
Rethinking Imitation-Based Planners for Autonomous Driving
Traffic Flow-Based Crowdsourced Mapping in Complex Urban Scenario
Scene Informer: Anchor-Based Occlusion Inference and Trajectory Prediction in Partially Observa
Monocular Localization with Semantics Map for Autonomous Vehicles
DiPA: Probabilistic Multi-Modal Interactive Prediction for Autonomous Driving
MARC: Multipolicy and Risk-Aware Contingency Planning for Autonomous Driving
Graph-Based Scenario-Adaptive Lane-Changing Trajectory Planning for Autonomous Driving
Toward Wheeled Mobility on Vertically Challenging Terrain: Platforms, Datasets, and Algorithms
Rethinking Social Robot Navigation: Leveraging the Best of Two Worlds
Continuous Robotic Tracking of Dynamic Targets in Complex Environments Based on Detectability
Talk2BEV: Language-Enhanced Bird’s-Eye View Maps for Autonomous Driving
DualAT: Dual Attention Transformer for End-To-End Autonomous Driving
SCALE: Self-Correcting Visual Navigation for Mobile Robots Via Anti-Novelty Estimation
UniGen: Unified Modeling of Initial Agent States and Trajectories for Generating Autonomous Driv
Implicit Point Function for LiDAR Super-Resolution in Autonomous Driving
Circular Accessible Depth: A Robust Traversability Representation for UGV Navigation
Robots That Can See: Leveraging Human Pose for Trajectory Prediction
Uncertainty-Aware Reinforcement Learning for Autonomous Driving with Multimodal Digital Driver
Boosting Offline Reinforcement Learning for Autonomous Driving with Hierarchical Latent Skills
A Framework for Real-Time Generation of Multi-Directional Traversability Maps in Unstructured E
Cross-Modal Registration Using Adaptive Modeling in Infrastructure-Based Vehicle Localization
Efficient Gas Source Active Search in Unfamiliar Environments
RGBD-Based Image Goal Navigation with Pose Drift: A Topo-Metric Graph Based Approach
OpenBot-Fleet: A System for Collective Learning with Real Robots
WOMD-LiDAR: Raw Sensor Dataset Benchmark for Motion Forecasting
Increasing the Absolute Position Accuracy of Industrial Robots by Means of a Deep Continual Evi
SpawnNet: Learning Generalizable Visuomotor Skills from Pre-Trained Network
RoboAgent: Generalization and Efficiency in Robot Manipulation Via Semantic Augmentations and
Dream2Real: Zero-Shot 3D Object Rearrangement with Vision-Language Models
Robot Fine-Tuning Made Easy: Pre-Training Rewards and Policies for Autonomous Real-World Re
Scaling Motion Forecasting Models with Ensemble Distillation
Is It a Bug? Understanding Physical Unit Mismatches in Robot Software
Ospreys-Inspired Self-Takeoff Strategy of an Eagle-Scale Flapping-Wing Robot: System Design an
Design and Analysis of Adaptive Flipper with Origami Structure for Frog-Inspired Swimming Robot
Model-Based Approach for Lateral Maneuvers of Bird-Size Ornithopter
A General Kinematic Model of Fish Locomotion Enables Robot Fish to Master Multiple Swimming
Adaptation of Flipper-Mud Interactions Enables Effective Terrestrial Locomotion on Muddy Substra
RoboTwin: A Platform to Study Hydrodynamic Interactions in Schooling Fish
Real-Time Estimation for the Swimming Direction of Robotic Fish Based on IMU Sensors
Tunable Stiffness Caudal Peduncle Leads to Higher Swimming Speed without Extra Energy
A Novel Fish-Inspired Self-Adaptive Approach to Collective Escape of Swarm Robots Based on
Visual CPG-RL: Learning Central Pattern Generators for Visually-Guided Quadruped Locomotion
Form Closure for Fully Actuated and Robust Obstacle-Aided Locomotion in Snake Robots
Self-Righting Shell for Robotic Hexapod
Quadruped-Frog: Rapid Online Optimization of Continuous Quadruped Jumping
AeroDima: Cheetah-Inspired Aerodynamic Tail Design for Rapid Maneuverability
Spined Torso Renders Advanced Mobility for Quadrupedal Locomotion
Pegasus: A Novel Bio-Inspired Quadruped Robot with Underactuated Wheeled-Legged Mechanis
LeapRun: A Dynamic Soft Robot with Running and Jumping Capabilities
Machine Learning-Driven Burrowing with a Snake-Like Robot
Takeoff of a 2.1g Fully Untethered Tailless Flapping-Wing Micro Aerial Vehicle with Integrated Batt
Design and Optimization of a Miniature Locust-Inspired Stable Jumping Robot
Multi-Modal Jumping and Crawling in an Autonomous, Springtail-Inspired Microrobot
High-Speed Interfacial Flight of an Insect-Scale Robot
VLEIBot: A New 45-Mg Swimming Microrobot Driven by a Bioinspired Anguilliform Propulsor
Direct Learning of Home Vector Direction for Insect-Inspired Robot Navigation
A Dragonfly-Inspired Flapping Wing Robot Mimicking Force Vector Control Approach
Microrobotic Flight Enabled by Ultralight Ion Thrusters with High Thrust-To-Weight Ratio and Low
A Modular Biological Neural Network-Based Neuro-Robotic System Via Local Chemical Stimulatio
Geared Rod-Driven Continuum Robot with Woodpecker-Inspired Extension Mechanism and IMU-B
A Multi-Modal Hybrid Robot with Enhanced Traversal Performance
Anisotropic Body Compliance Facilitates Robotic Sidewinding in Complex Environments
Combining Tail and Reaction Wheel for Underactuated Spatial Reorientation in Robot Falling wit
Environment-Modulated Self-Assembly by Changes in Modules' Buoyancy
Analysis and Validation of Stiffness and Payload of Nematode-Inspired Cable Routing Method for
Design, Implementation, and Observer-Based Output Control of a Super-Coiled Polymer-Driven
EasyHeC: Accurate and Automatic Hand-Eye Calibration Via Differentiable Rendering and Space E
Zero-Training LiDAR-Camera Extrinsic Calibration Method Using Segment Anything Model
Dive Deeper into Rectifying Homography for Stereo Camera Online Self-Calibration
Online Camera-LiDAR Calibration Monitoring and Rotational Drift Tracking
PeLiCal: Targetless Extrinsic Calibration Via Penetrating Lines for RGB-D Cameras with Limited Co-
A Novel, Efficient and Accurate Method for Lidar Camera Calibration
An Extrinsic Calibration Method between LiDAR and GNSS/INS for Autonomous Driving
SGCalib: A Two-Stage Camera-LiDAR Calibration Method Using Semantic Information and Geomet
EWand: An Extrinsic Calibration Framework for Wide Baseline Frame-Based and Event-Based C
LCCRAFT: LiDAR and Camera Calibration Using Recurrent All-Pairs Field Transforms without Preci
Physics-Informed Neural Network for Model Prediction and Dynamics Parameter Identification of Co
Estimating Material Properties of Interacting Objects Using Sum-GP-UCB
LiDAR-Camera Extrinsic Calibration with Hierachical and Iterative Feature Matching
GBEC: Geometry-Based Hand-Eye Calibration
A Learning-Based Approach for Estimating Inertial Properties of Unknown Objects from Encoder D
CalibFormer: A Transformer-Based Automatic LiDAR-Camera Calibration Network
Target-Free Extrinsic Calibration of Event-LiDAR Dyad Using Edge Correspondences
LB-R2R-Calib: Accurate and Robust Extrinsic Calibration of Multiple Long Baseline 4D Imaging Ra
Learning Vision-Based Bipedal Locomotion for Challenging Terrain
Resilient Legged Local Navigation: Learning to Traverse with Compromised Perception End-To-En
Vision-Language Frontier Maps for Zero-Shot Semantic Navigation
NoMaD: Goal Masked Diffusion Policies for Navigation and Exploration
Learning Continuous Control with Geometric Regularity from Robot Intrinsic Symmetry
CollisionGP: Gaussian Process-Based Collision Checking for Robot Motion Planning
Probabilistic Motion Planning and Prediction Via Partitioned Scenario Replay
Prescient Collision-Free Navigation of Mobile Robots with Iterative Multimodal Motion Prediction
GPU-Accelerated Optimization-Based Collision Avoidance
Learn to Navigate in Dynamic Environments with Normalized LiDAR Scans
Learning Terminal State of the Trajectory Planner: Application for Collision Scenarios of Autonomo
Informed Steiner Trees: Sampling and Pruning for Multi-Goal Path Finding in High Dimensions
History-Aware Planning for Risk-Free Autonomous Navigation on Unknown Uneven Terrain
DiPPeR: Diffusion-Based 2D Path Planner Applied on Legged Robots
Efficient Polynomial Sum-Of-Squares Programming for Planar Robotic Arms
PathRL: An End-To-End Path Generation Method for Collision Avoidance Via Deep Reinforcement
ZAPP! Zonotope Agreement of Prediction and Planning for Continuous-Time Collision Avoidance
Certifying Bimanual RRT Motion Plans in a Second
Cross View Capture for Distributed Image Compression with Decoder Side Information
Planning with Learned Subgoals Selected by Temporal Information
Neural Potential Field for Obstacle-Aware Local Motion Planning
Unconstrained Model Predictive Control for Robot Navigation under Uncertainty
MIM: Indoor and Outdoor Navigation in Complex Environments Using Multi-Layer Intensity Maps
Gaussian Process-Based Traversability Analysis for Terrain Mapless Navigation
Active Collision-Based Navigation for Wheeled Robots
Graph-Based 3D Collision-Distance Estimation Network with Probabilistic Graph Rewiring
Jump Over Block (JOB): An Efficient Line-Of-Sight Checker for Grid/voxel Maps with Sparse Obsta
Conformal Predictive Safety Filter for RL Controllers in Dynamic Environments
GrASPE: Graph Based Multimodal Fusion for Robot Navigation in Outdoor Environments
E-RRT*: Path Planning for Hyper-Redundant Manipulators
LiDAR-Based Online Control Barrier Function Synthesis for Safe Navigation in Unknown Environm
Physically Grounded Vision-Language Models for Robotic Manipulation
How Many Views Are Needed to Reconstruct an Unknown Object Using NeRF?
Active Implicit Reconstruction Using One-Shot View Planning
Multimodal Object Query Initialization for 3D Object Detection
SM^3: Self-Supervised Multi-Task Modeling with Multi-View 2D Images for Articulated Objects
MF-MOS: A Motion-Focused Model for Moving Object Segmentation
Improving Neural Indoor Surface Reconstruction with Mask-Guided Adaptive Consistency Constrai
NFL: Normal Field Learning for 6-DoF Grasping of Transparent Objects
TRTM: Template-Based Reconstruction and Target-Oriented Manipulation of Crumpled Cloths
Amodal Optical Flow
AutoGraph: Predicting Lane Graphs from Traffic Observations
3DSF-MixNet: Mixer-Based Symmetric Scene Flow Estimation from 3D Point Clouds
CVFormer: Learning Circum-View Representation and Consistency Constraints for Vision-Based O
Lightweight Event-Based Optical Flow Estimation Via Iterative Deblurring
ActFormer: Scalable Collaborative Perception Via Active Queries
LiDAR Data Synthesis with Denoising Diffusion Probabilistic Models
Multi-Task Learning for Real-Time Autonomous Driving Leveraging Task-Adaptive Attention Gener
LiDARFormer: A Unified Transformer-Based Multi-Task Network for LiDAR Perception
Intrinsic Contact Sensing and Object Perception of an Adaptive Fin-Ray Gripper Integrating Compa
Incipient Slip-Based Rotation Measurement Via Visuotactile Sensing During In-Hand Object Pivoti
Leveraging Compliant Tactile Perception for Haptic Blind Surface Reconstruction
Differentiable Compliant Contact Primitives for Estimation and Model Predictive Control
TacShade: A New 3D-Printed Soft Optical Tactile Sensor Based on Light, Shadow and Grey Scale
Physics-Encoded Graph Neural Networks for Deformation Prediction under Contact
Unwieldy Object Delivery with Nonholonomic Mobile Base: A Stable Pushing Approach
Robotic Contact Juggling
Beyond Coulomb: Stochastic Friction Models for Practical Grasping and Manipulation
LEAGUE: Guided Skill Learning and Abstraction for Long-Horizon Manipulation
Test-Time Adaptation in the Dynamic World with Compound Domain Knowledge Management
VERSE: Virtual-Gradient Aware Streaming Lifelong Learning with Anytime Inference
Experience Consistency Distillation Continual Reinforcement Learning for Robotic Manipulation Ta
Adapting to the “Open World”: The Utility of Hybrid Hierarchical Reinforcement Learning an
Lifelong Robot Library Learning: Bootstrapping Composable and Generalizable Skills for Embodie
Lifelong Robot Learning with Human Assisted Language Planners
Probabilistic Spiking Neural Network for Robotic Tactile Continual Learning
LOTUS: Continual Imitation Learning for Robot Manipulation through Unsupervised Skill Discovery
AirTwins: Modular Bi-Copters Capable of Splitting from Their Combined Quadcopter in Midair
ArrayBot: Reinforcement Learning for Generalizable Distributed Manipulation through Touch
Optimizing Modular Robot Composition: A Lexicographic Genetic Algorithm Approach
WiBot 1.0: A Modular Reconfigurable Glass Cleaning Robot for High-Rise Buildings
Collaborative Manipulation of Deformable Objects with Predictive Obstacle Avoidance
D-Lite: Navigation-Oriented Compression of 3D Scene Graphs for Multi-Robot Collaboration
ColAG: A Collaborative Air-Ground Framework for Perception-Limited UGVs' Navigation
GRF-Based Predictive Flocking Control with Dynamic Pattern Formation
Exploring the Needle Tip Interaction Force with Retinal Tissue Deformation in Vitreoretinal Surgery
PanNote: An Automatic Tool for Panoramic Image Annotation of People's Positions
A Multimodal Handover Failure Detection Dataset and Baselines
Introducing CEA-IMSOLD: An Industrial Multi-Scale Object Localization Dataset
PointSSC: A Cooperative Vehicle-Infrastructure Point Cloud Benchmark for Semantic Scene Comp
Close the Sim2real Gap Via Physically-Based Structured Light Synthetic Data Simulation
Interacting Objects: A Dataset of Object-Object Interactions for Richer Dynamic Scene Representa
RTS-GT: Robotic Total Stations Ground Truthing Dataset
RaSim: A Range-Aware High-Fidelity RGB-D Data Simulation Pipeline for Real-World Applications
Mobile Robot Oriented Large-Scale Indoor Dataset for Dynamic Scene Understanding
InteRACT: Transformer Models for Human Intent Prediction Conditioned on Robot Actions
Towards Learning-Based Planning: The nuPlan Benchmark for Real-World Autonomous Driving
TBD Pedestrian Data Collection: Towards Rich, Portable, and Large-Scale Natural Pedestrian Dat
RoboVQA: Multimodal Long-Horizon Reasoning for Robotics
CC-SGG: Corner Case Scenario Generation Using Learned Scene Graphs
RH20T: A Comprehensive Robotic Dataset for Learning Diverse Skills in One-Shot
SACSoN: Scalable Autonomous Control for Social Navigation
RiskBench: A Scenario-Based Benchmark for Risk Identification
Enhancing Inland Water Safety: The Lake Constance Obstacle Detection Benchmark
IDD-X: A Multi-View Dataset for Ego-Relative Important Object Localization and Explanation in De
LiDAR-CS Dataset: LiDAR Point Cloud Dataset with Cross-Sensors for 3D Object Detection
ROV6D: 6D Pose Estimation Benchmark Dataset for Underwater Remotely Operated Vehicles
The GOOSE Dataset for Perception in Unstructured Environments
CoAS-Net: Context-Aware Suction Network with a Large-Scale Domain Randomized Synthetic Dat
NYC-Indoor-VPR: A Long-Term Indoor Visual Place Recognition Dataset with Semi-Automatic Ann
TreeScope: An Agricultural Robotics Dataset for LiDAR-Based Mapping of Trees in Forests and O
Learning-Based Reflection-Aware Virtual Point Removal for Large-Scale 3D Point Clouds
Multi-Confidence Guided Source-Free Domain Adaption Method for Point Cloud Primitive Segment
FF-LOGO: Cross-Modality Point Cloud Registration with Feature Filtering and Local to Global Opti
CAPT: Category-Level Articulation Estimation from a Single Point Cloud Using Transformer
Energy-Based Detection of Adverse Weather Effects in LiDAR Data
EdgePoint: Efficient Point Detection and Compact Description Via Distillation
Fast and Robust Point Cloud Registration with Tree-Based Transformer
EgoPAT3Dv2: Predicting 3D Action Target from 2D Egocentric Vision for Human-Robot Interaction
Distribution-Aware Continual Test-Time Adaptation for Semantic Segmentation
STNet: Spatio-Temporal Fusion-Based Self-Attention for Slip Detection in Visuo-Tactile Sensors
Commonsense Spatial Knowledge-Aware 3-D Human Motion and Object Interaction Prediction
High-Degrees-Of-Freedom Dynamic Neural Fields for Robot Self-Modeling and Motion Planning
Language-Conditioned Affordance-Pose Detection in 3D Point Clouds
Multi-Object RANSAC: Efficient Plane Clustering Method in a Clutter
Utilizing Inpainting for Training Keypoint Detection Algorithms towards Markerless Visual Servoing
UAV-Sim: NeRF-Based Synthetic Data Generation for UAV-Based Perception
Contrastive Learning for Enhancing Robust Scene Transfer in Vision-Based Agile Flight
Watching the Air Rise: Learning-Based Single-Frame Schlieren Detection
High-Throughput Visual Nano-Drone to Nano-Drone Relative Localization Using Onboard Fully Con
End-To-End Semi-Supervised 3D Instance Segmentation with PCTeacher
PAg-NeRF: Towards Fast and Efficient End-To-End Panoptic 3D Representations for Agricultural
GAM-Depth: Self-Supervised Indoor Depth Estimation Leveraging a Gradient-Aware Mask and Sem
RoboKeyGen: Robot Pose and Joint Angles Estimation Via Diffusion-Based 3D Keypoint Generati
Advancements in 3D Lane Detection Using LiDAR Point Clouds: From Data Collection to Model D
PCPNet: An Efficient and Semantic-Enhanced Transformer Network for Point Cloud Prediction
N2M2: Learning Navigation for Arbitrary Mobile Manipulation Motions in Unseen and Dynamic Env
The Foreseeable Future: Self-Supervised Learning to Predict Dynamic Scenes for Indoor Navigati
The Treachery of Images: Bayesian Scene Keypoints for Deep Policy Learning in Robotic Manipula
PBP: Path-Based Trajectory Prediction for Autonomous Driving
Sensorless Estimation of Contact Using Deep-Learning for Human-Robot Interaction
Merging Decision Transformers: Weight Averaging for Forming Multi-Task Policies
Pedestrian Trajectory Prediction Using Dynamics-Based Deep Learning
POAQL: A Partially Observable Altruistic Q-Learning Method for Cooperative Multi-Agent Reinfor
Statler: State-Maintaining Language Models for Embodied Reasoning
Multi-Granular Transformer for Motion Prediction with LiDAR
What Matters for Active Texture Recognition with Vision-Based Tactile Sensors
Learning with Chemical versus Electrical Synapses - Does It Make a Difference?
Distill-Then-Prune: An Efficient Compression Framework for Real-Time Stereo Matching Network
Incremental 3D Reconstruction through a Hybrid Explicit-And-Implicit Representation
Sim2Real Bilevel Adaptation for Object Surface Classification Using Vision-Based Tactile Sensors
Robot Interaction Behavior Generation Based on Social Motion Forecasting for Human-Robot Inter
SPCGC: Scalable Point Cloud Geometry Compression for Machine Vision
CppFlow: Generative Inverse Kinematics for Efficient and Robust Cartesian Path Planning
Safe Deep Policy Adaptation
Physics-Informed Neural Networks for Continuum Robots: Towards Fast Approximation of Static
Fast Kinodynamic Planning on the Constraint Manifold with Deep Neural Networks
MANER: Multi-Agent Neural Rearrangement Planning of Objects in Cluttered Environments
Benchmarking Actor-Critic Deep Reinforcement Learning Algorithms for Robotics Control with Acti
Uncertainty-Driven Exploration Strategies for Online Grasp Learning
Pseudo-Labeling and Contextual Curriculum Learning for Online Grasp Learning in Robotic Bin Pic
PoseFusion: Multi-Scale Keypoint Correspondence for Monocular Camera-To-Robot Pose Estimatio
Online Fault Detection in Manipulation Tasks Via Generative Models
One-Shot Learning for Task-Oriented Grasping
Multi-Level Reasoning for Robotic Assembly: From Sequence Inference to Contact Selection
Learning to Design 3D Printable Adaptations on Everyday Objects for Robot Manipulation
Evaluating Robustness of Visual Representations for Object Assembly Task Requiring Spatio-Geo
DeRi-Bot: Learning to Collaboratively Manipulate Rigid Objects Via Deformable Objects
Visual-Policy Learning through Multi-Camera View to Single-Camera View Knowledge Distillation f
Symmetric Models for Visual Force Policy Learning
Out of Sight, Still in Mind: Reasoning and Planning about Unobserved Objects with Video Tracki
Learning to Dexterously Pick or Separate Tangled-Prone Objects for Industrial Bin Picking
Learning Fabric Manipulation in the Real World with Human Videos
HAGrasp: Hybrid Action Grasp Control in Cluttered Scenes Using Deep Reinforcement Learning
Dual-Critic Deep Reinforcement Learning for Push-Grasping Synergy in Cluttered Environment
DefGoalNet: Contextual Goal Learning from Demonstrations for Deformable Object Manipulation
Decomposing the Generalization Gap in Imitation Learning for Visual Robotic Manipulation
STOPNet: Multiview-Based 6-DoF Suction Detection for Transparent Objects on Production Lines
RoboTAP: Tracking Arbitrary Points for Few-Shot Visual Imitation
Learning Extrinsic Dexterity with Parameterized Manipulation Primitives
Learning Active Manipulation to Target Shapes with Model-Free, Long-Horizon Deep Reinforcemen
GAMMA: Generalizable Articulation Modeling and Manipulation for Articulated Objects
Efficient End-To-End Detection of 6-DoF Grasps for Robotic Bin Picking
Contact Energy Based Hindsight Experience Prioritization
ASGrasp: Generalizable Transparent Object Reconstruction and 6-DoF Grasp Detection from RG
An Offline Learning of Behavior Correction Policy for Vision-Based Robotic Manipulation
RGBManip: Monocular Image-Based Robotic Manipulation through Active Object Pose Estimation
Part-Guided 3D RL for Sim2Real Articulated Object Manipulation
MORPH: Design Co-Optimization with Reinforcement Learning Via a Differentiable Hardware Mode
Mastering Stacking of Diverse Shapes with Large-Scale Iterative Reinforcement Learning on Real
Information-Driven Affordance Discovery for Efficient Robotic Manipulation
HybGrasp: A Hybrid Learning-To-Adapt Architecture for Efficient Robot Grasping
Efficient Heatmap-Guided 6-Dof Grasp Detection in Cluttered Scenes
A Hyper-Network Based End-To-End Visual Servoing with Arbitrary Desired Poses
6-DoF Closed-Loop Grasping with Reinforcement Learning
ATPPNet: Attention Based Temporal Point Cloud Prediction Network
Transformer-CNN Cohort: Semi-Supervised Semantic Segmentation by the Best of Both Students
CrackNex: A Few-Shot Low-Light Crack Segmentation Model Based on Retinex Theory for UAV In
FBPT: A Fully Binary Point Transformer
Diffusion-Based Point Cloud Super-Resolution for mmWave Radar Data
Visual Noun Modifiers: The Problem of Binding Visual and Linguistic Cues
Cycle-Correspondence Loss: Learning Dense View-Invariant Visual Features from Unlabeled an
End-To-End RGB-D SLAM with Multi-MLPs Dense Neural Implicit Representations
Closing the Visual Sim-To-Real Gap with Object-Composable NeRFs
See to Touch: Learning Tactile Dexterity through Visual Incentives
Real-Time Contact State Estimation in Shape Control of Deformable Linear Objects under Small E
Self-Supervised Learning for Joint Pushing and Grasping Policies in Highly Cluttered Environment
A Robust Model Predictive Controller for Tactile Servoing
Harnessing the Synergy between Pushing, Grasping, and Throwing to Enhance Object Manipulation
Direct Self-Identification of Inverse Jacobians for Dexterous Manipulation through Particle Filtering
Masked Visual-Tactile Pre-Training for Robot Manipulation
Tactile Estimation of Extrinsic Contact Patch for Stable Placement
Robotic Manipulation of Hand Tools: The Case of Screwdriving
Helical Control in Latent Space: Enhancing Robotic Craniotomy Precision in Uncertain Environmen
1 kHz Behavior Tree for Self-Adaptable Tactile Insertion
DexDLO: Learning Goal-Conditioned Dexterous Policy for Dynamic Manipulation of Deformable Li
Everyday Finger: A Robotic Finger That Meets the Needs of Everyday Interactive Manipulation
Quadratic Programming Based Inverse Kinematics for Precise Bimanual Manipulation
Model-Free 3D Shape Control of Deformable Objects Using Novel Features Based on Modal Analy
Global Planning for Contact-Rich Manipulation Via Local Smoothing of Quasi-Dynamic Contact Mo
Enhancing Dexterity in Robotic Manipulation Via Hierarchical Contact Exploration
Inter-Finger Small Object Manipulation with DenseTact Optical Tactile Sensor
Dynamic Multi-Agent Deep Deterministic Policy Gradient for Autonomous Navigation of Reconfigu
FogROS2-LS: A Location-Independent Fog Robotics Framework for Latency Sensitive ROS2 Appli
Leveraging Tethers for Distributed Formation Control of Simple Robots
Distributed Differential Dynamic Programming Architectures for Large-Scale Multi-Agent Control
Accelerated K-Serial Stable Coalition for Dynamic Capture and Resource Defense
Sensor-Based Multi-Robot Coverage Control with Spatial Separation in Unstructured Environment
Localized and Incremental Probabilistic Inference for Large-Scale Networked Dynamical Systems
Comparison of Distributed Task Allocation Algorithms Considering Non-Ideal Communication Factors
Model Predictive Control for an Autonomous Underwater Robot with Fully Vectored Propulsion
Attitude Control for Morphing Quadrotor through Model Predictive Control with Constraints
NNgTL: Neural Network Guided Optimal Temporal Logic Task Planning for Mobile Robots
Synthesis of Temporally-Robust Policies for Signal Temporal Logic Tasks Using Reinforcement Le
A Deep Learning Framework for Non-Symmetrical Coulomb Friction Identification of Robotic Manip
Implicit Time Integration Simulation of Robots with Rigid Bodies and Cosserat Rods Based on a N
Efficient Constrained Dynamics Algorithms Based on an Equivalent LQR Formulation Using Gauss' P
Model-Based Co-Simulation of Flexible Mechanical Systems with Contacts Using Reduced Interfa
RIDER: Reinforcement-Based Inferred Dynamics Via Emulating Rehearsals for Robot Navigation i
Bridging the Sim-To-Real Gap with Dynamic Compliance Tuning for Industrial Insertion
Compliant Peg-In-Hole Assembly Using a Very Soft Wrist
6D Pose Estimation Based on 3D Edge Binocular Reprojection Optimization for Robotic Assembly
ASAP: Automated Sequence Planning for Complex Robotic Assembly with Physical Feasibility
Simulation-Based Approach for Automatic Roadmap Design in Multi-AGV Systems
MM4MM: Map Matching Framework for Multi-Session Mapping in Ambiguous and Perceptually-De
Learning Generalizable Patrolling Strategies through Domain Randomization of Attacker Behaviors
Combining Coordination and Independent Coverage in Multirobot Graph Patrolling
Longitudinal Control Volumes: A Novel Centralized Estimation and Control Framework for Distribu
Relaxed Hover Solution Based Control for a Bi-Copter with Rotor and Servo Stuck Failure
Aim-Aware Collision Monitoring: Discriminating between Expected and Unexpected Post-Impact B
The Voraus-AD Dataset for Anomaly Detection in Robot Applications
Multimodal Detection and Classification of Robot Manipulation Failures
FT-Net: Learning Failure Recovery and Fault-Tolerant Locomotion for Quadruped Robots
Utilizing a Malfunctioning 3D Printer by Modeling Its Dynamics with Machine Learning
A Novel Metric for Detecting Quadrotor Loss-Of-Control
Specifying and Monitoring Safe Driving Properties with Scene Graphs
Using Large Language Models to Generate and Apply Contingency Handling Procedures in Collabo
Topological Exploration Using Segmented Map with Keyframe Contribution in Subterranean Envir
A Powerline Inspection UAV Equipped with Dexterous, Lockable Gripping Mechanisms for Autonom
GIRA: Gaussian Mixture Models for Inference and Robot Autonomy
Hybrid Trajectory Optimization for Autonomous Terrain Traversal of Articulated Tracked Robots
X-ICP: Localizability-Aware LiDAR Registration for Robust Localization in Extreme Environments
Seabed Intervention with an Underwater Legged Robot
Predicting against the Flow: Boosting Source Localization by Means of Field Belief Modeling Usin
A Turning Radius Prediction Scheme for Sailing Robots under Complex Marine Environment
A Vision-Based Autonomous UAV Inspection Framework for Unknown Tunnel Construction Sites w
SOL: A Compact, Portable, Telescopic, Soft-Robotic Sun-Tracking Mechanism for Improved Solar
Measuring Ball Joint Faults in Parabolic-Trough Solar Plants with Data Augmentation and Deep Le
ECDP: Energy Consumption Disaggregation Pipeline for Energy Optimization in Lightweight Robot
Autonomous UAV Mission Cycling: A Mobile Hub Approach for Precise Landings and Continuous O
Low-To-High Resolution Path Planner for Robotic Gas Distribution Mapping
Persistent Monitoring of Large Environments with Robot Deployment Scheduling in between Remo
System Calibration of a Field Phenotyping Robot with Multiple High-Precision Profile Laser Scanne
Atmospheric Aerosol Diagnostics with UAV-Based Holographic Imaging and Computer Vision
WayFASTER: A Self-Supervised Traversability Prediction for Increased Navigation Awareness
ViTacTip: Design and Verification of a Novel Biomimetic Physical Vision-Tactile Fusion Sensor
A Large-Area Tactile Sensor for Distributed Force Sensing Using Highly Sensitive Piezoresistive
A Neuromorphic System for the Real-Time Classification of Natural Textures
An Electromagnetism-Inspired Method for Estimating In-Grasp Torque from Visuotactile Sensors
Sim-To-Real Model-Based and Model-Free Deep Reinforcement Learning for Tactile Pushing
Learning Contact for Haptic Feedback: Switching X-Lateral Teleoperators
L3 F-TOUCH: A Wireless GelSight with Decoupled Tactile and Three-Axis Force Sensing
GelLink: A Compact Multi-Phalanx Finger with Vision-Based Tactile Sensing and Proprioception
RainbowSight: A Family of Generalizable, Curved, Camera-Based Tactile Sensors for Shape Recon
Tactile Embeddings for Multi-Task Learning
AllSight: A Low-Cost and High-Resolution round Tactile Sensor with Zero-Shot Learning Capability
9DTact: A Compact Vision-Based Tactile Sensor for Accurate 3D Shape Reconstruction and Gener
GelFinger: A Novel Visual-Tactile Sensor with Multi-Angle Tactile Image Stitching
StereoTac: A Novel Visuotactile Sensor That Combines Tactile Sensing with 3D Vision
An Investigation of Multi-Feature Extraction and Super-Resolution with Fast Microphone Arrays
Model-Based Compliance Discrimination Via Soft Tactile Optical Sensing and Optical Flow Compu
Bi-Touch: Bimanual Tactile Manipulation with Sim-To-Real Deep Reinforcement Learning
AcTExplore: Active Tactile Exploration on Unknown Objects
Contrastive Learning-Based Attribute Extraction Method for Enhanced Terrain Classification
Enhancing Generalizable 6D Pose Tracking of an In-Hand Object with Tactile Sensing
UnfoldIR: Tactile Robotic Unfolding of Cloth
3D Force and Contact Estimation for a Soft-Bubble Visuotactile Sensor Using FEM
A Detachable FBG-Based Contact Force Sensor for Capturing Gripper-Vegetable Interactions
SATac: A Thermoluminescence Enabled Tactile Sensor for Concurrent Perception of Temperature,
Optimizing Multi-Touch Textile and Tactile Skin Sensing through Circuit Parameter Estimation
CushSense: Soft, Stretchable, and Comfortable Tactile Sensing Skin for Physical Human-Robot Int
Augmenting Tactile Simulators with Real-Like and Zero-Shot Capabilities
GelRoller: A Rolling Vision-Based Tactile Sensor for Large Surface Reconstruction Using Self-Su
Marker-Embedded Tactile Image Generation Via Generative Adversarial Networks
TEXterity: Tactile Extrinsic DeXterity
Optimization of Flexible Bronchoscopy Shape Sensing Using Fiber Optic Sensors
Tactile-Informed Action Primitives Mitigate Jamming in Dense Clutter
Crosstalk-Free Impedance-Separating Array Measurement for Iontronic Tactile Sensors
Visual-Tactile Learning of Garment Unfolding for Robot-Assisted Dressing
Multimodal Visual-Tactile Representation Learning through Self-Supervised Contrastive Pre-Traini
A Hierarchical Framework for Robot Safety Using Whole-Body Tactile Sensors
Robot-Camera Calibration in Tightly Constrained Environment Using Interactive Perception
Degenerate Motions of Multisensor Fusion-Based Navigation
Interaction Control for Tool Manipulation on Deformable Objects Using Tactile Feedback
Development of an Easy-To-Cut Six-Axis Force Sensor
An Ultra-Fast Intrinsic Contact Sensing Method for Medical Instruments with Arbitrary Shape
Proprioceptive-Based Whole-Body Disturbance Rejection Control for Dynamic Motions in Legged
Contact Force Estimation of Robot Manipulators with Imperfect Dynamic Model: On Gaussian Proc
Tactile Robot Programming: Transferring Task Constraints into Constraint-Based Unified Force-I
Online Modifications for Event-Based Signal Temporal Logic Specifications
Sampling-Based Reactive Synthesis for Nondeterministic Hybrid Systems
Safety Verification of Closed-Loop Control System with Anytime Perception
Model Predictive Robustness of Signal Temporal Logic Predicates
Unraveling the Single Tangent Space Fallacy: An Analysis and Clarification for Applying Riemann
Optimal Control Synthesis with Relaxed Global Temporal Logic Specifications for Homogeneous M
An Iterative Approach for Heterogeneous Multi-Agent Route Planning with Temporal Logic Goals a
Safe Networked Robotics with Probabilistic Verification
Robust MITL Planning under Uncertain Navigation Times
Exploiting Transformer in Sparse Reward Reinforcement Learning for Interpretable Temporal Logi
Stochastic Games for Interactive Manipulation Domains
Active Inference for Reactive Temporal Logic Motion Planning
Fast Task Allocation of Heterogeneous Robots with Temporal Logic and Inter-Task Constraints
Skill Transfer for Temporal Task Specification
High Precision Paint Deposition Modeling Considering Variable Posture of Spray Painting Robot
Verifiable Learned Behaviors Via Motion Primitive Composition: Applications to Scooping of Granu
Knowledge Acquisition Plans: Generation, Combination, and Execution
Grasp It Like a Pro 2.0: A Data-Driven Approach Exploiting Basic Shapes Decomposition and Hu
Visual-Tactile Fusion for Transparent Object Grasping in Complex Backgrounds
Variable Stiffness Soft Robotic Fingers Using Snap-Fit Kinematic Reconfiguration
Grasp Transfer Based on Self-Aligning Implicit Representations of Local Surfaces
Amortized Inference for Efficient Grasp Model Adaptation
Learning Realistic and Reasonable Grasps for Anthropomorphic Hand in Cluttered Scenes
FuncGrasp: Learning Object-Centric Neural Grasp Functions from Single Annotated Example Obje
Physical and Digital Adversarial Attacks on Grasp Quality Networks
Scaling Object-Centric Robotic Manipulation with Multimodal Object Identification
Grasp Manipulation Relationship Detection Based on Graph Sample and Aggregation
Acoustic Soft Tactile Skin (AST Skin)
Domain Randomization for Sim2real Transfer of Automatically Generated Grasping Datasets
Kinematic Synergy Primitives for Human-Like Grasp Motion Generation
VFAS-Grasp: Closed Loop Grasping with Visual Feedback and Adaptive Sampling
The Fractal Hand-II: Reviving a Classic Mechanism for Contemporary Grasping Challenges
ICGNet: A Unified Approach for Instance-Centric Grasping
The Grasp Reset Mechanism: An Automated Apparatus for Conducting Grasping Trials
Model-Based Runtime Monitoring with Interactive Imitation Learning
InterRep: A Visual Interaction Representation for Robotic Grasping
MoDem-V2: Visuo-Motor World Models for Real-World Robot Manipulation
Towards Feasible Dynamic Grasping: Leveraging Gaussian Process Distance Field, SE(3) Equiva
A Surprisingly Efficient Representation for Multi-Finger Grasping
GrainGrasp: Dexterous Grasp Generation with Fine-Grained Contact Guidance
Regrasping on Printed Circuit Boards with the Smart Suction Cup
Anthropomorphic Grasping with Neural Object Shape Completion
Statistical Stratification and Benchmarking of Robotic Grasping Performance
The Hydra Hand: A Mode-Switching Underactuated Gripper with Precision and Power Grasping M
Single-Motor Robotic Gripper with Three Functional Modes for Grasping in Confined Spaces
A Force-Controlled Gripper Capable of Measuring Mechanical Properties of an Object
Optimal Design of a Highly Self-Adaptive Gripper with Multi-Phalange Compliant Fingers for Grasp
Accelerating Robotic Picking of Rigid Objects with a Compliant Pneumatic Gripper and an Impact-
Vertical Vibratory Transport of Grasped Parts Using Impacts
Bionic Soft Fingers with Hybrid Variable Stiffness Mechanisms for Multimode Grasping
Design and Fabrication of a Novel Miniature Magnetic Gripper
Design of Highly Repeatable and Multi-Functional Grippers for Precision Handling with Articulated
Generalized Partially Destructive Disassembly Planning for Robotic Disassembly
The Fractal Hand--I: A Non-Anthropomorphic, but Synergistic, Adaptable Gripper
The Double-Scoop Gripper: A Tendon-Driven Soft-Rigid End-Effector for Food Handling Exploiting
Co-Designing Manipulation Systems Using Task-Relevant Constraints
Squirrel-Inspired Tendon-Driven Passive Gripper for Agile Landing
HASHI: Highly Adaptable Seafood Handling Instrument for Manipulation in Industrial Settings
All the Feels: A Dexterous Hand with Large-Area Tactile Sensing
Soft and Rigid Object Grasping with Cross-Structure Hand Using Bilateral Control-Based Imitation
GRASP: Grocery Robot’s Adhesion and Suction Picker
Improved Generalization of Probabilistic Movement Primitives for Manipulation Trajectories
X-Tacformer : Spatio-Tempral Attention Model for Tactile Recognition
The Joint-Space Reconstruction of Human Fingers by Using a Highly Under-Actuated Exoskeleton
Prosthetic Upper-Limb Sensory Enhancement (PULSE): A Dual Haptic Feedback Device in a Prost
Point-Wise Vibration Pattern Production Via a Sparse Actuator Array for Surface Tactile Feedback
Implementation and Assessment of an Augmented Training Curriculum for Surgical Robotics
Hapstick: A Soft Flexible Joystick for Stiffness Rendering Via Fiber Jamming
Fingertip Ultrasonic Array for Tactile Rendering
Active Exploration for Real-Time Haptic Training
A Multi-Stable Curved Line Shape Display
VIDAR: Data Quality Improvement for Monocular 3D Reconstruction through In-Situ Visual Interact
Transparency Control of a 1-DoF Knee Exoskeleton Via Human-In-The-Loop Velocity Optimisation
Towards Enhanced Human Activity Recognition for Real-World Human-Robot Collaboration
Self-Supervised Regression of sEMG Signals Combining Non-Negative Matrix Factorization with D
Maximising Coefficiency of Human-Robot Handovers through Reinforcement Learning
Jacquard V2: Refining Datasets Using the Human in the Loop Data Correction Method
Decision Making for Human-In-The-Loop Robotic Agents Via Uncertainty-Aware Reinforcement Le
Building User Proficiency in Piloting Small Unmanned Aerial Vehicles (sUAV)
A Probabilistic Model for Cobot Decision Making to Mitigate Human Fatigue in Repetitive Co-Manip
SEQUEL: Semi-Supervised Preference-Based RL with Query Synthesis Via Latent Interpolation
Learning When to Ask for Help: Efficient Interactive Navigation Via Implicit Uncertainty Estimation
JaywalkerVR: A VR System for Collecting Safety-Critical Pedestrian-Vehicle Interactions
Human Preference-Aware Rebalancing and Charging for Shared Electric Micromobility Vehicles
A Trajectory-Based Flight Assistive System for Novice Pilots in Drone Racing Scenario
A Power-Aware Control Strategy for an Elbow Effort-Compensation Device
A Planar Compliant Contact Control Applied to Multi-Dimensional Elastic Gripper for Unexpected C
Enabling Passivity for Cartesian Workspace Restrictions
Virtual Borders in 3D: Defining a Drone’s Movement Space Using Augmented Reality
Trajectory Prediction for Robot Navigation Using Flow-Guided Markov Neural Operator
Stranger Danger! Identifying and Avoiding Unpredictable Pedestrians in RL-Based Social Robot Na
Robot Navigation in Risky, Crowded Environments: Understanding Human Preferences
Real-Time Human Presence Estimation for Indoor Robots
MAC-ID: Multi-Agent Reinforcement Learning with Local Coordination for Individual Diversity
Interactive Joint Planning for Autonomous Vehicles
Improve Computing Efficiency and Motion Safety by Analyzing Environment with Graphics
Generating Environment-Based Explanations of Motion Planner Failure: Evolutionary and Joint-Opt
Self-Supervised 6-DoF Robot Grasping by Demonstration Via Augmented Reality Teleoperation S
Trust Recognition in Human-Robot Cooperation Using EEG
Multimodal Multi-User Surface Recognition with the Kernel Two-Sample Test
Learning User Preferences for Complex Cobotic Tasks: Meta-Behaviors and Human Groups
Learning Self-Confidence from Semantic Action Embeddings for Improved Trust in Human-Robot In
Interactive Navigation in Environments with Traversable Obstacles Using Large Language and Vi
From Unstable Electrode Contacts to Reliable Control: A Deep Learning Approach for HD-sEMG in
Enhanced Human-Robot Collaboration with Intent Prediction Using Deep Inverse Reinforcement L
ToP-ToM: Trust-Aware Robot Policy with Theory of Mind
Shaping Social Robot to Play Games with Human Demonstrations and Evaluative Feedback
Running Guidance for Visually Impaired People Using Sensory Augmentation Technology Based
Rider Cooperative Control of Rear-Wheel-Swing Motorcycle Based on Divergent Component of Mo
MORPHeus: A Multimodal One-Armed Robot-Assisted Peeling System with Human Users In-The-
MIntNet: Rapid Motion Intention Forecasting of Coupled Human-Robot Systems with Simulation-T
Language to Map: Topological Map Generation from Natural Language Path Instructions
Human-Centered Autonomy for UAS Target Search
Gaze-Based Human-Robot Interaction System for Infrastructure Inspections
Facile Integration of Robots into Experimental Orchestration at Scientific User Facilities
Usability Evaluation Framework for Close-Proximity Collaboration with Large Industrial Manipulator
MyoPassivity Map: Does Multi-Channel sEMG Correlate with the Energetic Behavior of Upper-Lim
Language-Guided Active Sensing of Confined, Cluttered Environments Via Object Rearrangement
Feedforward Control of Lower Limb Exoskeletons: Which Torque Profile Should We Use?
Design of Two Morphing Robot Surfaces and Results from a User Study on What People Want
Automatic Trust Estimation from Movement Data in Industrial Human-Robot Collaboration Based
A Dual Closed-Loop Control Strategy for Human-Following Robots Respecting Social Space
A Bayesian Optimization Framework for the Automatic Tuning of MPC-Based Shared Controllers
Trust-Aware Motion Planning for Human-Robot Collaboration under Distribution Temporal Logic Spe
Towards Proactive Safe Human-Robot Collaborations Via Data-Efficient Conditional Behavior Predi
Risk-Bounded Online Team Interventions Via Theory of Mind
Human-Robot Complementary Collaboration for Flexible and Precision Assembly
HAC-SLAM: Human Assisted Collaborative 3D-SLAM through Augmented Reality
GAN-Based Semi-Supervised Training of LSTM Nets for Intention Recognition in Cooperative Tas
CoBT: Collaborative Programming of Behaviour Trees from One Demonstration for Robot Manipula
A Dynamic Planner for Safe and Predictable Human-Robot Collaboration
Towards Human-Robot Collaborative Surgery: Trajectory and Strategy Learning in Bimanual Peg T
SIREN: Underwater Robot-To-Human Communication Using Audio
Shared Autonomy Via Variable Impedance Control and Virtual Potential Fields for Encoding Huma
Robustifying a Policy in Multi-Agent RL with Diverse Cooperative Behaviors and Adversarial Style
Human Robot Shared Control in Surgery: A Performance Assessment
Distilling and Retrieving Generalizable Knowledge for Robot Manipulation Via Language Correctio
An Intuitive Manual Guidance Scheme to Operate Rotation and Translation Simultaneously
An Ergo-Interactive Framework for Human-Robot Collaboration Via Learning from Demonstration
A User-Centered Shared Control Scheme with Learning from Demonstration for Robotic Surgery
Robust and Dexterous Dual-Arm Tele-Cooperation Using Adaptable Impedance Control
PlanCollabNL: Leveraging Large Language Models for Adaptive Plan Generation in Human-Robot
Multi-Agent Strategy Explanations for Human-Robot Collaboration
Efficient ISO/TS 15066 Compliance through Model Predictive Control
Dual-Mode Human-Robot Collaboration with Guaranteed Safety Using Time-Varying Zeroing Contro
A Time-Optimal Energy Planner for Safe Human-Robot Collaboration
Discuss before Moving: Visual Language Navigation Via Multi-Expert Discussions
Personality and Memory-Based Software Framework for Human-Robot Interaction
POLITE: Preferences Combined with Highlights in Reinforcement Learning
CoFRIDA: Self-Supervised Fine-Tuning for Human-Robot Co-Painting
MateRobot: Material Recognition in Wearable Robotics for People with Visual Impairments
Robot-Assisted Navigation for Visually Impaired through Adaptive Impedance and Path Planning
Incremental Learning of Full-Pose Via-Point Movement Primitives on Riemannian Manifolds
Supernumerary Robotic Limbs to Support Post-Fall Recoveries for Astronauts
Robust Body Exposure (RoBE): A Graph-Based Dynamics Modeling Approach to Manipulating Bla
Recency Bias in Task Performance History Affects Perceptions of Robot Competence and Trustwo
LaCE-LHMP: Airflow Modelling-Inspired Long-Term Human Motion Prediction by Enhancing Lamina
Interactive Continual Learning Architecture for Long-Term Personalization of Home Service Robot
Human-Robot Interactive Creation of Artistic Portrait Drawings
High Stimuli Virtual Reality Training for a Brain Controlled Robotic Wheelchair
Automatic Captioning Based on Visible and Infrared Images
A Semi-Automatic Oriental Ink Painting Framework for Robotic Drawing from 3D Models
A 3D Mixed Reality Interface for Human-Robot Teaming
Reinforcement Learning with Energy-Exchange Dynamics for Spring-Loaded Biped Robot Walking
Foot Shape-Dependent Resistive Force Model for Bipedal Walkers on Granular Terrains
Adaptive Passive Biped Dynamic Walking on Unknown Uneven Terrain
HumanMimic: Learning Natural Locomotion and Transitions for Humanoid Robot Via Wasserstein Ad
Online Adaptive Motion Generation for Humanoid Locomotion on Non-Flat Terrain Via Template B
A Bio-Plausible Approach to Realizing Heat-Evoked Nociceptive Withdrawal Reflex on the Upper
Fall Prediction for Bipedal Robots: The Standing Phase
Shape-Changing Robotic Mannequin Shoulder with Bio-Inspired Layered Structure
UKF-Based Sensor Fusion for Joint-Torque Sensorless Humanoid Robots
S2R-ViT for Multi-Agent Cooperative Perception: Bridging the Gap from Simulation to Reality
Eliminating Cross-Modal Conflicts in BEV Space for LiDAR-Camera 3D Object Detection
EMIFF: Enhanced Multi-Scale Image Feature Fusion for Vehicle-Infrastructure Cooperative 3D Obj
Vehicle Intention Classification Using Visual Clues
CFDNet: A Generalizable Foggy Stereo Matching Network with Contrastive Feature Distillation
Multi-Class Road Defect Detection and Segmentation Using Spatial and Channel-Wise Attention
HPL-ViT: A Unified Perception Framework for Heterogeneous Parallel LiDARs in V2V
FastOcc: Accelerating 3D Occupancy Prediction by Fusing the 2D Bird's-Eye View and Perspectiv
LPFormer: LiDAR Pose Estimation Transformer with Multi-Task Network
MonoOcc: Digging into Monocular Semantic Occupancy Prediction
ShaSTA: Modeling Shape and Spatio-Temporal Affinities for 3D Multi-Object Tracking
Breaking Data Silos: Cross-Domain Learning for Multi-Agent Perception from Independent Private
AdvGPS: Adversarial GPS for Multi-Agent Perception Attack
Towards Motion Forecasting with Real-World Perception Inputs: Are End-To-End Approaches Comp
QUEST: Query Stream for Practical Cooperative Perception
Towards Visibility Estimation and Noise-Distribution-Based Defogging for LiDAR in Autonomous Dr
CenterCoop: Center-Based Feature Aggregation for Communication-Efficient Vehicle-Infrastructur
Probabilistic 3D Multi-Object Cooperative Tracking for Autonomous Driving Via Differentiable Multi
Overparametrization Helps Offline-To-Online Generalization of Closed-Loop Control from Pixels
Hierarchical Human-To-Robot Imitation Learning for Long-Horizon Tasks Via Cross-Domain Skill A
Policy Optimization by Looking Ahead for Model-Based Offline RL
DINOBot: Robot Manipulation Via Retrieval and Alignment with Vision Foundation Models
Rank2Reward: Learning Shaped Reward Functions from Passive Video
A Generalized Acquisition Function for Preference-Based Reward Learning
Human-Robot Deformation Manipulation Skill Transfer: Sequential Fabric Unfolding Method for Ro
Model Optimization in Deep Learning Based Robot Control for Autonomous Driving
Quasi-Static Soft Fixture Analysis of Rigid and Deformable Objects
In-Hand Rolling Manipulation Based on Ball-On-Cloth System
A Linkage-Driven Underactuated Robotic Hand for Adaptive Grasping and In-Hand Manipulation
Curriculum-Based Sensing Reduction in Simulation to Real-World Transfer for In-Hand Manipulati
Geometric Fabrics: A Safe Guiding Medium for Policy Learning
Robust In-Hand Manipulation with Extrinsic Contacts
Dexterous In-Hand Manipulation by Guiding Exploration with Simple Sub-Skill Controllers
Robot Synesthesia: In-Hand Manipulation with Visuotactile Sensing
Adaptive Fingers Coordination for Robust Grasp and In-Hand Manipulation under Disturbances
A Large-Scale Suction-Based Climbing Parallel Robot for Wall Painting Application
A Collision-Aware Cable Grasping Method in Cluttered Environment
A Simple Computationally Efficient Path ILC for Industrial Robotic Manipulators
RoboGrind: Intuitive and Interactive Surface Treatment with Industrial Robots
An LLM-Driven Framework for Multiple-Vehicle Dispatching and Navigation in Smart City Landsca
SCRNet: A Retinex Structure-Based Low-Light Enhancement Model Guided by Spatial Consistenc
Autonomous Field-Of-View Adjustment Using Adaptive Kinematic Constrained Control with Robot
RoSSO: A High-Performance Python Package for Robotic Surveillance Strategy Optimization Usi
Semi-Autonomous Surface-Tracking Tasks Using Omnidirectional Mobile Manipulators
Probably Approximately Correct Nonlinear Model Predictive Control (PAC-NMPC)
QuAD: Query-Based Interpretable Neural Motion Planning for Autonomous Driving
Safe Receding Horizon Motion Planning with Infinitesimal Update Interval
NPC: Neural Predictive Control for Fuel-Efficient Autonomous Trucks
Robustified Time-Optimal Collision-Free Motion Planning for Autonomous Mobile Robots under Di
Learning-Aided Warmstart of Model Predictive Control in Uncertain Fast-Changing Traffic
Co-Learning Planning and Control Policies Constrained by Differentiable Logic Specifications
Output-Sampled Model Predictive Path Integral Control (o-MPPI) for Increased Efficiency
The Virtues of Laziness: Multi-Query Kinodynamic Motion Planning with Lazy Methods
A Tube-Based Reinforcement Learning Approach for Optimal Motion Planning in Unknown Works
Task-Oriented Active Learning of Model Preconditions for Inaccurate Dynamics Models
Risk-Predictive Planning for Off-Road Autonomy
BeBOP -- Combining Reactive Planning and Bayesian Optimization to Solve Robotic Manipulation
Motion Memory: Leveraging past Experiences to Accelerate Future Motion Planning
Mitigating Causal Confusion in Vector-Based Behavior Cloning for Safer Autonomous Planning
Learning-Based Motion Planning with Mixture Density Networks
Subgoal Diffuser: Coarse-To-Fine Subgoal Generation to Guide Model Predictive Control for Robot
Motion Planning As Online Learning: A Multi-Armed Bandit Approach to Kinodynamic Sampling-Ba
Fixture Calibration with Guaranteed Bounds from a Few Correspondence-Free Surface Points
Data-Driven Virtual Sensing for Probabilistic Condition Monitoring of Solenoid Valves
Digital Robot Judge (DR.J): Building a Task-Centric Performance Database of Real-World Manipula
Semi-Analytical Design of PDE Endpoint Controller for Flexible Manipulator with Non-Homogenou
Automated Sewing System Enabled by Machine Vision for Smart Garment Manufacturing
Segmentation and Coverage Planning of Freeform Geometries for Robotic Surface Finishing
Integrating Robot Assignment and Maintenance Management: A Multi-Agent Reinforcement Learning
Towards Fault-Tolerant Deployment of Mobile Robot Navigation in the Edge: An Experimental Stu
Towards Optimal Lane-Changing Coordination of CAVs in Multi-Lane Mixed Traffic Scenarios
Reducing Non-IID Effects in Federated Autonomous Driving with Contrastive Divergence Loss
ODD-Based Query-Time Scenario Mutation Framework for Autonomous Driving Scenario Databas
Cooperation for Scalable Supervision of Autonomy in Mixed Traffic
Hierarchical Learned Risk-Aware Planning Framework for Human Driving Modeling
DESTINE: Dynamic Goal Queries with Temporal Transductive Alignment for Trajectory Prediction
Parallel Optimization with Hard Safety Constraints for Cooperative Planning of Connected Autono
Editing Driver Character: Socially-Controllable Behavior Generation for Interactive Traffic Simulatio
Cognitive-Digital-Twin-Based Driving Assistance
A Safety-Adapted Loss for Pedestrian Detection in Autonomous Driving
PCB-RandNet: Rethinking Random Sampling for LiDAR Semantic Segmentation in Autonomous Dr
STT: Stateful Tracking with Transformers for Autonomous Driving
SmartCooper: Vehicle Collaborative Perception under Adaptive Fusion and Judger Mechanism
A Neural-Evolutionary Algorithm for Autonomous Transit Network Design
UDE-Based Robust Control of a Quadrotor-Slung-Load System
Are You a Robot? Detecting Autonomous Vehicles from Behavior Analysis
RaTrack: Moving Object Detection and Tracking with 4D Radar Point Cloud
Mixed Traffic Control and Coordination from Pixels
Prompting Multi-Modal Tokens to Enhance End-To-End AutonomousDriving Imitation Learning wit
Efficient and Differentiable Joint Conditional Prediction and Cost Evaluation for Tree-Structured P
SIMMF: Semantics-Aware Interactive Multiagent Motion Forecasting for Autonomous Vehicle Drivi
CausalAgents: A Robustness Benchmark for Motion Forecasting
Highway-Driving with Safe Velocity Bounds on Occluded Traffic
Generalizing Cooperative Eco-Driving Via Multi-Residual Task Learning
Approximate Multiagent Reinforcement Learning for On-Demand Urban Mobility Problem on a Lar
Continual Driving Policy Optimization with Closed-Loop Individualized Curricula
Task-Driven Domain-Agnostic Learning with Information Bottleneck for Autonomous Steering
MacFormer: Map-Agent Coupled Transformer for Real-Time and Robust Trajectory Prediction
Real-Time Capable Decision Making for Autonomous Driving Using Reachable Sets
Vehicle Behavior Prediction by Episodic-Memory Implanted NDT
Optimal Driver Warning Generation in Dynamic Driving Environment
Active Learning with Dual Model Predictive Path-Integral Control for Interaction-Aware Autono
Informed Reinforcement Learning for Situation-Aware Traffic Rule Exceptions
Chance-Aware Lane Change with High-Level Model Predictive Control through Curriculum Reinfor
Human Observation-Inspired Trajectory Prediction for Autonomous Driving in Mixed-Autonomy Tra
Context-Aware Timewise VAEs for Real-Time Vehicle Trajectory Prediction
Subject-Independent Estimation of Continuous Movements Using CNN-LSTM for a Home-Based Up
Robot Trajectron: Trajectory Prediction-Based Shared Control for Robot Manipulation
On the Feasibility of EEG-Based Motor Intention Detection for Real-Time Robot Assistive Control
Microexpression to Macroexpression: Facial Expression Magnification by Single Input
Looking Inside Out: Anticipating Driver Intent from Videos
CARTIER: Cartographic lAnguage Reasoning Targeted at Instruction Execution for Robots
A Novel Hybrid Unsupervised Domain Adaptation Method for Cross-Subject Joint Angle Estimatio
A Novel Benchmarking Paradigm and a Scale and Motion-Aware Model for Egocentric Pedestrian Tr
A 3D Vector Field and Gaze Data Fusion Framework for Hand Motion Intention Prediction in Huma
Perching and Grasping Using a Passive Dynamic Bioinspired Gripper
Self-Sensing Feedback Control of an Electrohydraulic Robotic Shoulder
Design and Validation of a Variable Stiffness Spiral Cam Actuator
Hybrid Force-Position Control of an Elastic Tendon-Driven Scrubbing Robot (TEDSR)
Investigation on the Multi-Solution Problem of the Kinetostatics of Cable-Driven Continuum Manipu
Optimization Design Method of Tendon-Sheath Transmission Path under Curvature Constraint
Stability Analysis of Tendon Driven Continuum Robots and Application to Active Softening
Elasto-Static Modelling and Identification of a Deployable Cable-Driven Parallel Robot with Compli
Torque Transmission in Double-Tendon Sheath Driven Actuators for Application in Exoskeletons
Accurate Kinematic Modeling Using Autoencoders on Differentiable Joints
A Miniature Water Jumping Robot Based on Accurate Interaction Force Analysis
Jerk-Limited Traversal of One-Dimensional Paths and Its Application to Multi-Dimensional Path Tr
The Kinematics of Constant Curvature Continuum Robots through Three Segments
An Analytic Solution to the 3D CSC Dubins Path Problem
Polytope-Based Continuous Scalar Performance Measure with Analytical Gradient for Effective Ro
Kinematic Optimization of a Robotic Arm for Automation Tasks with Human Demonstration
Enhancing Motion Trajectory Segmentation of Rigid Bodies Using a Novel Screw-Based Trajector
Model Reduction in Soft Robotics Using Locally Volume-Preserving Primitives
Motion Planning and Inertia Based Control for Impact Aware Manipulation
RASCAL: A Scalable, High-Redundancy Robot for Automated Storage and Retrieval Systems
Virtual Passive-Joint Space Based Time-Optimal Trajectory Planning for a 4-DOF Parallel Manipul
Direct Kinematic Singularities and Stability Analysis of Sagging Cable-Driven Parallel Robots
Towards Solving Cable-Driven Parallel Robot Inaccuracy Due to Cable Elasticity
Wrench and Twist Capability Analysis for Cable-Driven Parallel Robots with Consideration of the
RicMonk: A Three-Link Brachiation Robot with Passive Grippers for Energy-Efficient Brachiation
Gaussian Process-Enhanced, External and Internal Convertible Form-Based Control of Underactu
Synthesize Efficient Safety Certificates for Learning-Based Safe Control Using Magnitude Regulari
On the Optimality, Stability, and Feasibility of Control Barrier Functions: An Adaptive Learning-B
Learning Failure Prevention Skills for Safe Robot Manipulation
GG-LLM: Geometrically Grounding Large Language Models for Zero-Shot Human Activity Forecas
Modality Attention for Prediction-Based Robot Motion Generation: Improving Interpretability and R
Adaptive Whole-Body Robotic Tool-Use Learning on Low-Rigidity Plastic-Made Humanoids Using V
Generating and Transferring Priors for Causal Bayesian Network Parameter Estimation in Robotic
Training Diverse High-Dimensional Controllers by Scaling Covariance Matrix Adaptation MAP-Anne
Robotic Constrained Imitation Learning for the Peg Transfer Task in Fundamentals of Laparoscopi
Stable Motion Primitives Via Imitation and Contrastive Learning
GAN-Based Editable Movement Primitive from High-Variance Demonstrations
One-Shot Imitation Learning with Graph Neural Networks for Pick-And-Place Manipulation Tasks
Unsupervised Human Motion Segmentation Based on Characteristic Force Signals of Contact Eve
Robotic Skill Mutation in Robot-To-Robot Propagation During a Physically Collaborative Sawing Ta
Few-Shot Learning of Force-Based Motions from Demonstration through Pre-Training of Haptic Re
Learning-Based Risk-Bounded Path Planning under Environmental Uncertainty
ConBaT: Control Barrier Transformer for Safe Robot Learning from Demonstrations
Unsupervised Learning of Neuro-Symbolic Rules for Generalizable Context-Aware Planning in Ob
A Combination of a Controllable Clutch and an Oscillating Slider Crank Mechanism for Ease of Dir
WayEx: Waypoint Exploration Using a Single Demonstration
Generating Robotic Elliptical Excisions with Human-Like Tool-Tissue Interactions
Fitting Parameters of Linear Dynamical Systems to Regularize Forcing Terms in Dynamical Movem
AirExo: Low-Cost Exoskeletons for Learning Whole-Arm Manipulation in the Wild
Bayesian Constraint Inference from User Demonstrations Based on Margin-Respecting Preferenc
Instructing Robots by Sketching: Learning from Demonstration Via Probabilistic Diagrammatic Tea
Learning Distributional Demonstration Spaces for Task-Specific Cross-Pose Estimation
Globally Stable Neural Imitation Policies
Learning Barrier-Certified Polynomial Dynamical Systems for Obstacle Avoidance with Robots
Domain Adaptation of Visual Policies with a Single Demonstration
Learning Complex Motion Plans Using Neural ODEs with Safety and Stability Guarantees
Learning a Stable Dynamic System with a Lyapunov Energy Function for Demonstratives Using Ne
Learning a Flexible Neural Energy Function with a Unique Minimum for Globally Stable and Accur
Inverse Constraint Learning and Generalization by Transferable Reward Decomposition
Learning Robot Motion in a Cluttered Environment Using Unreliable Human Skeleton Data Collec
IQL-TD-MPC: Implicit Q-Learning for Hierarchical Model Predictive Control
SLIM: Skill Learning with Multiple Critics
Learning Robotic Milling Strategies Based on Passive Variable Operational Space Interaction Cont
SPRINT: Scalable Policy Pre-Training Via Language Instruction Relabeling
Effective Representation Learning Is More Effective in Reinforcement Learning Than You Think
Learning Highly Dynamic Behaviors for Quadrupedal Robots
TWIST: Teacher-Student World Model Distillation for Efficient Sim-To-Real Transfer
Learning Vision-Based Pursuit-Evasion Robot Policies
Weighting Online Decision Transformer with Episodic Memory for Offline-To-Online Reinforcement
COMPOSER: Scalable and Robust Modular Policies for Snake Robots
Barrier Functions Inspired Reward Shaping for Reinforcement Learning
AdaptAUG: Adaptive Data Augmentation Framework for Multi-Agent Reinforcement Learning
HyperPPO: A Scalable Method for Finding Small Policies for Robotic Control
Grow Your Limits: Continuous Improvement with Real-World RL for Robotic Locomotion
Torque-Based Deep Reinforcement Learning for Task-And-Robot Agnostic Learning on Bipedal Ro
Decentralized Motor Skill Learning for Complex Robotic Systems
TartanDrive 2.0: More Modalities and Better Infrastructure to Further Self-Supervised Learning Re
EnYOLO: A Real-Time Framework for Domain-Adaptive Underwater Object Detection with Image
GS-PKNN: An Efficient and High-Fidelity Mobility Prediction Method for Unmanned Ground Vehicl
UNRealNet: Learning Uncertainty-Aware Navigation Features from High-Fidelity Scans of Real En
Geo-Localization with Transformer-Based 2D-3D Match Network
Robot-Dependent Traversability Estimation for Outdoor Environments Using Deep Multimodal Vari
F3DMP: Foresighted 3D Motion Planning of Mobile Robots in Wild Environments
MATRIX: Multi-Agent Trajectory Generation with Diverse Contexts
Efficient Multi-Task and Transfer Reinforcement Learning with Parameter-Compositional Framewo
Goal-Conditioned Reinforcement Learning with Disentanglement-Based Reachability Planning
KINet: Unsupervised Forward Models for Robotic Pushing Manipulation
Intrinsic Language-Guided Exploration for Complex Long-Horizon Robotic Manipulation Tasks
Touch-Based Manipulation with Multi-Fingered Robot Using Off-Policy RL and Temporal Contrasti
Learning Language-Conditioned Deformable Object Manipulation with Graph Dynamics
Grasp Anything: Combining Teacher-Augmented Policy Gradient Learning with Instance Segmentati
Composable Interaction Primitives: A Structured Policy Class for Efficiently Learning Sustained-Con
Learning to Catch Reactive Objects with a Behavior Predictor
Enhancing Task Performance of Learned Simplified Models Via Reinforcement Learning
Leveraging the Efficiency of Multi-Task Robot Manipulation Via Task-Evoked Planner and Reinfor
Generalize by Touching: Tactile Ensemble Skill Transfer for Robotic Furniture Assembly
Sim2Real Manipulation on Unknown Objects with Tactile-Based Reinforcement Learning
Synchronized Dual-Arm Rearrangement Via Cooperative MTSP
EquivAct: SIM(3)-Equivariant Visuomotor Policies Beyond Rigid Object Manipulation
Trade-Off between Robustness and Rewards Adversarial Training for Deep Reinforcement Learnin
Multi Actor-Critic DDPG for Robot Action Space Decomposition: A Framework to Control Large 3D
BioSLAM: A Bio-Inspired Lifelong Memory System for General Place Recognition
Efficient Hierarchical Reinforcement Learning for Mapless Navigation with Predictive Neighbourin
Learning Diverse Skills for Local Navigation under Multi-Constraint Optimality
Snake Robot with Tactile Perception Navigates on Large-Scale Challenging Terrain
RaLF: Flow-Based Global and Metric Radar Localization in LiDAR Maps
VPE-SLAM: Neural Implicit Voxel-Permutohedral Encoding for SLAM
Zero-Shot Wireless Indoor Navigation through Physics-Informed Reinforcement Learning
An Environmental-Complexity-Based Navigation Method Based on Hierarchical Deep Reinforceme
Pre-Trained Masked Image Model for Mobile Robot Navigation
Human-Robot Gym: Benchmarking Reinforcement Learning in Human-Robot Collaboration
Improving the Generalization of Unseen Crowd Behaviors for Reinforcement Learning Based Local
Human-Aligned Longitudinal Control for Occluded Pedestrian Crossing with Visual Attention
Projection-Based Fast and Safe Policy Optimization for Reinforcement Learning
Symmetry Considerations for Learning Task Symmetric Robot Policies
Learning Dual-Arm Object Rearrangement for Cartesian Robots
Guided Online Distillation: Promoting Safe Reinforcement Learning by Offline Demonstration
Sample-Efficient Learning to Solve a Real-World Labyrinth Game Using Data-Augmented Model-B
Active Neural Topological Mapping for Multi-Agent Exploration
Seeing through the Grass: Semantic Pointcloud Filter for Support Surface Learning
Manipulator As a Tail: Promoting Dynamic Stability for Legged Locomotion
Two-Stage Learning of Highly Dynamic Motions with Rigid and Articulated Soft Quadrupeds
High-Dimensional Controller Tuning through Latent Representations
Expert Composer Policy: Scalable Skill Repertoire for Quadruped Robots
Learning Agile Bipedal Motions on a Quadrupedal Robot
LAGOON: Language-Guided Motion Control
Learning Quadrupedal Locomotion with Impaired Joints Using Random Joint Masking
Multi-Task Learning of Active Fault-Tolerant Controller for Leg Failures in Quadruped Robots
Cascaded Compositional Residual Learning for Complex Interactive Behaviors
Deep Compliant Control for Legged Robots
Imitating and Finetuning Model Predictive Control for Robust and Symmetric Quadrupedal Locomo
Learning Agile Locomotion and Adaptive Behaviors Via RL-Augmented MPC
Extreme Parkour with Legged Robots
Learning Risk-Aware Quadrupedal Locomotion Using Distributional Reinforcement Learning
Robust Quadrupedal Locomotion Via Risk-Averse Policy Learning
Maximizing Quadruped Velocity by Minimizing Energy
Dexterous Legged Locomotion in Confined 3D Spaces with Reinforcement Learning
Walking-By-Logic: Signal Temporal Logic-Guided Model Predictive Control for Bipedal Locomotion R
Seamless Reaction Strategy for Bipedal Locomotion Exploiting Real-Time Nonlinear Model Predicti
Synthesizing Robust Walking Gaits Via Discrete-Time Barrier Functions with Application to Multi-
Efficient, Dynamic Locomotion through Step Placement with Straight Legs and Rolling Contacts
Unified Motion Planner for Walking, Running, and Jumping Using the Three-Dimensional Divergen
Data-Driven Latent Space Representation for Robust Bipedal Locomotion Learning
LIKO: LiDAR, Inertial, and Kinematic Odometry for Bipedal Robots
Barry: A High-Payload and Agile Quadruped Robot
Terrestrial Locomotion of PogoX: From Hardware Design to Energy Shaping and Step-To-Step Dy
Learning Emergent Gaits with Decentralized Phase Oscillators: On the Role of Observations, Re
Bio-Inspired Gait Transitions for Quadruped Locomotion
Optimizing Dynamic Balance in a Rat Robot Via the Lateral Flexion of a Soft Actuated Spine
SLoMo: A General System for Legged Robot Motion Imitation from Casual Videos
Introducing the Carpal-Claw: A Mechanism to Enhance High-Obstacle Negotiation for Quadruped
SpaceHopper: A Small-Scale Legged Robot for Exploring Low-Gravity Celestial Bodies
ManyQuadrupeds: Learning a Single Locomotion Policy for Diverse Quadruped Robots
Investigating Stability Outcomes across Diverse Gait Patterns in Quadruped Robots: A Comparativ
Pedipulate: Enabling Manipulation Skills Using a Quadruped Robot's Leg
An Efficient Model Based Approach on Learning Agile Motor Skills without Reinforcement
Adaptive Model Predictive Control with Data-Driven Error Model for Quadrupedal Locomotion
Layered Control for Cooperative Locomotion of Two Quadrupedal Robots: Centralized and Distrib
RL + Model-Based Control: Using On-Demand Optimal Control to Learn Versatile Legged Locomot
Generation of Steady Wheel Gait for Planar X-Shaped Walker with Reaction Wheel
Trajectory Optimization Strategy That Considers Body Tip-Over Stability, Limb Dynamics, and Mot
Robust Jumping with an Articulated Soft Quadruped Via Trajectory Optimization and Iterative Lear
Unlocking Versatile Locomotion: A Novel Quadrupedal Robot with 4-DoFs Legs for Roller Skating
Efficient Terrain Map Using Planar Regions for Footstep Planning on Humanoid Robots
Convergent iLQR for Safe Trajectory Planning and Control of Legged Robots
Optimization Based Dynamic Skateboarding of Quadrupedal Robot
Hierarchical Experience-Informed Navigation for Multi-Modal Quadrupedal Rebar Grid Traversal
Learning-Based Propulsion Control for Amphibious Quadruped Robots with Dynamic Adaptation t
Reinforcement Learning for Blind Stair Climbing with Legged and Wheeled-Legged Robots
Modeling and Analysis of Combined Rimless Wheel with Tensegrity Spine
A Vision-Centric Approach for Static Map Element Annotation
VBR: A Vision Benchmark in Rome
Spatial-Aware Dynamic Lightweight Self-Supervised Monocular Depth Estimation
VDNA-PR: Using General Dataset Representations for Robust Sequential Visual Place Recognitio
NISB-Map: Scalable Mapping with Neural Implicit Spatial Block
Regressing Transformers for Data-Efficient Visual Place Recognition
On the Study of Data Augmentation for Visual Place Recognition
Enhancing Visual Place Recognition with Multi-Modal Features and Time-Constrained Graph Atten
MBFusion: A New Multi-Modal BEV Feature Fusion Method for HD Map Construction
Quantized Visual-Inertial Odometry
OCC-VO: Dense Mapping Via 3D Occupancy-Based Visual Odometry for Autonomous Driving
NF-Atlas: Multi-Volume Neural Feature Fields for Large Scale LiDAR Mapping
Dusk Till Dawn: Self-Supervised Nighttime Stereo Depth Estimation Using Visual Foundation Mode
SiLVR: Scalable Lidar-Visual Reconstruction with Neural Radiance Fields for Robotic Inspection
LESS-Map: Lightweight and Evolving Semantic Map in Parking Lots for Long-Term Self-Localizati
Observation Time Difference: An Online Dynamic Objects Removal Method for Ground Vehicles
RO-MAP: Real-Time Multi-Object Mapping with Neural Radiance Fields
OctoMap-RT: Fast Probabilistic Volumetric Mapping Using Ray-Tracing GPUs
An Onboard Framework for Staircases Modeling Based on Point Clouds
V-STRONG: Visual Self-Supervised Traversability Learning for Off-Road Navigation
Follow the Footprints: Self-Supervised Traversability Estimation for Off-Road Vehicle Navigation
Learning to Predict Navigational Patterns from Partial Observations
TSCM: A Teacher-Student Model for Vision Place Recognition Using Cross-Metric Knowledge Distil
3D-BBS: Global Localization for 3D Point Cloud Scan Matching Using Branch-And-Bound Algorith
DynaInsRemover: A Real-Time Dynamic Instance-Aware Static 3D LiDAR Mapping Framework for
Learning Semantic-Agnostic and Spatial-Aware Representation for Generalizable Visual-Audio Nav
Efficient 3D Instance Mapping and Localization with Neural Fields
Salience-Guided Ground Factor for Robust Localization of Delivery Robots in Complex Urban Env
Block-Map-Based Localization in Large-Scale Environment
Subsurface Feature-Based Ground Robot/Vehicle Localization Using a Ground Penetrating Radar
Colmap-PCD: An Open-Source Tool for Fine Image-To-Point Cloud Registration
COIN-LIO: Complementary Intensity-Augmented LiDAR Inertial Odometry
MegaParticles: Range-Based 6-DoF Monte Carlo Localization with GPU-Accelerated Stein Particle
Tightly Coupled Range Inertial Localization on a 3D Prior Map Based on Sliding Window Factor Gr
SPOT: Point Cloud Based Stereo Visual Place Recognition for Similar and Opposing Viewpoints
TP3M: Transformer-Based Pseudo 3D Image Matching with Reference Image
Adaptive Outlier Thresholding for Bundle Adjustment in Visual SLAM
From Satellite to Ground: Satellite Assisted Visual Localization with Cross-View Semantic Matchin
Self-Supervised Learning of Monocular Visual Odometry and Depth with Uncertainty-Aware Scale
Unifying Local and Global Multimodal Features for Place Recognition in Aliased and Low-Texture
RELEAD: Resilient Localization with Enhanced LiDAR Odometry in Adverse Environments
Semantic-Focused Patch Tokenizer with Multi-Branch Mixer for Visual Place Recognition
FF-LINS: A Consistent Frame-To-Frame Solid-State-LiDAR-Inertial State Estimator
VioLA: Aligning Videos to 2D LiDAR Scans
WayIL: Image-Based Indoor Localization with Wayfinding Maps
TransAPR: Absolute Camera Pose Regression with Spatial and Temporal Attention
Globalizing Local Features: Image Retrieval Using Shared Local Features with Pose Estimation for
Leveraging Neural Radiance Fields for Uncertainty-Aware Visual Localization
JIST: Joint Image and Sequence Training for Sequential Visual Place Recognition
CCL: Continual Contrastive Learning for LiDAR Place Recognition
OptiState: State Estimation of Legged Robots Using Gated Networks with Transformer-Based Visio
Pose-Graph Attentional Graph Neural Network for Lidar Place Recognition
ColonMapper: Topological Mapping and Localization for Colonoscopy
Simultaneous Localization and Actuation Using Electromagnetic Navigation Systems
A Coarse-To-Fine Place Recognition Approach Using Attention-Guided Descriptors and Overlap Es
LHMap-Loc: Cross-Modal Monocular Localization Using LiDAR Point Cloud Heat Map
LocNDF: Neural Distance Field Mapping for Robot Localization
Looking beneath More: A Sequence-Based Localizing Ground Penetrating Radar Framework
Increasing SLAM Pose Accuracy by Ground-To-Satellite Image Registration
EffLoc: Lightweight Vision Transformer for Efficient 6-DOF Camera Relocalization
SAGE-ICP: Semantic Information-Assisted ICP
HR-APR: APR-Agnostic Framework with Uncertainty Estimation and Hierarchical Refinement for C
Implicit Learning of Scene Geometry from Poses for Global Localization
Visual Localization in Repetitive and Symmetric Indoor Parking Lots Using 3D Key Text Graph
VOLoc: Visual Place Recognition by Querying Compressed Lidar Map
VPRTempo: A Fast Temporally Encoded Spiking Neural Network for Visual Place Recognition
17-Point Algorithm Revisited: Toward a More Accurate Way
AnyLoc: Towards Universal Visual Place Recognition
Lightweight Ground Texture Localization
NeRF-VINS: A Real-Time Neural Radiance Field Map-Based Visual-Inertial Navigation System
Night-Rider: Nocturnal Vision-Aided Localization in Streetlight Maps Using Invariant Extended Kalm
GPS-VIO Fusion with Online Rotational Calibration
Fully Onboard Low-Power Localization with Semantic Sensor Fusion on a Nano-UAV Using Floor
The LuViRA Dataset: Synchronized Vision, Radio, and Audio Sensors for Indoor Localization
Dual-IMU State Estimation for Relative Localization of Two Mobile Agents
Accurate Prior-Centric Monocular Positioning with Offline LiDAR Fusion
A Nonlinear Estimator for Dead Reckoning of Aquatic Surface Vehicles Using an IMU and a Dopple
Range-Visual-Inertial Sensor Fusion for Micro Aerial Vehicle Localization and Navigation
An Equivariant Approach to Robust State Estimation for the ArduPilot Autopilot System
Robust Indoor Localization with Ranging-IMU Fusion
Multimodal Indoor Localization Using Crowdsourced Radio Maps
CLIP-Loc: Multi-Modal Landmark Association for Global Localization in Object-Based Maps
Multiple Update Particle Filter: Position Estimation by Combining GNSS Pseudorange and Carrier
Robust Lifelong Indoor LiDAR Localization Using the Area Graph
Multi-Camera Asynchronous Ball Localization and Trajectory Prediction with Factor Graphs and
Doppler-Only Single-Scan 3D Vehicle Odometry
Do We Need Scan-Matching in Radar Odometry?
Outram: One-Shot Global Localization Via Triangulated Scene Graph and Global Outlier Pruning
Fast and Consistent Covariance Recovery for Sliding-Window Optimization-Based VINS
Adaptive Robot-Human Handovers with Preference Learning
NaviFormer: A Data-Driven Robot Navigation Approach Via Sequence Modeling and Path Planning w
Real-Time Adaptive Safety-Critical Control with Gaussian Processes in High-Order Uncertain Mode
DeformNet: Latent Space Modeling and Dynamics Prediction for Deformable Object Manipulation
Actor-Critic Model Predictive Control
Tractable Joint Prediction and Planning Over Discrete Behavior Modes for Urban Driving
GenDOM: Generalizable One-Shot Deformable Object Manipulation with Parameter-Aware Policy
Sim-To-Real Learning for Humanoid Box Loco-Manipulation
Hamiltonian Dynamics Learning from Point Cloud Observations for Nonholonomic Mobile Robot Co
Deep Model Predictive Optimization
Pay Attention to How You Drive: Safe and Adaptive Model-Based Reinforcement Learning for Off-
SERL: A Software Suite for Sample-Efficient Robotic Reinforcement Learning
Improving Out-Of-Distribution Generalization of Learned Dynamics by Learning Pseudometrics and
Robotic Offline RL from Internet Videos Via Value-Function Learning
Learning Manipulation of Steep Granular Slopes for Fast Mini Rover Turning
Safe Reinforcement Learning with Dead-Ends Avoidance and Recovery
Toward Optimal Tabletop Rearrangement with Multiple Manipulation Primitives
ReorientDiff: Diffusion Model Based Reorientation for Object Manipulation
ForceSight: Text-Guided Mobile Manipulation with Visual-Force Goals
Unknown Object Retrieval in Confined Space through Reinforcement Learning with Tactile Explora
The Grasp Loop Signature: A Topological Representation for Manipulation Planning with Ropes a
Articulated Object Manipulation with Coarse-To-Fine Affordance for Mitigating the Effect of Point
Improved M4M: Faster and Richer Planning for Manipulation among Movable Objects in Cluttere
Preprocessing-Based Kinodynamic Motion Planning Framework for Intercepting Projectiles Using
Augmenting Lane Perception and Topology Understanding with Standard Definition Navigation Ma
3QFP: Efficient Neural Implicit Surface Reconstruction Using Tri-Quadtrees and Fourier Feature P
Towards Large-Scale Incremental Dense Mapping Using Robot-Centric Implicit Neural Representa
Camera Relocalization in Shadow-Free Neural Radiance Fields
QuadricsNet: Learning Concise Representation for Geometric Primitives in Point Clouds
ERASOR++: Height Coding Plus Egocentric Ratio Based Dynamic Object Removal for Static Poin
H2-Mapping: Real-Time Dense Mapping Using Hierarchical Hybrid Representation
Uncertainty-Aware 3D Object-Level Mapping with Deep Shape Priors
RoboHop: Segment-Based Topological Map Representation for Open-World Visual Navigation
Scene Action Maps: Behavioural Maps for Navigation without Metric Information
Continuous Occupancy Mapping in Dynamic Environments Using Particles
Building Volumetric Beliefs for Dynamic Environments Exploiting Map-Based Moving Object Segme
Fast and Robust Normal Estimation for Sparse LiDAR Scans
OmniColor: A Global Camera Pose Optimization Approach of LiDAR-360Camera Fusion for Coloriz
Gaussian Process Mapping of Uncertain Building Models with GMM As Prior
Occupancy Grid Mapping without Ray-Casting for High-Resolution LiDAR Sensors
RH-Map: Online Map Construction Framework of Dynamic Object Removal Based on 3D Region-W
Photometric LiDAR and RGB-D Bundle Adjustment
Towards Centimeter-Scale Underwater Mobile Robots: An Architecture for Capable µAUVs
Untethered Bimodal Robotic Fish with Tunable Bistability
Tendon-Driven Continuum Robot for Deep-Sea Application
WAVE: An Open-Source underWater Arm-Vehicle Emulator
Nezha-F: Design and Analysis of a Foldable and Self-Deployable HAUV
Snapp: An Agile Robotic Fish with 3-D Maneuverability for Open Water Swim
A Novel Omnidirectional Swimming Robot with Articulated-Compliant Legs
Marine Sediment Sampling with an Underwater Legged Robot
Terrain-Adaptive Locomotion Control for an Underwater Hexapod Robot: Sensing Leg-Terrain Inter
RUMP: Robust Underwater Motion Planning in Dynamic Environments of Fast Moving Obstacles
Metrically Scaled Monocular Depth Estimation through Sparse Priors for Underwater Robots
Model-Based Underwater 6D Pose Estimation from RGB
SONIC: Sonar Image Correspondence Using Pose Supervised Learning for Imaging Sonars
CVAE-SM: A Conditional Variational Autoencoder with Style Modulation for Efficient Uncertainty Qu
Beyond NeRF Underwater: Learning Neural Reflectance Fields for True Color Correction of Marin
CaveSeg: Deep Semantic Segmentation and Scene Parsing for Autonomous Underwater Cave Exp
Discovering Biological Hotspots with a Passively Listening AUV
A Du-Octree Based Cross-Attention Model for LiDAR Geometry Compression
An Augmented Catenary Model for Underwater Tethered Robots
A Hybrid Dynamical Model for Robotic Underwater Vehicles When Submerged or Surfaced: Approa
Surfing Algorithm: Agile and Safe Transition Strategy for Hybrid Aerial Underwater Vehicle in Wav
ReefGlider: A Highly Maneuverable Vectored Buoyancy Engine Based Underwater Robot
Robust Model Predictive Control with Control Barrier Functions for Autonomous Surface Vessels
Design and Optimization of a Multimode Amphibious Robot with Propeller-Leg
Underwater Dome-Port Camera Calibration: Modeling of Refraction and Offset through N-Sphere
Benchmarking Classical and Learning-Based Multibeam Point Cloud Registration
Angler: An Autonomy Framework for Intervention Tasks with Lightweight Underwater Vehicle Mani
A Model for Multi-Agent Autonomy That Uses Opinion Dynamics and Multi-Objective Behavior Opti
Convex Geometric Trajectory Tracking Using Lie Algebraic MPC for Autonomous Marine Vehicles
Mission Planning for Multiple Autonomous Underwater Vehicles with Constrained in Situ Rechargi
Decentralized Multi-Robot Navigation for Autonomous Surface Vehicles with Distributional Reinfo
Real-Time Planning under Uncertainty for AUVs Using Virtual Maps
Sea-U-Foil: A Hydrofoil Marine Vehicle with Multi-Modal Locomotion
Learning Which Side to Scan: Multi-View Informed Active Perception with Side Scan Sonar for A
Development of a Lightweight Underwater Manipulator for Delicate Structural Repair Operations
An Open-Source Solution for Fast and Accurate Underwater Mapping with a Low-Cost Mechanical
Boundary Factors for Seamless State Estimation between Autonomous Underwater Docking Phas
Vision-Based Water Clearance Determination in Maritime Environment
Adaptive Landmark Color for AUV Docking in Visually Dynamic Environments
Navigable Area Detection and Perception-Guided Model Predictive Control for Autonomous Navig
An Online Self-Calibrating Refractive Camera Model with Application to Underwater Odometry
Enhancing Visual Inertial SLAM with Magnetic Measurements
Acoustic-VINS: Tightly Coupled Acoustic-Visual-Inertial Navigation System for Autonomous Under
Underwater Volumetric Mapping Using Imaging Sonar and Free-Space Modeling Approach
Sequential Trajectory Optimization for Externally-Actuated Modular Manipulators with Joint Lockin
Rapid Resistography with Passive Overhead-Perching Mechanism in an Unmanned Aerial System f
Dual Quaternion Control of UAVs with Cable-Suspended Load
Design, Modeling and Control of a Top-Loading Fully-Actuated Cargo Transportation Multirotor
Aerial Interaction with Tactile Sensing
A Meter-Scale Ornithopter Capable of Jumping Take-Off
Autonomous Aerial Perching and Unperching Using Omnidirectional Tiltrotor and Switching Control
RotorTM: A Flexible Simulator for Aerial Transportation and Manipulation
Simulation and Experimental Validation of an Autonomous Perching and Takeoff Method for a Mult
Energy Consumption Modelling of Coaxial-Rotor in Vortex Ring State for Controllable High-Speed
Equilibria, Stability, and Sensitivity for the Aerial Suspended Beam Robotic System Subject to Par
Aerial Tensile Perching and Disentangling Mechanism for Long-Term Environmental Monitoring
Millimeter-Level Pick and Peg-In-Hole Task Achieved by Aerial Manipulator
Lumped Drag Model Identification and Real-Time External Force Detection for Rotary-Wing Micro A
Flight Validation of a Global Singularity-Free Aerodynamic Model for Flight Control of Tail Sitters
Empirical Study of Ground Proximity Effects for Small-Scale Electroaerodynamic Thrusters
The Weighted Markov-Dubins Problem
The Price of a Safe Flight: Risk Cost Based Path Planning
Hitchhiker: A Quadrotor Aggressively Perching on a Moving Inclined Surface Using Compliant Suc
Harnessing the Differential Flatness of Monocopter Dynamics for the Purpose of Trajectory Trackin
Passive Aligning Physical Interaction of Fully-Actuated Aerial Vehicles for Pushing Tasks
Modeling and Control of PADUAV: A Passively Articulated Dual UAVs Platform for Aerial Manipulat
Particle Filter with Stable Embedding for State Estimation of the Rigid Body Attitude System on the
End-To-End Reinforcement Learning for Time Optimal Quadcopter Flight
Quadrolltor: A Reconfigurable Quadrotor with Controlled Rolling and Turning
Co-Design Optimisation of Morphing Topology and Control of Winged Drones
Robust Control for Bidirectional Thrust Quadrotors under Instantaneously Drastic Disturbances
MPS: A New Method for Selecting the Stable Closed-Loop Equilibrium Attitude-Error Quaternion of
Realtime Brain-Inspired Adaptive Learning Control for Nonlinear Systems with Configuration Uncert
Safety-Conscious Pushing on Diverse Oriented Surfaces with Underactuated Aerial Vehicles
Geranos: A Novel Tilted-Rotors Aerial Robot for the Transportation of Poles
Robust and Energy-Efficient Control for Multi-Task Aerial Manipulation with Automatic Arm-Switchi
Optimal Collaborative Transportation for Under-Capacitated Vehicle Routing Problems Using Aeri
A Modular Aerial System Based on Homogeneous Quadrotors with Fault-Tolerant Control
Observer-Based Controller Design for Oscillation Damping of a Novel Suspended Underactuated Ae
MOAR Planner: Multi-Objective and Adaptive Risk-Aware Path Planning for Infrastructure Inspecti
Magnetic Gear-Based Actuator: A Framework of Design, Optimization, and Disturbance Observer-
Johnsen-Rahbek Capstan Clutch: A High Torque Electrostatic Clutch
Research on Bionic Foldable Wing for Flapping Wing Micro Air Vehicle
Magnetic Field-Driven Bristle-Bots
A Scalable Monolithic 3D Printable Variable Stiffness Mechanism
Modular Growing Mechanism with Multi-Axis Deformation
Design and Experimental Characterisation of a Novel Quasi-Direct Drive Actuator for Highly Dynam
Design and Evaluation of a Reconfigurable 7-DOF Upper Limb Rehabilitation Exoskeleton with Gr
Flexible Omnidirectional Driving Gear Mechanism with Adaptation Over Arbitrary Curvatures
ReC-Gripper: A Reconfigurable Combined Suction and Fingered Gripper for Various Logistics Pic
Development of the Assembling System for Structure Transformable Humanoid with Attach-Lock-
Design of a Deployable Continuum Robot Using Elastic Kirigami-Origami
Design and Control of a Transformable Multi-Mode Mobile Robot
HyperLeg: Biomechanics-Inspired High-DOF Leg and Toe Mechanism for Highly Dynamic Motions
Design of a Towing System by Multi Autonomous Sailboats
Non-Intrusive LiDAR Protection Module Emulating Bio-Inspired Wiping Motion for Outdoor Unman
Lightweight Human-Friendly Robotic Arm Based on Transparent Hydrostatic Transmissions
OSCaR: An Origami-Inspired Shape-Changing Robot for Ground Coverage Tasks
Lissajous Curve-Based Vibrational Orbit Control of a Flexible Vibrational Actuator with a Structural
Design and Modeling of a Nested Bi-Cavity-Based Soft Growing Robot for Grasping in Constraine
Dynamic Modeling of Wing-Assisted Inclined Running with a Morphing Multi-Modal Robot
Optimized Design and Fabrication of Skeletal Muscle Actuators for Bio-Syncretic Robots
Exoskeleton-Mediated Physical Human-Human Interaction for a Sit-To-Stand Rehabilitation Task
Intraoperatively Iterative Hough Transform Based In-Plane Hybrid Control of Arterial Robotic Ultra
Efficient Model Learning and Adaptive Tracking Control of Magnetic Micro-Robots for Non-Contact
Design and Implementation of a Robotized Hand-Held Dissector for Endoscopic Pulmonary Endar
Colibri5: Real-Time Monocular 5-DoF Trocar Pose Tracking for Robot-Assisted Vitreoretinal Surge
Hybrid Volitional Control of a Robotic Transtibial Prosthesis Using a Phase Variable Impedance Con
Uncertainty-Aware Contextual Visualization for Human Supervision of OCT-Guided Autonomous Rob
A Track-Based Colon Endoscopic Robot with Depth Perception Stereo Cameras for Haustral Fold D
Caveats on the First-Generation Da Vinci Research Kit: Latent Technical Constraints and Essential
A Passive Variable Impedance Control Strategy with Viscoelastic Parameters Estimation of Soft T
A Robotic System for Transanal Endoscopic Microsurgery: Design, Dexterity Optimization and Prot
Implicit Neural Representations for Breathing-Compensated Volume Reconstruction in Robotic Ult
Shadow-Based 3D Pose Estimation of Intraocular Instrument Using Only 2D Images
Skeleton Graph-Based Ultrasound-CT Non-Rigid Registration
Automated Image Acquisition of Parasternal Long-Axis View with Robotic Echocardiography
DopUS-Net: Quality-Aware Robotic Ultrasound Imaging Based on Doppler Signal
Robotic Craniomaxillofacial Osteotomy System Using Acoustic 3D Registration
Elliptical Torus-Based Six-Axis FBG Force Sensor with In-Situ Calibration for Condition Monitoring
Vision-And-Force-Based Compliance Control for a Posterior Segment Ophthalmic Surgical Robot
A Hybrid Admittance Control Algorithm for Automatic Robotic Cranium-Milling
Preliminary Study of Fingertip and Wrist Motion Based Haptic Controller for Robotically Assisted
Haptic-Assisted Collaborative Robot Framework for Improved Situational Awareness in Skull Base
Intelligent Disinfection Robot with High-Touch Surface Detection and Dynamic Pedestrian Avoidan
Toward a Framework Integrating Augmented Reality and Virtual Fixtures for Safer Robot-Assist
A Force-Driven and Vision-Driven Hybrid Control Method of Autonomous Laparoscope-Holding Ro
Inconstant Curvature Kinematics of Parallel Continuum Robot without Static Model
A Novel SEA-Based Haptic Interface for Robot-Assisted Vascular Interventional Surgery
A Miniature 1R1T Precision Manipulator with Remote Center of Motion for Minimally Invasive Surg
A Three-Dimensional Compliant Bowtie-Shaped Mechanical Amplifier to Magnify Coaxial Displace
A Magnetic Continuum Robot with In-Situ Magnetic Reprogramming Capability
Design and Control of a Magnetically-Actuated Anti-Interference Microrobot for Targeted Therapeut
Design and Visual Servoing Control of a Hybrid Dual-Segment Flexible Neurosurgical Robot for Intr
Uncertainty-Aware Shape Estimation of a Surgical Continuum Manipulator in Constrained Environ
On the Disentanglement of Tube Inequalities in Concentric Tube Continuum Robots
Design and Experimental Investigation of a Vibro-Impact Capsule Robot for Colonoscopy
3D Navigation of a Magnetic Swimmer Using a 2D Ultrasonography Probe Manipulated by a Roboti
An Intelligent Robotic Endoscope Control System Based on Fusing Natural Language Processing
AI-Assisted Dynamic Tissue Evaluation for Early Bowel Cancer Diagnosis Using a Vibrational Cap
AiAReSeg: Catheter Detection and Segmentation in Interventional Ultrasound Using Transformers
Hybrid Robot for Percutaneous Needle Intervention Procedures: Mechanism Design and Experiment
Envibroscope: Real-Time Monitoring and Prediction of Environmental Motion for Enhancing Safety
Cooperative vs. Teleoperation Control of the Steady Hand Eye Robot with Adaptive Sclera Force
Adaptive Motion Scaling for Robot-Assisted Microsurgery Based on Hybrid Offline Reinforcement
Chained Flexible Capsule Endoscope: Unraveling the Conundrum of Size Limitations and Functional I
Towards a Novel Soft Magnetic Laparoscope for Single Incision Laparoscopic Surgery
Magnetic-Guided Flexible Origami Robot Toward Long-Term Phototherapy of H. Pylori in the Sto
Flexible Tactile-Sensing Gripper Design and Excessive Force Protection Function for Endovascula
Simultaneous Estimation of Shape and Force Along Highly Deformable Surgical Manipulators U
Autonomous System for Tumor Resection (ASTR)-Dual-Arm Robotic Midline Partial Glossectomy
A Semi-Autonomous Data Driven Shared Control Framework for Robotic Manipulation and Cutting
Design and Evaluation of a Modular Robotic System for Microsurgery
Analyzing Accessibility in Robot-Assisted Vitreoretinal Surgery: Integrating Eye Posture and Robot
The Flux One Magnetic Navigation System: A Preliminary Assessment for Stent Implantation
Learning-Based Efficient Phase-Amplitude Modulation and Hybrid Control for MRI-Guided Focuse
Co-Axial Slender Tubular Robot (CAST): Towards Robotized Operation for Transorbital Neurosurge
Vascular Centerline-Guided Autonomous Navigation Methods for Robot-Lead Endovascular Interve
A Soft Micro-Robotic Catheter for Aneurysm Treatment: A New Design and Enhanced Euler-Bernoul
A Generic Modeling Framework for the Design of Tendon-Driven Continuum Manipulators with Flex
Bevel-Tip Needle Deflection Modeling, Simulation, and Validation in Multi-Layer Tissues
Excitation Trajectory Optimization for Dynamic Parameter Identification Using Virtual Constraints
Automated Assembly by Two-Fingered Microhand for Fabrication of Soft Magnetic Microrobots
Thin-Film NiTi Microactuator with a Magnetic Spring for a Tiny Launcher Mechanism
Nature-Inspired Bubble Magnetic Microrobots for Multimode Locomotion, Cargo Delivery, Imaging,
Magnetic Mobile Micro-Gripping MicroRobots (MMμGRs) with Two Independent Magnetic Actuat
Electroosmotic Self-Propelled Microswimmer with Magnetic Steering
A Flat Tendon-Driven Continuum Microrobot for Brain Interventions
A Selectively Controllable Triple-Helical Micromotor
A Robotic Surgery Platform for Automated Tissue Micromanipulation in Zebrafish Embryos
Skill Learning in Robot-Assisted Micro-Manipulation through Human Demonstrations with Attentio
Automated Surgical Knot Tying on Mini-Incision with Micro-Suture Based on Dual-Arm Nanorobot
Weakly-Supervised Depth Completion During Robotic Micromanipulation from a Monocular Micros
Dynamic Adaptive Imaging System on Optoelectronic Tweezers Platform
Automated Dissection of Intact Single Cell from Tissue Using Robotic Micromanipulation System
Development of a 3-RRS Micromanipulator Based on Origami-Inspired Spherical Joint
Singularity Analysis and Solutions for the Origami Transmission Mechanism of Fast-Moving Untet
A Theoretical Investigation of the Ability of Magnetic Miniature Robots to Exert Forces and Torques
Automated Non-Invasive Analysis of Motile Sperms Using Cross-Scale Guidance Network
Multi-Scale Visual Servoing Framework for Optical Microscopy Based on SIFT Matching
Robotic Capillary Insertion to the Xenopus Oocyte Using Microscopic Image Analysis and QCR Fo
Robotic Mosaic Atomic Force Microscopy through Sequential Imaging and Multiview Iterative Clos
Automated Sperm Immobilization with a Clinically-Compatible and Compact XYZ Stage
Automated Sperm Morphology Analysis Based on Instance-Aware Part Segmentation
Fast Photoacoustic Microscopy with Robot Controlled Microtrajectory Optimization
Acoustically Driven Micropipette for Hydrodynamic Manipulation of Mouse Oocytes
Remote Control of Untethered Magnetic Robots within a Lumen Using X-Ray-Guided Robotic Plat
Sim-Suction: Learning a Suction Grasp Policy for Cluttered Environments Using a Synthetic Benc
Robot Task Planning under Local Observability
Real-Time Whole-Body Motion Planning for Mobile Manipulators Using Environment-Adaptive Sear
Learning Hierarchical Interactive Multi-Object Search for Mobile Manipulation
Keep It Upright: Model Predictive Control for Nonprehensile Object Transportation with Obstacle
Gaussian Mixture Likelihood-Based Adaptive MPC for Interactive Mobile Manipulators
GAMMA: Graspability-Aware Mobile MAnipulation Policy Learning Based on Online Grasping Pose
Dynamic Interaction Control in Legged Mobile Manipulators: A Decoupled Approach
Active-Perceptive Motion Generation for Mobile Manipulation
Active Learning of Discrete-Time Dynamics for Uncertainty-Aware Model Predictive Control
SculptBot: Pre-Trained Models for 3D Deformable Object Manipulation
Learning-Based Model Predictive Control for an Autonomous Formula Student Racing Car
Learning Terrain-Aware Kinodynamic Model for Autonomous Off-Road Rally Driving with Model Pred
Adaptive Gait Modeling and Optimization for Principally Kinematic Systems
Recursive Least Squares with Log-Determinant Divergence Regularisation for Online Inertia Identif
Sequential Manipulation of Deformable Linear Object Networks with Endpoint Pose Measurements
Physics-Informed Neural Network for Multirotor Slung Load Systems Modeling
A Probabilistic Motion Model for Skid-Steer Wheeled Mobile Robot Navigation on Off-Road Terrain
Towards Unifying Human Likeness: Evaluating Metrics for Human-Like Motion Retargeting on Bim
MiBOT: A Head-Worn Robot That Modulates Cardiovascular Responses through Human-Like Sof
ESP: Extro-Spective Prediction for Long-Term Behavior Reasoning in Emergency Scenarios
SynthAct: Towards Generalizable Human Action Recognition Based on Synthetic Data
Visual-Tactile Robot Grasping Based on Human Skill Learning from Demonstrations Using a Weara
Workstation Suitability Maps: Generating Ergonomic Behaviors on a Population of Virtual Humans
Self Model for Embodied Intelligence: Modeling Full-Body Human Musculoskeletal System and Loc
Keypoints-Guided Lightweight Network for Single-View 3D Human Reconstruction
Moving Horizon Estimation of Human Kinematics and Muscle Forces
A Novel Model for Layer Jamming-Based Continuum Robots
Lumped Parameter Dynamic Model of an Eversion Growing Robot: Analysis, Simulation and Experi
Trajectory Tracking Control of Dual-PAM Soft Actuator with Hysteresis Compensator
Kinematic Modeling and Control of a Soft Robotic Arm with Non-Constant Curvature Deformation
CIDGIKc: Distance-Geometric Inverse Kinematics for Continuum Robots
Automatically-Tuned Model Predictive Control for an Underwater Soft Robot
Realtime Robust Shape Estimation of Deformable Linear Object
Learning-Based Tracking Control of Soft Robots
A Provably Stable Iterative Learning Controller for Continuum Soft Robots
Control and Implementation of a Fluidic Elastomer Actuator for Active Suppression of Hand Tremo
Adaptive State Estimation with Constant-Curvature Dynamics Using Force-Torque Sensors with App
SofToss: Learning to Throw Objects with a Soft Robot
A Soft Robot Inverse Kinematics for Virtual Reality
Toward the Use of Proxies for Efficient Learning Manipulation and Locomotion Strategies on Soft
Vision-Based Autonomous Steering of a Miniature Eversion Growing Robot
POE: Acoustic Soft Robotic Proprioception for Omnidirectional End-Effectors
A Data-Driven Approach to Geometric Modeling of Systems with Low-Bandwidth Actuator Dynami
Modeling and Control of Intrinsically Elasticity Coupled Soft-Rigid Robots
Design and Implementation of a Robotic Testbench for Analyzing Pincer Grip Execution in Huma
Human Gait Cost Function Varies with Walking Speed: An Inverse Optimal Control Study
LORIS: A Lightweight Free-Climbing Robot for Extreme Terrain Exploration
MST-Q: Micro Suction Tape Quadruped Robot with High Payload Capacity
Floating-Base Manipulation on Zero-Perturbation Manifolds
Towards Geometric Motion Planning for High-Dimensional Systems: Gait-Based Coordinate Optimi
Autonomous Navigation with Online Replanning and Recovery Behaviors for Wheeled-Legged Rob
Signal Temporal Logic Neural Predictive Control
Multi-Query TDSP for Path Planning in Time-Varying Flow Fields
CTopPRM: Clustering Topological PRM for Planning Multiple Distinct Paths in 3D Environments
Stein Variational Guided Model Predictive Path Integral Control: Proposal and Experiments with F
An Efficient Solution to the 2D Visibility Problem in Cartesian Grid Maps and Its Application in Heur
Efficient Clothoid Tree-Based Local Path Planning for Self-Driving Robots
Decentralized Lifelong Path Planning for Multiple Ackerman Car-Like Robots
Energy-Aware Ergodic Search: Continuous Exploration for Multi-Agent Systems with Battery Const
RBI-RRT*: Efficient Sampling-Based Path Planning for High-Dimensional State Space
Quasi-Static Path Planning for Continuum Robots by Sampling on Implicit Manifold
Reconfiguration of a 2D Structure Using Spatio-Temporal Planning and Load Transferring
Neural Informed RRT*: Learning-Based Path Planning with Point Cloud State Representations under
Motions in Microseconds Via Vectorized Sampling-Based Planning
Gathering Data from Risky Situations with Pareto-Optimal Trajectories
RETRO: Reactive Trajectory Optimization for Real-Time Robot Motion Planning in Dynamic Envir
WiTHy A*: Winding-Constrained Motion Planning for Tethered Robot Using Hybrid A*
Differentiable Boustrophedon Paths That Enable Optimization Via Gradient Descent
Constant-Time Motion Planning with Anytime Refinement for Manipulation
VAPOR: Legged Robot Navigation in Unstructured Outdoor Environments Using Offline Reinforce
EDMP: Ensemble-Of-Costs-Guided Diffusion for Motion Planning
Approximating Robot Configuration Spaces with Few Convex Sets Using Clique Covers of Visibilit
Asymptotically-Optimal Multi-Robot Visibility-Based Pursuit-Evasion
APP: A* Post-Processing Algorithm for Robots with Bidirectional Shortcut and Path Perturbation
Scaling Infeasibility Proofs Via Concurrent, Codimension-One, Locally-Updated Coxeter Triangulat
Skeleton Disk-Graph Roadmap: A Sparse Deterministic Roadmap for Safe 2D Navigation and Expl
SMUG Planner: A Safe Multi-Goal Planner for Mobile Robots in Challenging Environments
Light-Weight Approach for Safe Landing in Populated Areas
Time-Optimal Gate-Traversing Planner for Autonomous Drone Racing
Design and Evaluation of Motion Planners for Quadrotors in Environments with Varying Complexiti
AutoTrans: A Complete Planning and Control Framework for Autonomous UAV Payload Transporta
Bat Planner: Aggressive Flying Ball Player
An NMPC Framework for Tracking and Releasing a Cable-Suspended Load to a Ground Target Usi
Multi-Vehicle Dynamic Water Surface Monitoring
Aerial Physical Human Robot Interaction for Payload Transportation
On Experimental Emulation of Printability and Fleet Aware Generic Mesh Decomposition for Enablin
A Robotic Manipulator Using Dual-Motor Joints: Prototype Design and Anti-Backlash Control
CoNi-MPC: Cooperative Non-Inertial Frame Based Model Predictive Control
Uniform Passive Fault-Tolerant Control of a Quadcopter with One, Two, or Three Rotor Failure
Geometric Slosh-Free Tracking for Robotic Manipulators
Development of a Four-Wheel Steering Scale Vehicle for Research and Education on Autonomous
Online-Learning-Based Distributionally Robust Motion Control with Collision Avoidance for Mobile
Prediction of Pose Errors Implied by External Forces Applied on Robots: Towards a Metric for the
Iterative Learning Control for Deformable Open-Frame Cable-Driven Parallel Robots
A Study of Force-Free Control Framework for Industrial Manipulator Tasks Based on High-Pass Fil
Safety-Critical Coordination of Legged Robots Via Layered Controllers and Forward Reachable Set
Safety-Critical Control of Quadrupedal Robots with Rolling Arms for Autonomous Inspection of C
Robust and Remote Center of Cyclic Motion Control for Redundant Robots with Partially Unknown
Safe Risk-Averse Bayesian Optimization for Controller Tuning
Phase Synthesis for Spatial Locomotion Control of Retractable Worm Robots
Enhanced Robust Motion Control Based on Unknown System Dynamics Estimator for Robot Manip
Model-Free Control of a Class of High-Precision Scanning Motion Systems with Piezoceramic Actu
Constrained Nonlinear Disturbance Observer for Robotic Systems
An Integrated Position-Velocity-Force Method for Safety-Enhanced Shared Control in Robot-Assist
Hybrid Force-Impedance Control for Fast End-Effector Motions
A Reinforcement Learning-Based Control Strategy for Robust Interaction of Robotic Systems with
Stiffness-Based Hybrid Motion/ Force Control for Cable-Driven Serpentine Manipulator
Reinforcement Learning for Reduced-Order Models of Legged Robots
K-BMPC: Derivative-Based Koopman Bilinear Model Predictive Control for Tractor-Trailer Traject
Learning to Shape by Grinding: Cutting-Surface-Aware Model-Based Reinforcement Learning
Adaptive Contact-Implicit Model Predictive Control with Online Residual Learning
DRIVE: Data-Driven Robot Input Vector Exploration
Redundancy Resolution at Position Level
Solving Rearrangement Puzzles Using Path Defragmentation in Factored State Spaces
Leveraging Opportunism in Sample-Based Motion Planning
SE(2) Assembly Planning for Magnetic Modular Cubes
A Constrained Path Following Method for Snake-Like Manipulators Via Controlled Winding Uncoilin
Multi-Profile Quadratic Programming (MPQP) for Optimal Gap Selection and Speed Planning of A
IKLink: End-Effector Trajectory Tracking with Minimal Reconfigurations
A Control Barrier Function-Based Motion Planning Scheme for a Quadruped Robot
Physics-Informed Neural Motion Planning on Constraint Manifolds
Practical and Safe Navigation Function Based Motion Planning of UAVs
Characterizing Physical Adversarial Attacks on Robot Motion Planners
Eclares: Energy-Aware Clarity-Driven Ergodic Search
Tight Motion Planning by Riemannian Optimization for Sliding and Rolling with Finite Number of Co
Online Trajectory Deformation and Tracking for Self-Entanglement-Free Differential-Driven Robots
Efficient Motion Planning for Manipulators with Control Barrier Function-Induced Neural Controller
Planning Optimal Trajectories for Mobile Manipulators under End-Effector Trajectory Continuity Con
Zero-Shot Constrained Motion Planning Transformers Using Learned Sampling Dictionaries
LSTP: Long Short-Term Motion Planning for Legged and Legged-Wheeled Systems
Risk-Inspired Aerial Active Exploration for Enhancing Autonomous Driving of UGV in Unknown Of
The Un-Kidnappable Robot: Acoustic Localization of Sneaking People
Predicting the Intention to Interact with a Service Robot: The Role of Gaze Cues
Non-Verbal Cues on Robot-Group Persuasion
Naturalistic Robot-To-Human Bimanual Handover in Complex Environments through Multi-Sensor
Language and Sketching: An LLM-Driven Interactive Multimodal Multitask Robot Navigation Fram
Dual-Modal Tactile E-Skin: Enabling Bidirectional Human-Robot Interaction Via Integrated Tactile
Close-Range Human Following Control on a Cane-Type Robot with Multi-Camera Fusion
CAMInterHand: Cooperative Attention for Multi-View Interactive Hand Pose and Mesh Reconstruct
Action Segmentation Using 2D Skeleton Heatmaps and Multi-Modality Fusion
Vision and Tactile-Based Continuous Multimodal Intention and Attention Recognition for Safer Phy
Towards Unified Interactive Visual Grounding in the Wild
Think, Act and Ask: Open-World Interactive Personalized Robot Navigation
PROGrasp: Pragmatic Human-Robot Communication for Object Grasping
Enhancing Tactile Sensing in Robotics: Dual-Modal Force and Shape Perception with EIT-Base
CLARA: Classifying and Disambiguating User Commands for Reliable Interactive Robotic Agents
Assisting Group Discussions Using Desktop Robot Haru
Assessment and Benchmarking of XoNLI: A Natural Language Processing Interface for Industrial
Advancing Virtual Reality Interaction: A Ring-Shaped Controller and Pose Tracking
Tight Fusion of Odometry and Kinematic Constraints for Multiple Aerial Vehicles in Physical Interc
Robust Multi-Robot Global Localization with Unknown Initial Pose Based on Neighbor Constraints
Swarm-SLAM: Sparse Decentralized Collaborative Simultaneous Localization and Mapping Framew
AutoFusion: Autonomous Visual Geolocation and Online Dense Reconstruction for UAV Cluster
CoLRIO: LiDAR-Ranging-Inertial Centralized State Estimation for Robotic Swarms
Relative Localization Estimation for Multiple Robots Via the Rotating Ultra-Wideband Tag
Asynchronous Multiple LiDAR-Inertial Odometry Using Point-Wise Inter-LiDAR Uncertainty Propag
AutoMerge: A Framework for Map Assembling and Smoothing in City-Scale Environments
Observer-Based Distributed MPC for Collaborative Quadrotor-Quadruped Manipulation of a Cabl
Learning for Dynamic Subteaming and Voluntary Waiting in Heterogeneous Multi-Robot Collaborat
Uncertainty-Bounded Active Monitoring of Unknown Dynamic Targets in Road-Networks with Mini
Do We Run Large-Scale Multi-Robot Systems on the Edge? More Evidence for Two-Phase Perform
Asynchronous Distributed Smoothing and Mapping Via On-Manifold Consensus ADMM
Phasic Diversity Optimization for Population-Based Reinforcement Learning
VO-Safe Reinforcement Learning for Drone Navigation
RoCo: Dialectic Multi-Robot Collaboration with Large Language Models
Collision Avoidance and Navigation for a Quadrotor Swarm Using End-To-End Deep Reinforcemen
C3F: Constant Collaboration and Communication Framework for Graph-Representation Dynamic M
Multi-Level Action Tree Rollout (MLAT-R): Efficient and Accurate Online Multiagent Policy Improv
Stimulate the Potential of Robots Via Competition
Multi-Agent Visual Coordination Using Optical Wireless Communication
Distributed Control of a Limited Angular Field-Of-View Multi-Robot System in Communication-Deni
Assessing Reputation to Improve Team Performance in Heterogeneous Multi-Robot Coverage
A Robot Web for Distributed Many-Device Localisation
Learning Decentralized Flocking Controllers with Spatio-Temporal Graph Neural Network
Simultaneous Time Synchronization and Mutual Localization for Multi-Robot System
Enabling Large-Scale Heterogeneous Collaboration with Opportunistic Communications
AG-CVG: Coverage Planning with a Mobile Recharging UGV and an Energy-Constrained UAV
A Non-Cubic Space-Filling Modular Robot
Optimal Containment Control of Multiple Quadrotors Via Reinforcement Learning
Behavior Tree Capabilities for Dynamic Multi-Robot Task Allocation with Heterogeneous Robot T
Multi-Robot Cooperative Navigation in Crowds: A Game-Theoretic Learning-Based Model Predicti
Multi-Robot Human-In-The-Loop Control under Spatiotemporal Specifications
Hypergraph-Based Multi-Robot Task and Motion Planning
Measurement-Limited Multi-Agent, Relative Pose Estimation for On-Orbit Inspection
Dynamic Targeting of Satellite Observations Incorporating Slewing Costs and Complex Observation
RecNet: An Invertible Point Cloud Encoding through Range Image Embeddings for Multi-Robot Ma
Automatic Configuration of Multi-Agent Model Predictive Controllers Based on Semantic Graph Wo
Meta-Reinforcement Learning Based Cooperative Surface Inspection of 3D Uncertain Structures U
Decentralized Multi-Agent Trajectory Planning in Dynamic Environments with Spatiotemporal Occ
Communicating Intent As Behaviour Trees for Decentralised Multi-Robot Coordination
Partial Belief Space Planning for Scaling Stochastic Dynamic Games
Decentralized Multi-Agent Active Search and Tracking When Targets Outnumber Agents
Multi-Robot Autonomous Exploration and Mapping under Localization Uncertainty with Expectation
Optimal Task Allocation for Heterogeneous Multi-Robot Teams with Battery Constraints
Bigraph Matching Weighted with Learnt Incentive Function for Multi-Robot Task Allocation
Automation and Artificial Intelligence Technology in Surface Mining: State of the Art, Challenges a
Hierarchical Traffic Management of Multi-AGV Systems with Deadlock Prevention Applied to Indust
Task Allocation in Heterogeneous Multi-Robot Systems Based on Preference-Driven Hedonic Ga
Persistent Monitoring of Multiple Moving Targets Using High Order Control Barrier Functions
Singularity Analysis of Rigid Directed Bearing Graphs for Quadrotor Formations
EM-Patroller: Entropy Maximized Multi-Robot Patrolling with Steady State Distribution Approximati
Behavioral-Based Circular Formation Control for Robot Swarms
Optimization and Evaluation of a Multi Robot Surface Inspection Task through Particle Swarm Opti
Hierarchical Planning for Long-Horizon Multi-Agent Collective Construction
Fast Force-Closure Grasp Synthesis with Learning-Based Sampling
The New Dexterity Modular, Dexterous, Anthropomorphic, Open-Source, Bimanual Manipulation Pla
Fully 3D Printable Robot Hand and Soft Tactile Sensor Based on Air-Pressure and Capacitive Prox
TPGP: Temporal-Parametric Optimization with Deep Grasp Prior for Dexterous Motion Planning
A Wearable Robotic Hand for Hand-Over-Hand Imitation Learning
WARABI Hand: Five-Fingered Robotic Hand with Flexible Skin and Force Sensors for Social Intera
Sensorized Soft Skin for Dexterous Robotic Hands
Identifying Expert Behavior in Offline Training Datasets Improves Behavioral Cloning of Robotic Ma
Development of a Versatile Robotic Hand Toward Jig-Less Assembly of a Shaft-Shaped Part
Coordinated Landing Control for Cross-Domain UAV-USV Fleets Using Heterogeneous-Feature Ma
Distributed Control Barrier Functions for Global Connectivity Maintenance
Stability Analysis of Distance-Angle Leader-Follower Formation Control
A Distributed Multi-Robot Framework for Exploration, Information Acquisition and Consensus
Environmental Awareness Dynamic 5G QoS for Retaining Real Time Constraints in Robotic Applic
CloudGripper: An Open Source Cloud Robotics Testbed for Robotic Manipulation Research, Bench
FogROS2-Config: A Toolkit for Choosing Server Configurations for Cloud Robotics
Opportunistic Communication in Robot Teams
Enhancing the Tracking Performance of Passivity-Based High-Frequency Robot Cloud Control
IFFNeRF: Initialisation Free and Fast 6DoF Pose Estimation from a Single Image and a NeRF Mod
VeloVox: A Low-Cost and Accurate 4D Object Detector with Single-Frame Point Cloud of Livox Li
MTRadSSD: A Multi-Task Single-Stage Detector for Object Detection and Free Space Analysis in
Toward Accurate Camera-Based 3D Object Detection Via Cascade Depth Estimation and Calibrati
DA-RAW: Domain Adaptive Object Detection for Real-World Adverse Weather Conditions
SyMFM6D: Symmetry-Aware Multi-Directional Fusion for Multi-View 6D Object Pose Estimation
Mutual Information-Calibrated Conformal Feature Fusion for Uncertainty-Aware Multimodal 3D Obje
RGB-Based Category-Level Object Pose Estimation Via Decoupled Metric Scale Recovery
Implicit Coarse-To-Fine 3D Perception for Category-Level Object Pose Estimation from Monocul
Road Obstacle Detection Based on Unknown Objectness Scores
PVTransformer: Point-To-Voxel Transformer for Scalable 3D ObjectDetection
Object-Centric Cross-Modal Feature Distillation for Event-Based Object Detection
Hierarchical Point Attention for Indoor 3D Object Detection
Frame Fusion with Vehicle Motion Prediction for 3D Object Detection
FG-PFE: Fine-Grained Pillar Feature Encoding Via Spatio-Temporal Virtual Grid for 3D Object Det
Efficient Semantic Segmentation for Compressed Video
Cross-Cluster Shifting for Efficient and Effective 3D Object Detection in Autonomous Driving
BEE-Net: Bridging Semantic and Instance with Gated Encoding and Edge Constraint for Efficient
Robust 3D Object Detection from LiDAR-Radar Point Clouds Via Cross-Modal Feature Augmentati
Predicting Class Distribution Shift for Reliable Domain Adaptive Object Detection
LSSAttn: Towards Dense and Accurate View Transformation for Multi-Modal 3D Object Detection
Learning Temporal Cues by Predicting Objects Move for Multi-Camera 3D Object Detection
Improved Yolov5: HIC-Yolov5 for Small Object Detection
CLIPUNetr: Assisting Human-Robot Interface for Uncalibrated Visual Servoing Control with CLIP-
C2FDrone: Coarse-To-Fine Drone-To-Drone Detection Using Vision Transformer Networks
Better Monocular 3D Detectors with LiDAR from the Past
A Metacognitive Approach to Out-Of-Distribution Detection for Segmentation
RISeg: Robot Interactive Object Segmentation Via Body Frame-Invariant Features
PUMA: Fully Decentralized Uncertainty-Aware Multiagent Trajectory Planner with Real-Time Ima
Open-Vocabulary Affordance Detection Using Knowledge Distillation and Text-Point Correlation
MEDL-U: Uncertainty-Aware 3D Automatic Annotation Based on Evidential Deep Learning
HandyPriors: Physically Consistent Perception of Hand-Object Interactions with Differentiable Prior
Dynamic Occupancy Grids for Object Detection: A Radar-Centric Approach
Dynamic Object Classification of Low-Resolution Point Clouds: An LSTM-Based Ensemble Learni
Dynablox: Real-Time Detection of Diverse Dynamic Objects in Complex Environments
3D Object Detection with VI-SLAM Point Clouds: The Impact of Object and Environment Character
Uplifting Range-View-Based 3D Semantic Segmentation in Real-Time with Multi-Sensor Fusion
Radar Tracker: Moving Instance Tracking in Sparse and Noisy Radar Point Clouds
ProEqBEV: Product Group Equivariant BEV Network for 3D Object Detection in Road Scenes of A
Orientation-Aware Multi-Modal Learning for Road Intersection Identification and Mapping
Masked Gamma-SSL: Learning Uncertainty Estimation Via Masked Image Modeling
EfficientDPS: Efficient and End-To-End Depth-Aware Panoptic Segmentation
Robust Collaborative Perception against Temporal Information Disturbance
Concavity-Induced Distance for Unoriented Point Cloud Decomposition
FocoTrack: Multi Object Tracking by Focusing on Overlap at Low Frame Rate
WLST: Weak Labels Guided Self-Training for Weakly-Supervised Domain Adaptation on 3D Object
Towards a Robust Sensor Fusion Step for 3D Object Detection on Corrupted Data
TerrainSense: Vision-Driven Mapless Navigation for Unstructured Off-Road Environments
RCM-Fusion: Radar-Camera Multi-Level Fusion for 3D Object Detection
One-Vs-All Semi-Automatic Labeling Tool for Semantic Segmentation in Autonomous Driving
LiRaFusion: Deep Adaptive LiDAR-Radar Fusion for 3D Object Detection
Discwise Active Learning for LiDAR Semantic Segmentation
BaSAL: Size Balanced Warm Start Active Learning for LiDAR Semantic Segmentation
Trajectory-Prediction-Based Dynamic Tracking of a UGV to a Moving Target under Multi-Disturbed
Motion Planning for 4WS Vehicle with Autonomous Selection of Steering Modes Via an MIQP-MPC
On the Performance of Jerk-Constrained Time-Optimal Trajectory Planning for Industrial Manipulat
Symmetric Stair Preconditioning of Linear Systems for Parallel Trajectory Optimization
MPCGPU: Real-Time Nonlinear Model Predictive Control through Preconditioned Conjugate Gradi
Efficient and Robust Time-Optimal Trajectory Planning and Control for Agile Quadrotor Flight
Invariant Descriptors of Motion and Force Trajectories for Interpreting Object Manipulation Tasks i
Generalizing Trajectory Retiming to Quadratic Objective Functions
A Distributed Processing Approach for Smooth Task Transitioning in Strict Hierarchical Control
Risk-Averse Trajectory Optimization Via Sample Average Approximation
Optimal Control for Clutched-Elastic Robots: A Contact-Implicit Approach
Optimal Control for Articulated Soft Robots
Force Feedback Model-Predictive Control Via Online Estimation
Geometric Algebra for Optimal Control with Applications in Manipulation Tasks
Trajectory Tracking Runtime Assurance for Systems with Partially Unknown Dynamics
How to Train Your Neural Control Barrier Function: Learning Safety Filters for Complex Input-Con
Stable, Safe, and Passive Teleoperation of Multi-Robot Systems
Approximate Optimal Controller Synthesis for Cart-Poles and Quadrotors Via Sums-Of-Squares
Online Multi-Contact Feedback Model Predictive Control for Interactive Robotic Tasks
Optimal Control of Granular Material
CACTO: Continuous Actor-Critic with Trajectory Optimization---Towards Global Optimality
Learning Model Predictive Control with Error Dynamics Regression for Autonomous Racing
Safe Non-Stochastic Control of Control-Affine Systems: An Online Convex Optimization Approach
Robust Balancing Control of Biped Robots for External Forces
Robust Policy Iteration of Uncertain Interconnected Systems with Imperfect Data
Robust Co-Design of Canonical Underactuated Systems for Increased Certifiable Stability
Whole-Body Ergodic Exploration with a Manipulator Using Diffusion
ReLU-QP: A GPU-Accelerated Quadratic Programming Solver for Model-Predictive Control
Collision Detection and Avoidance for Black Box Multi-Robot Navigation
Stackelberg Game-Theoretic Trajectory Guidance for Multi-Robot Systems with Koopman Operato
Optimal Path Planning for a Convoy-Support Vehicle Pair through a Repairable Network
Learning Heterogeneous Multi-Agent Allocations for Ergodic Search
Multi-Robot Cooperative Socially-Aware Navigation Using Multi-Agent Reinforcement Learning
Multi-Agent Path Finding for Cooperative Autonomous Driving
Communication-Aware Map Compression for Online Path-Planning
Wind Field Modeling for Formation Planning in Multi-Drone Systems
Multi-Robot Informative Path Planning from Regression with Sparse Gaussian Processes
Benchmarking Multi-Robot Coordination in Realistic, Unstructured Human-Shared Environments
Conflict Area Prediction for Boosting Search-Based Multi-Agent Pathfinding Algorithms
AMSwarmX: Safe Swarm Coordination in CompleX Environments Via Implicit Non-Convex Decompo
Conflict-Based Model Predictive Control for Scalable Multi-Robot Motion Planning
Db-CBS: Discontinuity-Bounded Conflict-Based Search for Multi-Robot Kinodynamic Motion Plann
ALPHA Attention-Based Long-Horizon Pathfinding in Highly-Structured Areas
Online On-Demand Multi-Robot Coverage Path Planning
Scalable Multi-Robot Motion Planning for Congested Environments with Topological Guidance
Mixed Integer Programming for Time-Optimal Multi-Robot Coverage Path Planning with Efficient He
N-MPC for Deep Neural Network-Based Collision Avoidance Exploiting Depth Images
Aerial Transportation of Cable-Suspended Loads with an Event Camera
Autonomous Overhead Powerline Recharging for Uninterrupted Drone Operations
LPNet: A Reaction-Based Local Planner for Autonomous Collision Avoidance Using Imitation Lear
Autonomous Exploration of Unknown 3D Environments Using a Frontier-Based Collector Strategy
Learning Multi-Scale Context Mask-RCNN Network for Slant Angled Aerial Imagery in Instance S
Real-Time Multi-Modal Active Vision for Object Detection on UAVs Equipped with Limited Field o
High-Speed Stereo Visual SLAM for Low-Powered Computing Devices
Air Bumper: A Collision Detection and Reaction Framework for Autonomous MAV Navigation
Safety-Aware Perception for Autonomous Collision Avoidance in Dynamic Environments
Incremental Multimodal Surface Mapping Via Self-Organizing Gaussian Mixture Models
Learning to Explore Indoor Environments Using Autonomous Micro Aerial Vehicles
FC-Planner: A Skeleton-Guided Planning Framework for Fast Aerial Coverage of Complex 3D Sce
Multi-Robot Multi-Room Exploration with Geometric Cue Extraction and Circular Decomposition
Fast Multi-UAV Decentralized Exploration of Forests
Reinforcement Learning for Collision-Free Flight Exploiting Deep Collision Encoding
Learning Agile Flights through Narrow Gaps with Varying Angles Using Onboard Sensing
Control-Barrier-Aided Teleoperation with Visual-Inertial SLAM for Safe MAV Navigation in Comple
Flow Shadowing: A Method to Detect Multiple Flow Headings Using an Array of Densely Packed W
Onboard Dynamic-Object Detection and Tracking for Autonomous Robot Navigation with RGB-D
APACE: Agile and Perception-Aware Trajectory Generation for Quadrotor Flights
SpECULARIA: Towards Fully Autonomous Robotic Indoor Farming System
End-To-End Thermal Updraft Detection and Estimation for Autonomous Soaring Using Temporal C
SANet: Small but Accurate Detector for Aerial Flying Object
N-QR: Natural Quick Response Codes for Multi-Robot Instance Correspondence
You Only Scan Once: A Dynamic Scene Reconstruction Pipeline for 6-DoF Robotic Grasping of No
Thermoformed Electronic Skins for Conformal Tactile Sensor Arrays
The GEM-C Controller for Load Compensation in Object Manipulation
Sim-To-Real Grasp Detection with Global-To-Local RGB-D Adaptation
Residual-NeRF : Learning Residual NeRFs for Transparent Object Manipulation
Online 3D Edge Reconstruction of Wiry Structures from Monocular Image Sequences
Kitchen Artist: Precise Control of Liquid Dispensing for Gourmet Plating
HandNeRF: Learning to Reconstruct Hand-Object Interaction Scene from a Single RGB Image
Robotic Grasping of Harvested Tomato Trusses Using Vision and Online Learning
ViHOPE: Visuotactile In-Hand Object 6D Pose Estimation with Shape Completion
VERGNet: Visual Enhancement Guided Robotic Grasp Detection under Low-Light Condition
TactileAR: Active Tactile Pattern Reconstruction
Online Estimation of Articulated Objects with Factor Graphs Using Vision and Proprioceptive Sens
Learning to Estimate Incipient Slip with Tactile Sensing to Gently Grasp Objects
Learning Interaction Constraints for Robot Manipulation Via Set Correspondences
Joint-Loss Enhanced Self-Supervised Learning for Refinement-Coupled Object 6D Pose Estimatio
Force-Based Semantic Representation and Estimation of Feature Points for Robotic Cable Manipul
AnyGrasp: Robust and Efficient Grasp Perception in Spatial and Temporal Domains
AHPPEBot: Autonomous Robot for Tomato Harvesting Based on Phenotyping and Pose Estimatio
Unknown Object Grasping for Assistive Robotics
Liquids Identification and Manipulation Via Digitally Fabricated Impedance Sensors
Learning to Grasp in Clutter with Interactive Visual Failure Prediction
Kinesthetic-Based In-Hand Object Recognition with an Underactuated Robotic Hand
Fit-NGP: Fitting Object Models to Neural Graphics Primitives
Efficient Object Rearrangement Via Multi-View Fusion
EDOPT: Event-Camera 6-DoF Dynamic Object Pose Tracking
Attention-Based Cloth Manipulation from Model-Free Topological Representation
A Group Theoretic Metric for Robot State Estimation Leveraging Chebyshev Interpolation
AD4RL: Autonomous Driving Benchmarks for Offline Reinforcement Learning with Value-Based Da
The Cluttered Environment Picking Benchmark (CEPB) for Advanced Warehouse Automation
SceneReplica: Benchmarking Real-World Robot Manipulation by Creating Replicable Scenes
CRITERIA: A New Benchmarking Paradigm for Evaluating Trajectory Prediction Models for Autono
LET-3D-AP: Longitudinal Error Tolerant 3D Average Precision for Camera-Only 3D Detection
HuNavSim: A ROS 2 Human Navigation Simulator for Benchmarking Human-Aware Robot Navigat
RobotPerf: An Open-Source, Vendor-Agnostic, Benchmarking Suite for Evaluating Robotics Comp
Standardization of Cloth Objects and Its Relevance in Robotic Manipulation
Unified Power and Admittance Adaptation for Safe and Effective Physical Interaction with Unmod
Towards Robot to Human Skill Coaching: A ML-Powered IoT and HRI Platform for Martial Arts Trai
Towards Robo-Coach: Robot Interactive Stiffness/Position Adaptation for Human Strength and Cond
Predictive and Robust Robot Assistance for Sequential Manipulation
Nonlinear Subsystem-Based Adaptive Impedance Control of Physical Human-Robot-Environment In
Model Predictive Control with Graph Dynamics for Garment Opening Insertion During Robot-Assis
Force Feedback-Based Gamification: Performance Validation of Squat Exergame Using Pneumatic
External Dynamics Dependent Human Gait Adaptation Using a Cable-Driven Exoskeleton
Comparison of Rating Scale and Pairwise Comparison Methods for Measuring Human Co-Worker Su
Transformer-Based Prediction of Human Motions and Contact Forces for Physical Human-Robot In
SynH2R: Synthesizing Hand-Object Motions for Learning Human-To-Robot Handovers
Proactive Robot Control for Collaborative Manipulation Using Human Intent
Human Modeling in Physical Human-Robot Interaction: A Brief Survey
Exploring Transformers and Visual Transformers for Force Prediction in Human-Robot Collaborati
Exploring the Effect of Base Compliance on Physical Human-Robot Collaboration
Experimental and Simulation-Based Estimation of Interface Power During Physical Human-Robot I
A Personalizable Controller for the Walking Assistive omNi-Directional Exo-Robot (WANDER)
MUI-TARE: Cooperative Multi-Agent Exploration with Unknown Initial Position
Safe Planning in Dynamic Environments Using Conformal Prediction
Wasserstein Distributionally Robust Chance Constrained Trajectory Optimization for Mobile Robots
Shield Model Predictive Path Integral: A Computationally Efficient Robust MPC Method Using Contr
Distributionally Robust CVaR-Based Safety Filtering for Motion Planning in Uncertain Environment
Monte Carlo Planning in Hybrid Belief POMDPs
Data Association Aware POMDP Planning with Hypothesis Pruning Performance Guarantees
Safe POMDP Online Planning Via Shielding
Generating Sparse Probabilistic Graphs for Efficient Planning in Uncertain Environments
Stochastic Implicit Neural Signed Distance Functions for Safe Motion Planning under Sensing Unce
Constrained Hierarchical Monte Carlo Belief-State Planning
Estimating 3D Uncertainty Field: Quantifying Uncertainty for Neural Radiance Fields
Online Adaptation of Sampling-Based Motion Planning with Inaccurate Models
Autonomous 3D Exploration in Large-Scale Environments with Dynamic Obstacles
MTG: Mapless Trajectory Generator with Traversability Coverage for Outdoor Navigation
IBBT: Informed Batch Belief Trees for Motion Planning under Uncertainty
Integrating Predictive Motion Uncertainties with Distributionally Robust Risk-Aware Control for Sa
A GP-Based Robust Motion Planning Framework for Agile Autonomous Robot Navigation and Rec
Multi-Sample Long Range Path Planning under Sensing Uncertainty for Off-Road Autonomous Driv
Perceptual Factors for Environmental Modeling in Robotic Active Perception
Weathering Ongoing Uncertainty: Learning and Planning in a Time-Varying Partially Observable E
Choosing the Right Tool for the Job: Online Decision Making Over SLAM Algorithms
ASPIRe: An Informative Trajectory Planner with Mutual Information Approximation for Target Sear
Preprocessing-Based Planning for Utilizing Contacts in Semi-Structured High-Precision Insertion
Vision-Based Uncertainty-Aware Motion Planning Based on Probabilistic Semantic Segmentation
Chance-Constrained Multi-Robot Motion Planning under Gaussian Uncertainties
Uncertainty-Aware Trajectory Planning: Using Uncertainty Quantification and Propagation in Travers
Quadcopter Trajectory Time Minimization and Robust Collision Avoidance Via Optimal Time Alloca
Scaling Team Coordination on Graphs with Reinforcement Learning
Dynamic Crane Scheduling with Reinforcement Learning for a Steel Coil Warehouse
Tree-Based Representation of Locally Shortest Paths for 2D K-Shortest Non-Homotopic Path Plan
Well-Connected Set and Its Application to Multi-Robot Path Planning
Dynamic Coalition Formation and Routing for Multirobot Task Allocation Via Reinforcement Learni
Multi-Robot Task Allocation under Uncertainty Via Hindsight Optimization
Traffic Flow Learning Enhanced Large-Scale Multi-Robot Cooperative Path Planning under Uncerta
Accounting for Travel Time and Arrival Time Coordination During Task Allocations in Legged-Rob
A Transtibial Prosthesis Using a Parallel Spring Mechanism
Deep Learning Based Acoustic Measurement Approach for Robotic Applications on Orthopedics
EMG-Based Intention Detection Using Deep Learning for Shared Control in Upper-Limb Assistive
Gait Phase Detection Based on LSTM-CRF for Stair Ambulation
Towards a Unified Approach for Continuously-Variable Impedance Control of Powered Prosthetic
Design, Simulation and Kinematic Verification of a Multi-Loop Ankle-Foot Prosthetic Mechanism
Short Term After-Effects of Small Force Fields Applied by an Upper-Limb Exoskeleton on Inter-Joi
Real-Time Dexterous Prosthesis Hand Control by Decoding Neural Information Based on EMG De
A Novel Back-Support Exoskeleton with a Differential Series Elastic Actuator for Lifting Assistance
Lightweight and Flexible Prosthetic Wrist with Shape Memory Alloy (SMA)-Based Artificial Muscle and
Ankle Exoskeleton with a Symmetric 3 DoF Structure for Plantarflexion Assistance
Design of a Front-Enveloping Powered Exoskeleton Considering Optimal Distribution of Actuating
Real-Time Locomotion Transitions Detection: Maximizing Performances with Minimal Resources
ExoRecovery: Push Recovery with a Lower-Limb Exoskeleton Based on Stepping Strategy
Pilot Comparison of Customized and Generalized Hip-Knee-Ankle Exoskeleton Torque Profiles
Task-Space Control of a Powered Ankle Prosthesis
Integrating Computer Vision in Exosuits for Adaptive Support and Reduced Muscle Strain in Indust
Geometry-Informed Distance Candidate Selection for Adaptive Lightweight Omnidirectional Stereo
On Camera Model Conversions
RadCloud: Real-Time High-Resolution Point Cloud Generation Using Low-Cost Radars for Aerial
That's My Point: Compact Object-Centric LiDAR Pose Estimation for Large-Scale Outdoor Localisa
A Point-To-Distribution Degeneracy Detection Factor for LiDAR SLAM Using Local Geometric Mod
I-Octree: A Fast, Lightweight, and Dynamic Octree for Proximity Search
Self-Supervised Depth Correction of Lidar Measurements from Map Consistency Loss
GroundGrid: LiDAR Point Cloud Ground Segmentation and Terrain Estimation
Morphable-SfS: Enhancing Shape-From-Silhouette Via Morphable Modeling
Waverider: Leveraging Hierarchical, Multi-Resolution Maps for Efficient and Reactive Obstacle Av
Whisker-Based Tactile Navigation Algorithm for Underground Robots
Spline-Interpolated Model Predictive Path Integral Control with Stein Variational Inference for Reac
RETOM: Leveraging Maneuverability for Reactive Tool Manipulation Using Wrench-Fields
Optimal Prescribed-Time Control Based Reactive Planning System for Quadruped Robot Navigati
On the Fly Robotic-Assisted Medical Instrument Planning and Execution Using Mixed Reality
K-VIL: Keypoints-Based Visual Imitation Learning
Circular Field Motion Planning for Highly-Dynamic Multi-Robot Systems with Application to Robot
Autonomous Mapless Navigation on Uneven Terrains
The Impact of Evolutionary Computation on Robotic Design: A Case Study with an Underactuated
Design & Systematic Evaluation of Power Transmission Efficiency of an Ankle Exoskeleton for Wal
Achieving Mechanical Transparency Using Fusion Hybrid Linear Actuator for Shoulder Flexion and
Imitation Learning-Based System for the Execution of Self-Paced Robotic-Assisted Passive Rehabil
An Adaptable Ankle Trajectory Generation Method for Lower-Limb Exoskeletons by Means of Safe
Controlling FES of Arm Movements Using Physics-Informed Reinforcement Learning Via Co-Krigin
Adaptive Control for Triadic Human-Robot-FES Collaboration in Gait Rehabilitation: A Pilot Study
Stretch with Stretch: Physical Therapy Exercise Games Led by a Mobile Manipulator
Offline Goal-Conditioned Reinforcement Learning for Safety-Critical Tasks with Recovery Policy
Reinforcement Learning in a Safety-Embedded MDP with Trajectory Optimization
Distributional Reinforcement Learning with Sample-Set Bellman Update
Learning Adaptive Safety for Multi-Agent Systems
Contrastive Initial State Buffer for Reinforcement Learning
Safety Optimized Reinforcement Learning Via Multi-Objective Policy Optimization
Differentially Encoded Observation Spaces for Perceptive Reinforcement Learning
Projected Task-Specific Layers for Multi-Task Reinforcement Learning
Bi²Lane: Bi-Directional Temporal Refinement with Bi-Level Feature Aggregation for 3D Lane Dete
Active Automotive Augmented Reality Displays Using Reinforcement Learning
Extremum-Seeking Action Selection for Accelerating Policy Optimization
Privacy Risks in Reinforcement Learning for Household Robots
MAexp: A Generic Platform for RL-Based Multi-Agent Exploration
Improving Offline Reinforcement Learning with Inaccurate Simulators
REFORMA: Robust REinFORceMent Learning Via Adaptive Adversary for Drones Flying under Di
Brain-Inspired Hyperdimensional Computing in the Wild: Lightweight Symbolic Learning for Senso
Learning for Deformable Linear Object Insertion Leveraging Flexibility Estimation from Visual Cues
Reinforcement Learning of Action and Query Policies with LTL Instructions under Uncertain Event
MOSAIC: Learning Unified Multi-Sensory Object Property Representations for Robot Learning Via I
Neural Rearrangement Planning for Object Retrieval from Confined Spaces Perceivable by Rob
MMPI: A Flexible Radiance Field Representation by Multiple Multi-Plane Images Blending
Neural Implicit Swept Volume Models for Fast Collision Detection
Reg-NF: Efficient Registration of Implicit Surfaces within Neural Fields
3D-OAE: Occlusion Auto-Encoders for Self-Supervised Learning on Point Clouds
Composing Pre-Trained Object-Centric Representations for Robotics from "What" and "Where" Fo
SKT-Hang: Hanging Everyday Objects Via Object-Agnostic Semantic Keypoint Trajectory Generati
CITR: A Coordinate-Invariant Task Representation for Robotic Manipulation
SlotGNN: Unsupervised Discovery of Multi-Object Representations and Visual Dynamics
What Do We Learn from a Large-Scale Study of Pre-Trained Visual Representations in Sim and R
L-DYNO: Framework to Learn Consistent Visual Features Using Robot’s Motion
Point Cloud Models Improve Visual Robustness in Robotic Learners
HIO-SDF: Hierarchical Incremental Online Signed Distance Fields
Recasting Generic Pretrained Vision Transformers As Object-Centric Scene Encoders for Manipulat
NeRF-Loc: Transformer-Based Object Localization within Neural Radiance Fields
Multi-Resolution Planar Region Extraction for Uneven Terrains
RIC: Rotate-Inpaint-Complete for Generalizable Scene Reconstruction
Exploiting Point-Wise Attention in 6D Object Pose Estimation Based on Bidirectional Prediction
Stereo-LiDAR Depth Estimation with Deformable Propagation and Learned Disparity-Depth Conver
Leveraging Cycle-Consistent Anchor Points for Self-Supervised RGB-D Registration
MMAUD: A Comprehensive Multi-Modal Anti-UAV Dataset for Modern Miniature Drone Threats
SupeRGB-D: Zero-Shot Instance Segmentation in Cluttered Indoor Environments
Mean Shift Mask Transformer for Unseen Object Instance Segmentation
SLCF-Net: Sequential LiDAR-Camera Fusion for Semantic Scene Completion Using a 3D Recurren
WeatherDepth: Curriculum Contrastive Learning for Self-Supervised Depth Estimation under Adve
Collaborative Decision-Making Using Spatiotemporal Graphs in Connected Autonomy
Multi-Model 3D Registration: Finding Multiple Moving Objects in Cluttered Point Clouds
A Robust Deformable Linear Object Perception Pipeline in 3D: From Segmentation to Reconstruct
Unlocking the Performance of Proximity Sensors by Utilizing Transient Histograms
VG4D: Vision-Language Model Goes 4D Video Recognition
ConceptGraphs: Open-Vocabulary 3D Scene Graphs for Perception and Planning
TrackDLO: Tracking Deformable Linear Objects under Occlusion with Motion Coherence
Long-Tailed 3D Semantic Segmentation with Adaptive Weight Constraint and Sampling
Torsion-Induced Compliant Joints and Its Application to Flat-Foldable and Self-Assembling Roboti
OriTrack: A Small, 3 Degree of Freedom, Origami Solar Tracker
Reinforcement Learning for Freeform Robot Design
A Helical Bistable Soft Gripper Enable by Pneumatic Actuation
Singularity Analysis of Kinova's Link 6 Robot~Arm Via Grassmann Line Geometry
Design and Testing of a Multi-Module, Tetherless, Soft Robotic Eel
Untethered Underwater Soft Robot with Thrust Vectoring
A Backdrivable Axisymmetric Kinematically Redundant (6+3)-Degree-Of-Freedom Hybrid Parallel
Design of a Fully Pulley-Guided Wire-Driven Prismatic Tensegrity Robot: Friction Impact to Robot
Open X-Embodiment: Robotic Learning Datasets and RT-X Models
SARA-RT: Scaling up Robotics Transformers with Self-Adaptive Robust Attention
DenseTact-Mini: An Optical Tactile Sensor for Grasping Multi-Scale Objects from Flat Surfaces
Hearing Touch: Audio-Visual Pretraining for Contact-Rich Manipulation
Towards Generalizable Zero-Shot Manipulation Via Translating Human Interaction Plans
Constrained Bimanual Planning with Analytic Inverse Kinematics
Fault Tolerant Neural Control Barrier Functions for Robotic Systems under Sensor Faults and Atta
Belief Control Barrier Functions for Risk-Aware Control
A Novel Algorithmic Approach to Obtaining Maneuverable Control-Invariant Sets
Safe Navigation and Obstacle Avoidance Using Differentiable Optimization Based Control Barrier
Achieving Autonomous Cloth Manipulation with Optimal Control Via Differentiable Physics-Aware R
Online Data-Driven Safety Certification for Systems Subject to Unknown Disturbances
Towards Standardized Disturbance Rejection Testing of Legged Robot Locomotion with Linear Impa
Detecting and Mitigating System-Level Anomalies of Vision-Based Controllers
Generative Modeling of Residuals for Real-Time Risk-Sensitive Safety with Discrete-Time Control
Learning-Based Inverse Perception Contracts and Applications
Safe Multi-Robot Exploration Using Symbolic Control
Receding-Constraint Model Predictive Control Using a Learned Approximate Control-Invariant Set
VBOC: Learning the Viability Boundary of a Robot Manipulator Using Optimal Control
Closing the Perception-Action Loop for Semantically Safe Navigation in Semi-Static Environments
Magnetorheological-Actuators: An Enabling Technology for Fast, Safe and Practical Collaborative
Risk-Aware Control for Robots with Non-Gaussian Belief Spaces
Conformal Decision Theory: Safe Autonomous Decisions from Imperfect Predictions
A Learning-Based Framework for Safe Human-Robot Collaboration with Multiple Backup Control Ba
HEGN: Hierarchical Equivariant Graph Neural Network for 9DoF Point Cloud Registration
Deep Evidential Uncertainty Estimation for Semantic Segmentation under Out-Of-Distribution Obst
Ultrafast Square-Root Filter-Based VINS
SeqTrack3D: Exploring Sequence Information for Robust 3D Point Cloud Tracking
NGEL-SLAM: Neural Implicit Representation-Based Global Consistent Low-Latency SLAM System
Universal Visual Decomposer: Long-Horizon Manipulation Made Easy
Dynamic Evaluation of a Suction Based Gripper for Fruit Picking Using a Physical Twin
Strawberry Weight Estimation Based on Plane-Constrained Binary Division Point Cloud Completio
AV4GAInsp: An Efficient Dual-Camera System for Identifying Defective Kernels of Cereal Grains
Aerial Image-Based Inter-Day Registration for Precision Agriculture
Inexpensive, Automated Pruning Weight Estimation in Vineyards
High Precision Leaf Instance Segmentation for Phenotyping in Point Clouds Obtained under Real F
Towards Robotic Tree Manipulation: Leveraging Graph Representations
Field Evaluation of a Prioritized Path-Planning Algorithm for Heterogeneous Agricultural Tasks of
Unsupervised Pre-Training for 3D Leaf Instance Segmentation
Robotic Assessment of a Crop's Need for Watering
Enhancement on Target-Gripper Alignment: A Tomato Harvesting Robot with Dual-Camera Image-
Few-Shot Fruit Segmentation Via Transfer Learning
Spatio-Temporal Correspondence Estimation of Growing Plants by Hausdorff Distance Based Skele
Field Robot for High-Throughput and High-Resolution 3D Plant Phenotyping
A Hybrid Controller Enhancing Transient Performance for an Aerial Manipulator Extracting a Wedg
VACNA: Visibility-Aware Cooperative Navigation with Application in Inventory Management
A CBF-Adaptive Control Architecture for Visual Navigation for UAV in the Presence of Uncertaintie
Osiris: Building Hierarchical Representations for Agricultural Environments
Streamlined Acquisition of Large Sensor Data for Autonomous Mobile Robots to Enable Efficient C
Development of an Automatic Sweet Pepper Harvesting Robot and Experimental Evaluation
LiDAR-Based Robot Transplanter
EdgeSoil 2.0 – Soil Analyzer Using Convolutional Neural Network and Camera Imaging for Agricu
Semiautonomous Precision Pruning of Upright Fruiting Offshoot Orchard Systems: An Integrated
On-The-Go Tree Detection and Geometric Traits Estimation with Ground Mobile Robots in Fruit T
Autonomous Apple Fruitlet Sizing with Next Best View Planning
Gradient-Based Local Next-Best-View Planning for Improved Perception of Targeted Plant Nodes
Containerized Vertical Farming Using Cobots
INoD: Injected Noise Discriminator for Self-Supervised Representation Learning in Agricultural Fiel
Unsupervised Generation of Labeled Training Images for Crop-Weed Segmentation in New Fields a
Log Loading Automation for Timber-Harvesting Industry
Region-Determined Localization Method for Unmanned Ground Vehicle under Pole-Like Feature E
Tree Instance Segmentation and Traits Estimation for Forestry Environments Exploiting LiDAR Dat
Automated Testing of Spatially-Dependent Environmental Hypotheses through Active Transfer Lea
Decentralized Multi-Phase Formation Control for Cattle Herding
Hyblock: Hardware Realization and Control of Modular Hydraulic Robots with Dowel Connectors
PLASTR: Planning for Autonomous Sampling-Based Trowelling
Self-Reconfigurable Robots for Collaborative Discrete Lattice Assembly
LiSTA: Geometric Object-Based Change Detection in Cluttered Environments
Scalable Underwater Assembly with Reconfigurable Visual Fiducials
Automatic Loading of Unknown Material with a Wheel Loader Using Reinforcement Learning
Learning Adaptive Policies for Autonomous Excavation under Various Soil Conditions by Adversar
Robotic Inspection and Subsurface Defect Mapping Using Impact-Echo and Ground Penetrating R
ERRA: An Embodied Representation and Reasoning Architecture for Long-Horizon Language-Cond
Grasp-Anything: Large-Scale Grasp Dataset from Foundation Models
Anticipate & Act : Integrating LLMs and Classical Planning for Efficient Task Execution in Househ
Conditionally Combining Robot Skills Using Large Language Models
Interactive Planning Using Large Language Models for Partially Observable Robotic Tasks
Optimal Scene Graph Planning with Large Language Model Guidance
CAPE: Corrective Actions from Precondition Errors Using Large Language Models
GraspGPT: Leveraging Semantic Knowledge from a Large Language Model for Task-Oriented Gra
DOS: A Deployment Operating System for RoboOps
Non-Singular Fast Terminal Adaptive Visual Tracking Control with Reduced Tuning Parameters for
Quadrotor Neural Network Adaptive Control: Design and Experimental Validation
Parameter Identifying Disturbance Rejection Control with Asymptotic Error Convergence
Smooth Computation without Input Delay: Robust Tube-Based Model Predictive Control for Robot
Nullspace Adaptive Model-Based Trajectory-Tracking Control for a 6-DOF Underwater Vehicle wit
Adaptive Planning and Control with Time-Varying Tire Models for Autonomous Racing Using Extr
Risk-Sensitive Extended Kalman Filter
Global Terminal Sliding Mode Control of Tethered Satellites Formation with Chattering Reduction
Robust Feedback Quadratic Programming for Kinematic-Controlled Robots
Safe-By-Design Digital Twins for Human-Robot Interaction: A Use Case for Humanoid Service Rob
Safe Execution of Learned Orientation Skills with Conic Control Barrier Functions
Real-Time Batched Distance Computation for Time-Optimal Safe Path Tracking
Overcoming Hand and Arm Occlusion in Human-To-Robot Handovers: Predicting Safe Poses with
Legible and Proactive Robot Planning for Prosocial Human-Robot Interactions
Integrated Data-Driven Inference and Planning-Based Human Motion Prediction for Safe Human-Ro
How Does Perception Affect Safety: New Metrics and Strategy
Constrained Passive Interaction Control: Leveraging Passivity and Safety for Robot Manipulators
Boosting Adversarial Training in Safety-Critical Systems through Boundary Data Selection
Autonomous Robotic Re-Alignment for Face-To-Face Underwater Human-Robot Interaction
Mechanism Design for New Sensors Field Deployment by LineRanger Powerline Robot
Translational Disturbance Rejection for Jet-Actuated Flying Continuum Robots on Mobile Bases
Cellular-Enabled Collaborative Robots Planning and Operations for Search-And-Rescue Scenarios
STAGE: Scalable and Traversability-Aware Graph Based Exploration Planner for Dynamically Var
Computation-Aware Multi-Object Search in 3D Space Using Submodular Tree
Multi-Robot Search in a 3D Environment with Intersection System Constraints
Wireless Communication Infrastructure Building for Mobile Robot Search and Inspection Missions
RB5 Low-Cost Explorer: Implementing Autonomous Long-Term Exploration on Low-Cost Robotic
Perception through Cognitive Emulation: “A Second Iteration of NaivPhys4RP for Learningless
Mapping High-Level Semantic Regions in Indoor Environments without Object Recognition
LLM-Grounder: Open-Vocabulary 3D Visual Grounding with Large Language Model As an Agent
Learning Off-Road Terrain Traversability with Self-Supervisions Only
Improving Radial Imbalances with Hybrid Voxelization and RadialMix for LiDAR 3D Semantic Segm
Few-Shot Panoptic Segmentation with Foundation Models
End-To-End Semantic Segmentation Network for Low-Light Scenes
DefFusion: Deformable Multimodal Representation Fusion for 3D Semantic Segmentation
Lifelong LERF: Local 3D Semantic Inventory Monitoring Using FogROS2
Translating Universal Scene Descriptions into Knowledge Graphs for Robotic Environment
SeMLaPS: Real-Time Semantic Mapping with Latent Prior Networks and Quasi-Planar Segmentati
Robotic Exploration through Semantic Topometric Mapping
Open-Fusion: Real-Time Open-Vocabulary 3D Mapping and Queryable Scene Representation
Mask4Former: Mask Transformer for 4D Panoptic Segmentation
Mask4D: End-To-End Mask-Based 4D Panoptic Segmentation for LiDAR Sequences
HSPNav: Hierarchical Scene Prior Learning for Visual Semantic Navigation towards Real Settings
Belief Scene Graphs: Expanding Partial Scenes with Objects through Computation of Expectation
A Guided Gaussian-Dirichlet Random Field for Scientist-In-The-Loop Inference in Underwater Robo
SG-RoadSeg: End-To-End Collision-Free Space Detection Sharing Encoder Representations Joint
Robust Few-Shot 3D Point Cloud Scene Segmentation
Radar Instance Transformer: Reliable Moving Instance Segmentation in Sparse Radar Point Cloud
On the Overconfidence Problem in Semantic 3D Mapping
Complementing Onboard Sensors with Satellite Maps: A New Perspective for HD Map Constructio
Complementary Random Masking for RGB-Thermal Semantic Segmentation
Collaborative Dynamic 3D Scene Graphs for Automated Driving
BroadBEV: Collaborative LiDAR-Camera Fusion for Broad-Sighted Bird's Eye View Map Construct
AGRNav: Efficient and Energy-Saving Autonomous Navigation for Air-Ground Robots in Occlusio
Enhancing mmWave Radar Point Cloud Via Visual-Inertial Supervision
Influence of Camera-LiDAR Configuration on 3D Object Detection for Autonomous Driving
Chasing Day and Night: Towards Robust and Efficient All-Day Object Detection Guided by an Eve
CMDFusion: Bidirectional Fusion Network with Cross-Modality Knowledge Distillation for LIDAR
SK-Net: Spectral-Based Knowledge Distillation in Low-Light Thermal Imagery for Robotic Percepti
Use Your Imagination: A Detector-Independent Approach for LiDAR Quality Booster
SuperFusion: Multilevel LiDAR-Camera Fusion for Long-Range HD Map Generation
Continuous-Time Ultra-Wideband-Inertial Fusion
GICI-LIB: A GNSS/INS/Camera Integrated Navigation Library
JSTR: Joint Spatio-Temporal Reasoning for Event-Based Moving Object Detection
AYDIV: Adaptable Yielding 3D Object Detection Via Integrated Contextual Vision Transformer
RadarCam-Depth: Radar-Camera Fusion for Depth Estimation with Learned Metric Scale
HabitatDyn 2.0: Dataset for Spatial Anticipation and Dynamic Object Localization
Attentive Multimodal Fusion for Optical and Scene Flow
LiDAR-Camera Calibration Using Intensity Variance Cost
SRFNet: Monocular Depth Estimation with Fine-Grained Structure Via Spatial Reliability-Oriented
Bayesian Filtering for Homography Estimation
Saturation-Aware Angular Velocity Estimation: Extending the Robustness of SLAM to Aggressive
Multi-Modal 3D Human Tracking for Robots in Complex Environment with Siamese Point-Video Tr
Efficient Gesture Recognition on Spiking Convolutional Networks through Sensor Fusion of Event
Smoothly Connected Preemptive Impact Reduction and Contact Impedance Control
Point Cloud-Based Control Barrier Function Regression for Safe and Efficient Vision-Based Contro
Stability Analysis of Plane-To-Plane Positioning by Proximity-Based Control
An Image Acquisition Scheme for Visual Odometry Based on Image Bracketing and Online Attribut
MagicTip: A Novel High-Resolution 3D Multi-Layer Grid-Based Tactile Sensor
Microphone Pair Training for Robust Sound Source Localization with Diverse Array Configurations
Mobile Bot Rotation Using Sound Source Localization and Distant Speech Recognition
CalliRewrite: Recovering Handwriting Behaviors from Calligraphy Images without Supervision
Efficient and Accurate Transformer-Based 3D Shape Completion and Reconstruction of Fruits for A
Censible: A Robust and Practical Global Localization Framework for Planetary Surface Missions
Learning to Walk in Confined Spaces Using 3D Representation
Self-Recovery Prompting: Promptable General Purpose Service Robot System with Foundation Mo
Autonomous Quilt Spreading for Caregiving Robots
CNS: Correspondence Encoded Neural Image Servo Policy
Adapting for Calibration Disturbances: A Neural Uncalibrated Visual Servoing Policy
ARIS 1.0: An Autonomous Multitasking Medical Service Robot for Hospital Environments
Design, Modeling and Analysis of a Spherical Parallel Continuum Manipulator for Nursing Robots
LeagTag: An Elongated High-Accuracy Fiducial Marker for Tight Spaces
An Open and Flexible Robot Perception Framework for Mobile Manipulation Tasks
Toward Mass Customization of a Robot's Morphology Design for Improving Area Coverage
CoPAL: Corrective Planning of Robot Actions with Large Language Models
A Convex Formulation of Frictional Contact between Rigid and Deformable Bodies
SocialGAIL: Faithful Crowd Simulation for Social Robot Navigation
MuRoSim – a Fast and Efficient Multi-Robot Simulation for Learning-Based Navigation
STARK: A Unified Framework for Strongly Coupled Simulation of Rigid and Deformable Bodies with
Hydrodynamic Interactions in Schooling Fish: Prioritizing Real Fish Kinematics Over Travelling-Wa
OmniLRS: A Photorealistic Simulator for Lunar Robotics
SceneControl: Diffusion for Controllable Traffic Scene Generation
Jade: A Differentiable Physics Engine for Articulated Rigid Bodies with Intersection-Free Frictional
Simulation Modeling of Highly Dynamic Omnidirectional Mobile Robots Based on Real-World Data
KDD-LOAM: Jointly Learned Keypoint Detector and Descriptors Assisted LiDAR Odometry and Ma
Campus Map: A Large-Scale Dataset to Support Multi-View VO, SLAM and BEV Estimation
DISO: Direct Imaging Sonar Odometry
CURL-MAP: Continuous Mapping and Positioning with CURL Representation
Degradation Resilient LiDAR-Radar-Inertial Odometry
Design and Evaluation of a Generic Visual SLAM Framework for Multi Camera Systems
Ground-Fusion: A Low-Cost Ground SLAM System Robust to Corner Cases
HERO-SLAM: Hybrid Enhanced Robust Optimization of Neural SLAM
ONeK-SLAM: A Robust Object-Level Dense SLAM Based on Joint Neural Radiance Fields and Key
A Two-Step Nonlinear Factor Sparsification for Scalable Long-Term SLAM Backend
Effectively Detecting Loop Closures Using Point Cloud Density Maps
LOG-LIO: A LiDAR-Inertial Odometry with Efficient Local Geometric Information Estimation
Radar-Only Odometry and Mapping for Autonomous Vehicles
IPC: Incremental Probabilistic Consensus-Based Consistent Set Maximization for SLAM Backends
Generalized Correspondence Matching Via Flexible Hierarchical Refinement and Patch Descriptor Di
VOOM: Robust Visual Object Odometry and Mapping Using Hierarchical Landmarks
Lite-SVO: Towards a Lightweight Self-Supervised Semantic Visual Odometry Exploiting Multi-Featu
Efficient Pose Prediction with Rational Regression Applied to VSLAM
IMU-Aided Event-Based Stereo Visual Odometry
S-Graphs+: Real-Time Localization and Mapping Leveraging Hierarchical Representations
Visual Place Recognition: A Tutorial
Multi-Radar Inertial Odometry for 3D State Estimation Using mmWave Imaging Radar
Semantically Guided Feature Matching for Visual SLAM
DVI-SLAM: A Dual Visual Inertial SLAM Network
DMSA - Dense Multi Scan Adjustment for LiDAR Inertial Odometry and Global Optimization
CTA-LO: Accurate and Robust LiDAR Odometry Using Continuous-Time Adaptive Estimation
LONER: LiDAR Only Neural Representations for Real-Time SLAM
LIO-EKF: High Frequency LiDAR-Inertial Odometry Using Extended Kalman Filters
Multi-LIO: A Lightweight Multiple LiDAR-Inertial Odometry System
The Importance of Coordinate Frames in Dynamic SLAM
VoxelMap++: Mergeable Voxel Mapping Method for Online LiDAR(-Inertial) Odometry
Efficient and Consistent Bundle Adjustment on Lidar Point Clouds
DORF: A Dynamic Object Removal Framework for Robust Static LiDAR Mapping in Urban Enviro
ImMesh: An Immediate LiDAR Localization and Meshing Framework
OASIS: Optimal Arrangements for Sensing in SLAM
HPF-SLAM: An Efficient Visual SLAM System Leveraging Hybrid Point Features
2D-3D Object Shape Alignment for Camera-Object Pose Compensation in Object-Visual SLAM
Spectral Trade-Off for Measurement Sparsification of Pose-Graph SLAM
Learning Covariances for Estimation with Constrained Bilevel Optimization
UWB Radar SLAM: An Anchorless Approach in Vision Denied Indoor Environments
Less Is More: Physical-Enhanced Radar-Inertial Odometry
Linear Four-Point LiDAR SLAM for Manhattan World Environments
IBoW3D: Place Recognition Based on Incremental and General Bag of Words in 3D Scans
Language-EXtended Indoor SLAM (LEXIS): A Versatile System for Real-Time Visual Scene Under
VICAN: Very Efficient Calibration Algorithm for Large Camera Networks
Tightly-Coupled LiDAR-Visual-Inertial SLAM and Large-Scale Volumetric-Occupancy Mapping
Active Visual Localization for Multi-Agent Collaboration: A Data-Driven Approach
Autonomous Implicit Indoor Scene Reconstruction with Frontier Exploration
Probabilistic Active Loop Closure for Autonomous Exploration
CARE: Confidence-Rich Autonomous Robot Exploration Using Bayesian Kernel Inference and Opti
Event-Based Stereo Visual Odometry with Native Temporal Resolution Via Continuous-Time Gaus
MSCEqF: A Multi State Constraint Equivariant Filter for Vision-Aided Inertial Navigation
L-VIWO: Visual-Inertial-Wheel Odometry Based on Lane Lines
Training a Non-Cooperator to Identify Vulnerabilities and Improve Robustness for Robot Navigatio
Toward Grounded Commonsense Reasoning
The Effect of Rejection Strategy on Trust and Shopping Choices in Robot-Assisted Shopping
Planning of Explanations for Robot Navigation
Learning Crowd Behaviors in Navigation with Attention-Based Spatial-Temporal Graphs
Grounding Conversational Robots on Vision through Dense Captioning and Large Language Mode
Exploring the Impact of Narrator Type on Response Latency and Utterance Length During Interactiv
Design of Embodied Mediator Haru for Remote Cross Cultural Communication
ChatAdp: ChatGPT-Powered Adaptation System for Human-Robot Interaction
Rotenna: Harnessing Seamless Integration of Continuum Robot for Dynamic Electromagnetic Reco
Efficient RRT*-Based Safety-Constrained Motion Planning for Continuum Robots in Dynamic Envi
Ultrafast Capturing In-Flight Objects with Reprogrammable Working Speed Ranges
A Restorable, Variable Stiffness Pneumatic Soft Gripper Based on Jamming of Strings of Beads
Hard Shell, Soft Core: Binary Actuators for Deep-Sea Applications
Tip-Clutching Winch for High Tensile Force Application with Soft Growing Robots
Symmetry-Aware Reinforcement Learning for Robotic Assembly under Partial Observability with a S
Force Estimation at the Bionic Soft Arm’s Tool-Center-Point During the Interaction with the En
Field-Evaluated Closed Structure Soft Gripper Enhances the Shelf Life of Harvested Blackberries
A Soft, Lightweight Flipping Robot with Versatile Motion Capabilities for Wall-Climbing Applications
Tetraflex: A Multigait Soft Robot for Object Transportation in Confined Environments
A Strong Underwater Soft Manipulator with Planarly-Bundled Actuators and Accurate Position Cont
Learning-Based Object Recognition Via a Eutectogel Electronic Skin Enabled Soft Robotic Gripper
Design and Validation of Slender Extensible Continuum Robot for Solar Wing Re-Unfolding in Aer
Bio-Inspired Pupal-Mode Actuator with Ultra-Crossing Capability for Soft Robots
Crawling Soft Robot Exploiting Wheel-Legs and Multimodal Locomotion for High Terrestrial Maneuv
Compliant Robotic Gripper with Integrated Ripeness Sensing for Blackberry Harvesting
Lightweight Untethered Soft Robotic Fish
Optimal Design of Flexible-Link Mechanisms with Desired Load-Displacement Profiles
A Scalable, Light-Controlled, Individually Addressable, Non-Metal Actuator Array
A Passively Bendable, Compliant Tactile Palm with RObotic Modular Endoskeleton Optical (ROME
Shape-Conformable Suction Cups with Controllable Adaptive Suction on Complex Surfaces
A Phase-Change Emulsion Jamming Gripper for Manipulation of Micro-Scale Textured Surfaces
Design and Fabrication of String-Driven Origami Robots
Design and Implementation of a Ferrofluid-Based Liquid Robot for Small-Scale Manipulation
Towards Optimal Design of Dielectric Elastomer Actuators Using a Graph Neural Network Encoder
Modeling and Design of Lattice-Reinforced Pneumatic Soft Robots
Design and Analysis of Soft Hybrid-Driven Manipulator with Variable Stiffness and Multiple Motion
Directly 3D Printed, Pneumatically Actuated Multi-Material Robotic Hand
Soft Hand Extension Glove with Thumb Abduction and Extension Assistance
Design and Characterization of a Soft Flat Tube Twisting Actuator
Self-Retractable Soft Growing Robots for Reliable and Fast Retraction While Preserving Their Inh
High-Curvature, High-Force, Vine Robot for Inspection
Robotic Modules for a Continuum Manipulator with Variable Stiffness Joints
A Modular, Tendon Driven Variable Stiffness Manipulator with Internal Routing for Improved Stabil
A Biomorphic Whisker Sensor for Aerial Tactile Applications
Embedded Air Channels Transform Soft Lattices into Sensorized Grippers
Towards Automatic Design of Soft Pneumatic Actuators: Inner Structure Design Using CNN Model
Design of a Rigid-Soft Hybrid Robotic Glove with Force Sensing Function
SoftER: A Spiral Soft Robotic Ejector for Sorting Applications
Design and Validation of Soft Sliding Structure with Adjustable Stiffness for Ankle Sprain Preventio
Capacitive Origami Sensing Modules for Measuring Force in a Neurosurgical, Soft Robotic Retract
A Soft Miniaturized Continuum Robot with 3D Shape Sensing Via Functionalized Soft Optical Wav
Continuously Estimate and Control Prosthetic Grip Force by an Optical Waveguide Sensor
Multi-Tap Resistive Sensing and FEM Modeling Enables Shape and Force Estimation in Soft Robo
Learning Motion Reconstruction from Demonstration Via Multi-Modal Soft Tactile Sensing
A Generalized Motion Control Framework of Dielectric Elastomer Actuators: Dynamic Modeling, Sl
Vision-Based Tip Force Estimation on a Soft Continuum Robot
Soft Bending Actuator with Fiber-Jamming Variable Stiffness and Fiber-Optic Proprioception
A Light and Heat-Seeking Vine-Inspired Robot with Material-Level Responsiveness
Morphological Design for Pneumatic Soft Actuators and Robots with Desired Deformation Behavior
Thermally-Activated Biochemically-Sustained Reactor for Soft Fluidic Actuation
Pulsating Fluidic Sensor for Sensing of Location, Pressure and Contact Area
MoRC - a Modular Robot Controller
Enabling the Deployment of Any-Scale Robotic Applications in Microservice Architectures through
Plug’n Play Task-Level Autonomy for Robotics Using POMDPs and Probabilistic Programs
CoBRA: A Composable Benchmark for Robotics Applications
GSL-Bench: High Fidelity Gas Source Localization Benchmarking Tool
Cook2LTL: Translating Cooking Recipes to LTL Formulae Using Large Language Models
Toward Automated Programming for Robotic Assembly Using ChatGPT
A Method for Multi-Robot Asynchronous Trajectory Execution in MoveIt2
Improving the ROS 2 Navigation Stack with Real-Time Local Costmap Updates for Agricultural Appl
Super-Resolution of Lunar Satellite Images for Enhanced Robotic Traverse Planning
PPO-Based Dynamic Control of Uncertain Floating Platforms in Zero-G Environment
Learning-Aided Control of Robotic Tether-Net with Maneuverable Nodes to Capture Large Space D
Online Supervised Training of Spaceborne Vision During Proximity Operations Using Adaptive Kalm
Towards Real-World Efficiency: Domain Randomization in Reinforcement Learning for Pre-Captur
SPADES: A Realistic Spacecraft Pose Estimation Dataset Using Event Sensing
Covariance Based Terrain Mapping for Autonomous Mobile Robots
VINSat: Solving the Lost-In-Space Problem with Visual-Inertial Navigation
System Identification of Space Manipulator Systems and Its Implications on Robust Control Perfo
Model Design and Concept of Operations of Standard Interface for On-Orbit Construction
Enabling Faster Locomotion of Planetary Rovers with a Mechanically-Hybrid Suspension
RoboBall: An All-Terrain Spherical Robot with a Pressurized Shell
Enhanced Multifunctional Interface for Reconfigurability of Robotic Teams in Planetary Application
Autonomous Perching on Flat Surfaces for Free-Flying Robots with Gecko Adhesive Gripper
VWDER: A Variable Wheel-Diameter Ellipsoidal Robot
On Robust Control Laws Trade-Off Analysis for Space Manipulators with Uncertain Parameters a
Markerless Ultrasound Probe Pose Estimation in Mini-Invasive Surgery
Occlusion-Robust Autonomous Robotic Manipulation of Human Soft Tissues with 3D Surface Fee
Sensorless Transparency Optimized Haptic Teleoperation on the Da Vinci Research Kit
Multimodal Transformers for Real-Time Surgical Activity Prediction
Learning Needle Pick-And-Place without Expert Demonstrations
Learning Nonprehensile Dynamic Manipulation: Sim2real Vision-Based Policy with a Surgical Robo
Realistic Data Generation for 6D Pose Estimation of Surgical Instruments
Surgical Gym: A High-Performance GPU-Based Platform for Reinforcement Learning with Surgical
Multi-Objective Cross-Task Learning Via Goal-Conditioned GPT-Based Decision Transformers for
A Novel Robotic Bronchoscope with a Spring-Based Extensible Segment for Improving Steering Abi
Robust Surgical Tool Tracking with Pixel-Based Probabilities for Projected Geometric Primitives
Ada-Tracker: Soft Tissue Tracking Via Inter-Frame and Adaptive-Template Matching
Real-To-Sim Deformable Object Manipulation: Optimizing Physics Models with Residual Mappings
Efficient and Accurate Mapping of Subsurface Anatomy Via Online Trajectory Optimization for Rob
A Cross-Entropy Motion Planning Framework for Hybrid Continuum Robots
Evaluating the Task Generalization of Temporal Convolutional Networks for Surgical Gesture and
Lens Capsule Tearing in Cataract Surgery Using Reinforcement Learning
ORBIT-Surgical: An Open-Simulation Framework for Learning Surgical Augmented Dexterity
Iterative PnP and Its Application in 3D-2D Vascular Image Registration for Robot Navigation
Sim2Real Transfer of Reinforcement Learning for Concentric Tube Robots
A Kinetostatic Model for Concentric Push-Pull Robots
Fully Distributed Shape Sensing of a Flexible Surgical Needle Using Optical Frequency Domain Ref
Integrated Magnetic Location Sensing and Actuation of Steerable Robotic Catheters for Peripheral
Semi-Autonomous Robotic Manipulator for Minimally Invasive Aortic Valve Replacement
Robotic Needle Insertion with 2D Ultrasound – 3D CT Fusion Guidance
Flock-Formation Control of Multi-Agent Systems Using Imperfect Relative Distance Measurements
From Shadows to Light: A Swarm Robotics Approach with Onboard Control for Seeking Dynamic S
Multi-Swarm Interaction through Augmented Reality for Kilobots
Morphobot: A Platform for Morphogenesis in Robot Swarm
FireAntV3: A Modular Self-Reconfigurable Robot towards Free-Form Self-Assembly Using Attach
Reciprocal and Non-Reciprocal Swarmalators with Programmable Locomotion and Formations for
Automatically Designing Robot Swarms in Environments Populated by Other Robots: An Experime
CrazySim: A Software-In-The-Loop Simulator for the Crazyflie Nano Quadrotor
Long-HOT: A Modular Hierarchical Approach for Long-Horizon Object Transport
COAST: Constraints and Streams for Task and Motion Planning
Extending the Cooperative Dual-Task Space in Conformal Geometric Algebra
D-LGP: Dynamic Logic-Geometric Program for Reactive Task and Motion Planning
Indoor Exploration and Simultaneous Trolley Collection through Task-Oriented Environment Partiti
Effort Level Search in Infinite Completion Trees with Application to Task-And-Motion Planning
Sense in Motion with Belief Clustering: Efficient Gas Source Localization with Mobile Robots
R-LGP: A Reachability-Guided Logic-Geometric Programming Framework for Optimal Task and Mot
Solving Sequential Manipulation Puzzles by Finding Easier Subproblems
When Prolog Meets Generative Models: A New Approach for Managing Knowledge and Planning in
HAPFI: History-Aware Planning Based on Fused Information
Non-Axiomatic Reasoning for an Autonomous Mobile Robot
Asynchronous Task Plan Refinement for Multi-Robot Task and Motion Planning
Optimal Planning for Timed Partial Order Specifications
On the Convergence of a Closed-Loop Inverse Kinematics Solver with Time-Varying Task Functio
PROTAMP-RRT: A Probabilistic Integrated Task and Motion Planner Based on RRT
Distributed Multi-Robot Online Sampling with Budget Constraints
Coupled Active Perception and Manipulation Planning for a Mobile Manipulator in Precision Agricul
RAMP: A Benchmark for Evaluating Robotic Assembly Manipulation and Planning
Towards Safe Robot Use with Edged or Pointed Objects: A Surrogate Study Assembling a Human
Accelerating Long-Horizon Planning with Affordance-Directed Dynamic Grounding of Abstract Stra
Long-Horizon Planning and Execution with Functional Object-Oriented Networks
From Cooking Recipes to Robot Task Trees – Improving Planning Correctness and Task Efficie
Stepwise Large-Scale Multi-Agent Task Planning Using Neighborhood Search
Bayesian-Guided Evolutionary Strategy with RRT for Multi-Robot Exploration
Wearable Haptics for a Marionette-Inspired Teleoperation of Highly Redundant Robotic Systems
NetLfD: Network-Aware Learning from Demonstration for In-Contact Skills Via Teleoperation
Lightweight and Compliant Bilateral Teleoperation System with Anthropomorphic Arms for Aerial
Intelligent Mode-Switching Framework for Teleoperation
Digital Twin-Driven Mixed Reality Framework for Immersive Teleoperation with Haptic Rendering
Design Octree-Based Method to Improve Model-Mediated Teleoperation in Tactile Internet
Autonomous and Teleoperation Control of a Drawing Robot Avatar
Adaptive Haptic Control Interface for Safeguarding Robotic Teleoperation in Hazardous Steelmak
A Probabilistic Approach for Learning and Adapting Shared Control Skills with the Human in the L
TELESIM: A Modular and Plug-And-Play Framework for Robotic Arm Teleoperation Using a Digital
Synchronized Human-Humanoid Motion Imitation
SPOTS: Stable Placement of Objects with Reasoning in Semi-Autonomous Teleoperation Systems
Online Minimization of the Robot Silhouette Viewed from Eye-To-Hand Camera
IRoCo: Intuitive Robot Control from Anywhere Using a Smartwatch
Integrating Open-World Shared Control in Immersive Avatars
Hierarchical Deep Learning for Intention Estimation of Teleoperation Manipulation in Assembly Ta
Dynamic Mobile Manipulation Via Whole-Body Bilateral Teleoperation of a Wheeled Humanoid
3D Autocomplete: Enhancing UAV Teleoperation with AI in the Loop
Fine-Tuning Point Cloud Transformers with Dynamic Aggregation
MoPA: Multi-Modal Prior Aided Domain Adaptation for 3D Semantic Segmentation
Cross Domain Policy Transfer with Effect Cycle-Consistency
Parameter-Efficient Prompt Learning for 3D Point Cloud Understanding
BEVUDA: Multi-Geometric Space Alignments for Domain Adaptive BEV 3D Object Detection
6-DOF Grasp Pose Evaluation and Optimization Via Transfer Learning from NeRFs
Multi-Level Progressive Reinforcement Learning for Control Policy in Physical Simulations
Kalman Filter-Based One-Shot Sim-To-Real Transfer Learning
Towards Inclusive Dance: A Robotic Approach for Enhancing Movement and Interaction for People w
Design of Humanoid Robot Foot to Absorb Ground Reaction Force by Mimicking Longitudinal Arc
A Novel Modular Robot Inspired by Rubik's Snake
A Negative Pressure Adsorption Climbing Robot with Independent Steering Wheel for Machining P
Posture Estimation and Motion Planning for Plate Insertion into the Frame
Project YORI: Automated Cooking through Proprioceptive Dual-Arm Manipulation within a Structu
Mono Wi-Fi Scene
Robot Guardians: The Artificial Immune System of Our Planet
Gesture Detection System for Collaborative Robots
Universal Monocular 3D Human Recovery Engine
Collaborative Robot Assistance for Assembly: A Deep Learning Approach
Intelligent Dextrous Manipulation: Humanoid Agile Justin Meets Learning AI
Knowledge-Based Reasoning and Learning in Ad Hoc Teamwork
Inertial-Only Positioning for Human and Car Localization
Open Sourse Underwater Robot: Easys
3D Tracking of a Moving Object by a Multibeam Imaging Sonar-Equipped Autonomous Underwater V
Cut and Fire Resistant Variable Stiffness Gripper Mechanisms
Introduction to the Music Intervention Using Socially Engaging Robotics (MUSE) System for Perso
Autonomous Robotics for Transcatheter Delivery System
Roomac: Affordable Open-Source Autonomous Mobile Manipulation Robot - Fetch a Bottle Demo
Human-Robot Skill Transfer with Enhanced Compliance Via Dynamic Movement Primitives
Investigation of Turning Strategies for a Buoyancy Assisted Biped
FiReSpARX – FinTech/RegTech in Space for Trustful Autonomous Robotic Interaction
Optical Flow Sensor Enables Automatic Grasping Assistance for Prosthetic Hands
Digital Twin of a Robot Cell with Projected AR of Dynamic Safety Zones in a Pick'n Place Task Con
EEWOC: Extended-Reach Enhanced Wheeled Orb for Climbing
A Day with GARMI: Robot Home Assistance in the Lives of Senior Citizens
A Nitinol-Embedded Wearable Soft Robotic Gripper for Deep-Sea Manipulation
Toward Autonomous Tensegrity Robots
Tip Growing Actuators with Steering or Lift-Up Function
Small Phase-Change Buoyancy Regulator for Deep-Sea Robot
The Efforts of the ILSR Laboratory in the Surgical Robot
A Series-Parallel-Reconfigurable Tendon-Driven Supernumerary Robotic Arms for Human Operati
Effects of Impedance Control Stiffness on Assisted Elbow Flexion
Human Operation Augmentation through Wearable Robotic Limb Integrated with Mixed Reality Dev
Ensemble Latent Space Roadmap for Improved Robustness in Visual Action Planning
Direct 3D Model-Based Object Tracking with Event Camera by Motion Interpolation
Using Specularities to Boost Non-Rigid Structure-From-Motion
Tracking Snake-Like Robots in the Wild Using Only a Single Camera
Multi-Object Tracking by Hierarchical Visual Representations
AgriSORT: A Simple Online Real-Time Tracking-By-Detection Framework for Robotics in Precision
Tightly Coupled Visual-Inertial-UWB Indoor Localization System with Multiple Position-Unknown A
Stereo-NEC: Enhancing Stereo Visual-Inertial SLAM Initialization with Normal Epipolar Constraints
Nvblox: GPU-Accelerated Incremental Signed Distance Field Mapping
Exploitation-Guided Exploration for Semantic Embodied Navigation
Teach and Repeat Navigation: A Robust Control Approach
LOC-ZSON: Language-Driven Object-Centric Zero-Shot Object Retrieval and Navigation
Real-Time Localization for Closed-Loop Control of Assistive Furniture
Uncertainty-Aware Hybrid Paradigm of Nonlinear MPC and Model-Based RL for Offroad Navigation:
Robot Navigation in Unseen Environments Using Coarse Maps
Bicode: A Hybrid Blinking Marker System for Event Cameras
RAPIDFlow: Recurrent Adaptable Pyramids with Iterative Decoding for Efficient Optical Flow Estim
Guided by the Way: The Role of On-The-Route Objects and Scene Text in Enhancing Outdoor Nav
PlaceNav: Topological Navigation through Place Recognition
Aligning Knowledge Graph with Visual Perception for Object-Goal Navigation
Probable Object Location (POLo) Score Estimation for Efficient Object Goal Navigation
Bridging Zero-Shot Object Navigation and Foundation Models through Pixel-Guided Navigation Skil
GeoAdapt: Self-Supervised Test-Time Adaptation in LiDAR Place Recognition Using Geometric Pri
ViPlanner: Visual Semantic Imperative Learning for Local Navigation
UIVNAV: Underwater Information-Driven Vision-Based Navigation Via Imitation Learning
Wait, That Feels Familiar: Learning to Extrapolate Human Preferences for Preference-Aligned Pat
ASP-LED: Learning Ambiguity-Aware Structural Priors for Joint Low-Light Enhancement and Deblu
AnyOKP: One-Shot and Instance-Aware Object Keypoint Extraction with Pretrained ViT
RenderOcc: Vision-Centric 3D Occupancy Prediction with 2D Rendering Supervision
DRO: Deep Recurrent Optimizer for Video to Depth
Doduo: Learning Dense Visual Correspondence from Unsupervised Semantic-Aware Flow
RoboLLM: Robotic Vision Tasks Grounded on Multimodal Large Language Models
CrossVideo: Self-Supervised Cross-Modal Contrastive Learning for Point Cloud Video Understand
FSNet: Redesign Self-Supervised MonoDepth for Full-Scale Depth Prediction for Autonomous Driv
V2CE: Video to Continuous Events Simulator
MOTPose: Multi-Object 6D Pose Estimation for Dynamic Video Sequences Using Attention-Based
Generalizable Thermal-Based Depth Estimation Via Pre-Trained Visual Foundation Model
OSSAR: Towards Open-Set Surgical Activity Recognition in Robot-Assisted Surgery
FSD: Fast Self-Supervised Single RGB-D to Categorical 3D Objects
Keypoint Detection and Tracking in Low-Quality Image Frames with Events
TiV-ODE: A Neural ODE-Based Approach for Controllable Video Generation from Text-Image Pair
TVFusionGAN: Thermal-Visible Image Fusion Based on Multi-Level Adversarial Network Strategy
You Only Label Once: 3D Box Adaptation from Point Cloud to Image with Semi-Supervised Learni
Self-Supervised Pretraining and Finetuning for Monocular Depth and Visual Odometry
From Bird’s-Eye to Street View: Crafting Diverse and Condition-Aligned Images with Latent Dif
Lightning NeRF: Efficient Hybrid Scene Representation for Autonomous Driving
Physical Priors Augmented Event-Based 3D Reconstruction
SLAM Based on Camera-2D LiDAR Fusion
NeRF-Enhanced Outpainting for Faithful Field-Of-View Extrapolation
DL-PoseNet: A Differential Lightweight Network for Pose Regression Over SE(3)
Crossway Diffusion: Improving Diffusion-Based Visuomotor Policy Via Self-Supervised Learning
Bi-KVIL: Keypoints-Based Visual Imitation Learning of Bimanual Manipulation Tasks
Neural Radiance Fields for Unbounded Lunar Surface Scene
Bag of Views: An Appearance-Based Approach to Next-Best-View Planning for 3D Reconstruction
CopperTag: A Real-Time Occlusion-Resilient Fiducial Marker
Robust Collaborative Perception without External Localization and Clock Devices
DerainNeRF: 3D Scene Estimation with Adhesive Waterdrop Removal
Learning Interaction Regions and Motion Trajectories Simultaneously from Egocentric Demonstrat
Marrying NeRF with Feature Matching for One-Step Pose Estimation
Occluded Part-Aware Graph Convolutional Networks for Skeleton-Based Action Recognition
MAL: Motion-Aware Loss with Temporal and Distillation Hints for Self-Supervised Depth Estimation
See through the Real World Haze Scenes: Navigating the Synthetic-To-Real Gap in Challenging
Semi-Supervised Learning for Visual Bird's Eye View Semantic Segmentation
OpenAnnotate3D: Open-Vocabulary Auto-Labeling System for Multi-Modal 3D Data
SAM-Event-Adapter: Adapting Segment Anything Model for Event-RGB Semantic Segmentation
BuFF: Burst Feature Finder for Light-Constrained 3D Reconstruction
Unsupervised Spike Depth Estimation Via Cross-Modality Cross-Domain Knowledge Transfer
PillarGen: Enhancing Radar Point Cloud Density and Quality Via Pillar-Based Point Generation Ne
Deep Dynamic Layout Optimisation of Photogrammetry Camera Position Based on Digital Twin
MMA-Net: Multiple Morphology-Aware Network for Automated Cobb Angle Measurement
Synset Boulevard: A Synthetic Image Dataset for VMMR
ContourPose: Monocular 6D Pose Estimation Method for Reflective Texture-Less Metal Parts
Soft Acoustic End-Effector
Deep Learning Based 6-DoF Antipodal Grasp Planning from Point Cloud in Random Bin-Picking Ta
Sim-To-Real Object Pose Estimation for Random Bin Picking
Action-By-Detection: Efficient Forklift Action Detection for Autonomous Mobile Robots in Warehou
RIDE: Self-Supervised Learning of Rotation-Equivariant Keypoint Detection and Invariant Descript
LLM-Assisted Multi-Teacher Continual Learning for Visual Question Answering in Robotic Surgery
Procedure Recognition by Knowledge-Driven Segmentation in Robotic-Assisted Vitreoretinal Surge
Stereo Image-Based Visual Servoing towards Feature-Based Grasping
Visual Feedback Control of an Underactuated Hand for Grasping Brittle and Soft Foods
Compositional Servoing by Recombining Demonstrations
Second-Order Position-Based Visual Servoing of a Robot Manipulator
Event-Triggered Image Moments Predictive Control for Tracking Evolving Features Using UAVs
Lattice-Based Shape Tracking and Servoing of Elastic Objects
DCPT: Darkness Clue-Prompted Tracking in Nighttime UAVs
Unifying Foundation Models with Quadrotor Control for Visual Tracking Beyond Object Categories
DroneMOT: Drone-Based Multi-Object Tracking Considering Detection Difficulties and Simultaneo
Object Permanence Filter for Robust Tracking with Interactive Robots
Zero-Shot Open-Vocabulary Tracking with Large-Scale Pre-Trained Models
Multi-Correlation Siamese Transformer Network with Dense Connection for 3D Single Object Track
Refining Pre-Trained Motion Models
SWTrack: Multiple Hypothesis Sliding Window 3D Multi-Object Tracking
UncertaintyTrack: Exploiting Detection and Localization Uncertainty in Multi-Object Tracking
Humanoid Loco-Manipulations Using Combined Fast Dense 3D Tracking and SLAM with Wide-Ang
LiteTrack: Layer Pruning with Asynchronous Feature Extraction for Lightweight and Efficient Visual
Field-VIO: Stereo Visual-Inertial Odometry Based on Quantitative Windows in Agricultural Open Fi
Online Calibration of a Single-Track Ground Vehicle Dynamics Model by Tight Fusion with Visual-I
VI-HSO: Hybrid Sparse Monocular Visual-Inertial Odometry
Square-Root Inverse Filter-Based GNSS-Visual-Inertial Navigation
Omnidirectional Dense SLAM for Back-To-Back Fisheye Cameras
Visual Inertial Odometry Using Focal Plane Binary Features (BIT-VIO)
PL-EVIO: Robust Monocular Event-Based Visual Inertial Odometry with Point and Line Features
JacobiGPU: GPU-Accelerated Numerical Differentiation for Loop Closure in Visual SLAM
MAVIS: Multi-Camera Augmented Visual-Inertial SLAM Using SE2(3) Based Exact IMU Pre-Integra
Vision-Based Wearable Steering Assistance for People with Impaired Vision in Jogging
Variable Grounding Flexible Limb Tracking Center of Gravity for Sit-To-Stand Transfer Assistance
Towards Enhanced Stability of Human Stance with a Supernumerary Robotic Tail
Active, Quasi-Passive, Pneumatic, and Portable Knee Exoskeleton with Bidirectional Energy Flow fo
Task-Based Human-Robot Collaboration Control of Supernumerary Robotic Limbs for Overhead T
Safe and Individualized Motion Planning for Upper-Limb Exoskeleton Robots Using Human Demonst
Pneumatic Back Exoskeleton for Lifting Posture Detection and Correction
Leaf-Inspired FSR Array and Insole-Type Sensor Module for Mobile Three-Dimensional Ground Re
Human-Exoskeleton Locomotion Interaction Experience Transfer: Speeding up and Improving the
Design of a Knee-Joint Exoskeleton to Reduce Misalignment in Both the Sagittal and Coronal Plan
Adaptive Active Disturbance Rejection Control of an Actuated Ankle Foot Orthosis for Ankle Move
A Novel Funnel-Based L1 Adaptive Fuzzy Approach for the Control of an Actuated Ankle Foot Orth
Online Camera Orientation Calibration Aided by a High-Speed Ground-View Camera
Fast Wheeled Driving to Legged Leaping Onto a Step in a Leg-Wheel Transformable Robot
Body Velocity Estimation in a Leg–Wheel Transformable Robot without a Priori Knowledge of
Rolling with Planar Parametric Curves for Real-Time Robot Locomotion Algorithms
Non-Smooth Trajectory Optimization for Wheeled Balancing Robots with Contact Switches and Im
Design and Central Pattern Generator Control of a New Transformable Wheel-Legged Robot
Planned Trajectory Classification for Wheeled Mobile Robots to Prevent Rollover and Slip
Mechanical Design and Kinematics of a Multimodal Two-Wheeled Robot
Global Tracking Control for Car-Like Mobile Robots with Zero-Crossing Driving Velocity
Reactive Landing Controller for Quadruped Robots
Hierarchical Optimization-Based Control for Whole-Body Loco-Manipulation of Heavy Objects
Toward Self-Righting and Recovery in the Wild: Challenges and Benchmarks
Design of Morphable StateNet Based on Pseudo-Generalization of Standing up Motions for Humano
Agile and Dynamic Standing-Up Control for Humanoids Using 3D Divergent Component of Motion i
Representing Robot Geometry As Distance Fields: Applications to Whole-Body Manipulation
Singularity-Robust Prioritized Whole-Body Tracking and Interaction Control with Smooth Task Tran
Learning Force Control for Legged Manipulation
A Study of Shared-Control with Bilateral Feedback for Obstacle Avoidance in Whole-Body Telelo
Authors Organisation Session
Alloy Das, Sanket Biswas, Umapada Pal, Josep LladósIndian Statistical Insti 2D/3D Visual Perce
Bowei Xing, Xianghua Ying, Ruibin Wang Peking University 2D/3D Visual Perce
Federico Rollo, Andrea Zunino, Nikos Tsagarakis, Enr Leonardo S.p.A,Leonardo 2D/3D Visual Perce
Kavisha Vidanapathirana, Peyman Moghadam, Sridha S Queensland University 2D/3D Visual Perce
Chao Wei, Zhidong Deng Tsinghua University 2D/3D Visual Perce
Chengxin Liu, Guiyou Chen, Ran Song Shandong University,Sh2D/3D Visual Perce
Biao Wang, Wenling Li, Bin Zhang, Yang Liu Beihang University,Beij 2D/3D Visual Perce
Philipp Ausserlechner, David Dylan Haberger, Stefan TU Wien,Vienna Univers2D/3D Visual Perce
Younggi Hong, Min Ju Kim, Isack Lee, Seok Bong Yoo Chonnam National Unive 2D/3D Visual Perce
Tianci Wang, Hao Wen, Zaixin Song, Zhiping Dong, ChCity University of Hong Actuation
Maxime Marchal, Raphaël Furnémont, Bram Vanderborg Vrije Universiteit Brus Actuation
Hongxi Zhu, Ulrike Thomas Chemnitz University of Actuation
Shuowen Yi, Siyu Liu, Junbei Liao, Zhao Guo Wuhan university,the S Actuation
Edmundo Pozo Fortunić, Mehmet Can Yildirim, Dennis Technical University o Actuation
Wesley Roozing, Glenn Roozing University of Twente,AutActuation
Huaiyuan Chen, Yiyuan Ma, Weidong Chen Shanghai Jiao Tong Univ Actuation
Timothée Portha, Laurent Barbé, Francois Geiskopf, J University of Strasbour Actuation
Sung Kim, David Braun Vanderbilt University Actuation
Simon Martinez-rozas, David Alejo, Fernando CaballeroUniversidad de AntofagaAerial Systems
Han Jiang, Yanchun Chang, Liying Yang, Xu Liu, Yuqi the State Key LaboratorAerial Systems
Reiya Takemura, Genya Ishigami Keio University Aerial Systems
Alessandro Garzelli, Kumud Darshan Yadav, Alessandro GRVC Robotics Lab, Unive Aerial Systems
Elia Cereda, Manuele Rusci, Alessandro Giusti, DanieleUSI and SUPSI,KU LeuvAerial Systems
Ezra Tal, Gilhyun Ryou, Sertac Karaman MIT,Massachusetts InstiAerial Systems
Jindou Jia, Wenyu Zhang, Kexin Guo, Jianliang Wang, Beihang University,Hangz Aerial Systems
Brady Moon, Sagar Sachdev, Junbin Yuan, Sebastian Carnegie Mellon UniversAerial Systems
Hanna Müller, Vlad Niculescu, Tommaso Polonelli, Mic ETH Zürich,ETH Zurich,U Aerial Systems
Biao Jia, Dinesh Manocha University of Maryland aAI Robotics
Francesco Cursi, Marcus Kalander, Shuang Wu, Xidi Xu Imperial College Londo AI Robotics
Michael Görner, Norman Hendrich, Jianwei Zhang University of Hamburg AI Robotics
Petre Gaskell, Jana Pavlasek, Tom Gao, Abhishek Naru University of Michigan AI Robotics
Henry Smith, Ajay Shankar, Jennifer Gielis, Jan Blu University of CambridgeAI Robotics
Fengze Xie, Guanya Shi, Michael O'connell, Yisong Y California Institute of AI Robotics
Xianlei Long, Hui Zhao, Chao Chen, Fuqiang Gu, QingyChongqing University,Co AI Robotics
Kangcheng Liu, Xinhu Zheng, Chaoqun Wang, Lin Wang ETH Zurich,The HongKon AI Robotics
Keisuke Shirai, Cristian Beltran-Hernandez, Masashi Kyoto University,OMROAI-Based Methods
Bingheng Wang, Xuyang Chen, Lin Zhao National University of S AI-Based Methods
He Zhu, Li Junyi, Hua Yang, Jiankui Chen, Zhouping YiHust,Huazhong UniversiAI-Based Methods
Wenke Xia, Dong Wang, Xincheng Pang, Zhigang Wang, Renmin University of ChiAI-Based Methods
Zhehua Zhou, Jiayang Song, Kunpeng Yao, Zhan Shu,University of Alberta,MaAI-Based Methods
Rizwana Kausar, Fakhreddine Zayer, Jaime Viegas, JoKhalifa University,khalif AI-Based Methods
Victoria Magdalena Dax, Jiachen Li, Enna Sachdeva, NStanford University,UnivAI-Based Methods
Qingwen Zhang, Yi Yang, Heng Fang, Ruoyu Geng, Patr KTH Royal Institute o AI-Based Methods
Haoyu Wang, Xiaoyu Tan, Xihe Qiu, Chao Qu Shanghai University of EAI-Based Methods
Qingyuan Lu, Justin Svegliato, Samer Nashed, ShlomoMassachusetts
Z Institute AI-Enabled Robotic
Jingkai Sun, Qiang Zhang, Yiqun Duan, Xiaoyang Jian The Hong Kong UniversiAI-Enabled Robotic
Feiyang Wu, Zhaoyuan Gu, Hanran Wu, Anqi Wu, Ye Georgia Institute of Tec AI-Enabled Robotic
Rachel Luo, Rohan Sinha, Yixiao Sun, Ali Hindy, Shen Stanford University AI-Enabled Robotic
Andrey Zhitnikov, Vadim Indelman Technion – Israel Instit AI-Enabled Robotic
Huang Huang, Satvik Sharma, Antonio Loquercio, Anast University of California AI-Enabled Robotic
Akhilan Boopathy, Aneesh Muppidi, Peggy Yang, Abhiram Massachusetts InstituteAI-Enabled Robotic
Wei Kong, Li Hu, Du Qianjin, Cao Huayang, Kuang XiaNational Key LaboratoryAI-Enabled Robotic
Gloria Beraldo, Riccardo De Benedictis, Amedeo Cesta,National Research Counc AI-Enabled Robotics
Cristian Beltran-Hernandez, Nicolas Erbetti, Masashi OMRON SINIC X,OMRON AI-Enabled Robotics
Guangyao Zhai, Xiaoni Cai, Dianye Huang, Yan Di, Fa Technical University o AI-Enabled Robotics
Yongchao Chen, Jacob Arkin, Yang Zhang, Nicholas RHarvard University,MassAI-Enabled Robotics
w lesjie, Yichen Zhu, Minjie Zhu, Jinming Li, Zhiyuan East China Normal Unive AI-Enabled Robotics
Janosch Moos, Cedric Derstroff, Niklas Schröder, DeboTU Darmstadt, Institute AI-Enabled Robotics
Minjie Zhu, Yichen Zhu, Jinming Li, w lesjie, Zhiyuan East China Normal Unive AI-Enabled Robotics
Montse Gonzalez Arenas, Ted Xiao, Sumeet Singh, Vidh Google Inc,Google,Carne AI-Enabled Robotics
Cyrus Neary, Christian Ellis, Aryaman Singh Samyal, The University of TexasAI-Enabled Robotics
Kohei Honda, Ryo Yonetani, Mai Nishimura, Tadashi KNagoya University,Cyb AI-Enabled Robotics
Hongzhou Li, Sijie Yu, Shengkai Zhang, Guang Tan Sun Yat-sen Universit AI-Enabled Robotics
Xin Meng, Hongtao Wu, Sipu Ruan, Gregory Chirikjian National University of AI-Enabled Robotics
Pushkal Katara, Zhou Xian, Aikaterini Fragkiadaki Carnegie Mellon UniversAI-Enabled Robotics
Jiatao Zhang, Lanling Tang, Yufan Song, Qiwei Meng, Zhejiang University,Uni AI-Enabled Robotics
Tsun-Hsuan Johnson Wang, Alaa Maalouf, Wei Xiao, YMassachusetts InstituteAI-Enabled Robotics
Yongchao Chen, Jacob Arkin, Charles Dawson, Yang ZHarvard University,MassAI-Enabled Robotics
Naoki Yokoyama, Alexander Clegg, Joanne Truong, EriGeorgia Institute of T AI-Enabled Robotics
Joana Plewnia, Fabian Peller-konrad, Tamim Asfour Karlsruhe Institute of T AI-Enabled Robotics
Luca Morando, Giuseppe Loianno New York University Applications
Huaxin Tang, John Henry Burns, Alexander Strong, Yu Binghamton University Applications
Yao Su, Jingwen Zhang, Ziyuan Jiao, Hang Li, Meng WBeijing Institute for General Artificial Intelligence （BIGAI）,U
Applications
H S Helson Go, Hugh Hong-Tao Liu University of Toronto Applications
Mael Feurgard, Gautier Hattenberger, Simon Lacroix Ecole Nationale de l'Av Applications
Haoyu Wei, Shuai Wang, Quan Quan Beihang University Applications
Pravin Kumar Jaisawal, Stephanos Papakonstantinou, V Hamburg University of Applications
Martin Krizek, Matous Vrba, Antonella Barisic, Stjepan Czech Technical Universi Applications
Zhefan Xu, Christopher Suzuki, Xiaoyang Zhan, Kenji Carnegie Mellon UniversApplications
Maurits Reitsma, Julian Keller, Kenneth Blomqvist, Rol ETH Zurich Automation
Ewerton Vieira, Aravind Sivaramakrishnan, Sumanth TaRutgers University,RutgAutomation
Khai Nguyen, Sam Schoedel, Anoushka Alavilli, Brian Carnegie Mellon UniversAutomation
Shuzhang Liang, Satoshi Amaya, Hirotaka Sugiura, Hao The University of TokyoAutomation
Tian Shi, Weibing Li, Haoyong Yu, Yongping Pan Sun Yat-Sen University,Automation
Yuan Shen, Bhargav Chandaka, Zhi-hao Lin, Albert Zh UIUC,University of Illin Autonomous Agents
Haolan Liu, Liangjun Zhang, Siva Kumar Sastry Hari, J University of Californi Autonomous Agents
Jiaxin Pan, Changyao Zhou, Mariia Gladkova, Qadeer Technical University of Autonomous Agents
Divya Thuremella, Lewis Ince, Lars Kunze University of Oxford, RobAutonomous Agents
Yulong Cao, Boris Ivanovic, Chaowei Xiao, Marco Pav NVIDIA,University of MiAutonomous Agents
Ziyi Yang, Shreyas Sundara Raman, Ankit Jayesh Shah, Brown University,Massac Autonomous Agents
Wentao Wang, Haoran Xu, Guang Tan Sun Yat-sen Universi Autonomous Agents
Sheng Wang, Yingbing Chen, Jie Cheng, Xiaodong MeiHong Kong University oAutonomous Agents
Sungmin Woo, Minjung Kim, Donghyeong Kim, SungjunYonsei University Autonomous Agents
Golam Md Muktadir, Jim Whitehead University of California Autonomous Agents
Sanjay Sarma Oruganti Venkata, Ramviyas ParasuramRensselaer Polytechnic Autonomous
I Agents
Arezoo Samiei, Liang Sun USA,New Mexico State Autonomous
U Agents
Sandesh Athni Hiremath, Chengyi Huang, Argtim Tika, TU Kaiserslautern,RheinAutonomous Agents
Vishnu Sashank Dorbala, James Mullen, Dinesh Mano University of Maryland, Autonomous Agents
Lennart Puck, Maximilian Schik, Tristan Schnell, Timo FZI ForschungszentrumAutonomous
I Agents
Haotian Zhou, Yunhan Lin, Longwu Yan, Jihong Zhu, Wuhan University of Scie Autonomous Agents
Weilong Lin, Xinhua Zeng, Jing Teng, Pang Chengxin, JFudan University,North C Autonomous Agents
Long Chen, Oleg Sinavski, Jan Hünermann, Alice KarnsWayve,Wayve Technolog Autonomous Vehicle
Hetian Guo, Yingzhi Peng, Zipei Fan, He Zhu, Xuan SoSouthern University of Autonomous Vehicle
Tim Brühl, Tim Dieter Eberhardt, Robin Schwager, Lu Dr. Ing. h.c. F. Porsche Autonomous Vehicle
Aditya N G, Dhruval Pb, Jehan Shalabi, Shubhankar JaPES University,Purdue U Autonomous Vehicle
Jie Cheng, Yingbing Chen, Xiaodong Mei, Bowen Yang,Hong Kong University oAutonomous Vehicle
Tong Qin, Haihui Huang, Ziqiang Wang, Tongqing CheShanghai Jiao Tong Univ Autonomous Vehicle
Bernard Lange, Jiachen Li, Mykel Kochenderfer Stanford University,UniveAutonomous Vehicle
Jixiang Wan, Xudong Zhang, Shuzhou Dong, Yuwei Zhan OPPO Research InstitutAutonomous Vehicle
Anthony Knittel, Majd Hawasly, Stefano V. Albrecht, Five AI,FiveAI Ltd,UniveAutonomous Vehicle
Tong Li, Lu Zhang, Sikang Liu, Shaojie Shen Hong Kong University oAutonomous Vehicle
Qing Dong, Zhanhong Yan, Kimihiko Nakano, Xuewu Ji,Tsinghua University,the Autonomous Vehicle
Aniket Datar, Chenhui Pan, Mohammad Nazeri, Xuesu George Mason UniversitAutonomous Vehicle
Amir Hossain Raj, Zichao Hu, Haresh Karnan, Rohan CGeorge Mason University Autonomous Vehicle
Zhihao Wang, Shixing Huang, Minghang Li, Junyuan Harbin Institute of Tec Autonomous Vehicle
Tushar Choudhary, Vikrant Dewangan, Shivam Chandhok International Institute Autonomous Vehicle
Zesong Chen, Ze Yu, Jun Li, Linlin You, Xiaojun Tan Sun Yat-sen University,Autonomous Vehicle
Chang Chen, Yuecheng Liu, Yuzheng Zhuang, Sitong The Chinese University Autonomous Vehicle
Reza Mahjourian, Rongbing Mu, Valerii Likhosherstov, Waymo,Waymo UK Ltd Autonomous Vehicle
Minseong Park, Haengseon Son, Euntai Kim Yonsei University,KoreaAutonomous Vehicle
Shikuan Xie, Ran Song, Yuenan Zhao, Xueqin Huang, Y Shandong University Autonomous Vehicle
Tim Salzmann, Lewis Chiang, Markus Ryll, Dorsa SadigTechnical University M Autonomous Vehicle
Wenhui Huang, Zitong Shan, Shanhe Lou, Chen Lv NanYang TechnologicalAutonomous
U Vehicle
Zenan Li, Fan Nie, Qiao Sun, Fang Da, Hang Zhao Tsinghua University,ShaAutonomous Vehicle
Tao Huang, Gang Wang, Hongliang Liu, Jun Luo, LangChongqing University,Hu Autonomous Vehicle
Fei Wang, Yuesheng He, Hanyang Zhuang, Chenxi YaShanghai JiaoTong Unive Autonomous Vehicle
Yu Zhai, Yanzi Miao China University of min Autonomous Vehicle
Shuhao Ye, Yuxiang Cui, Hao Sha, Sha Lu, Yu Zhang, Zhejiang University Autonomous Vehicle
Matthias Mueller, Samarth Brahmbhatt, Ankur Deka, Que Intel Labs,Intel CorporatBig Data in Roboti
Kan Chen, Runzhou Ge, Hang Qiu, Rami Al-rfou, CharlWaymo LLC,University of Big Data in Roboti
Eckart Uhlmann, Mitchel Polte, Julian Bllumberg, She TU Berlin, Institute fo Big Data in Roboti
Xingyu Lin, John Ian So, Sashwat Mahalingam, Fangche UC Berkeley,Stanford Uni Big Data in Roboti
Homanga Bharadhwaj, Jay Vakil, Mohit Sharma, AbhinCarnegie Mellon UniversBig Data in Roboti
Ivan Kapelyukh, Yifei Ren, Ignacio Alzugaray, Edward Imperial College LondonBig Data in Roboti
Jingyun Yang, Max Sobol Mark, Brandon Vu, Archit ShStanford University Big Data in Roboti
Scott Ettinger, Kratarth Goel, Avikalp Srivastava, Rami Waymo Big Data in Roboti
Paulo Canelas, Trenton Tabor, John-Paul Ore, Alcides Carnegie Mellon UniversBig Data in Roboti
Haoyu Wang, Wenfu Xu, Linpo Hou, Erzhen Pan Harbin Institute of Tec Bioinspired Flight
Shuqi Wang, Jizhuang Fan, Yitao Pan, Gangfeng Liu, YHarbin Institute of Tech Bioinspired Flight
Ernesto Sanchez-laulhe, Álvaro César Satué Crespo, SUniversity of Malaga,GRBioinspired Flight
Yong Zhong, Zicun Hong, Yuhan Li, Junzhi Yu South China University Bioinspired Flight
Shipeng Liu, Boyuan Huang, Feifei Qian University of Southern CBioinspired Flight
Liang Li, Li-ming Chao, Siyuan Wang, Oliver Deussen, Max-Planck Institute of Bioinspired Flight
Shikun Li, Yufan Zhai, Chen Wang, Guangming Xie Peking University Bioinspired Flight
Sijia Liu, Chunbao Liu, Yunhong Liang, Luquan Ren, L Jilin University,Universi Bioinspired Flight
Junfei Li, Simon X. Yang University of Guelph Bioinspired Flight
Guillaume Bellegarda, Milad Shafiee, Auke Ijspeert EPFL Bioinspired Locomo
Jostein Løwer, Irja Gravdahl, Damiano Varagnolo, ØyviNorwegian University o Bioinspired Locomo
Katelyn King, Shai Revzen University of Michigan Bioinspired Locomo
Guillaume Bellegarda, Milad Shafiee, Merih Ekin Özberk EPFL,École Polytechni Bioinspired Locomo
Daryn Bright, Stacey Leigh Shield, Amir Patel University of Cape Tow Bioinspired Locomo
Jichao Wang, Jinyu Cheng, Jiangtao Hu, Wei Gao, Sh University of Science a Bioinspired Locomo
Yuzhen Pan, Rezwan Al Islam Khan, Chenyun Zhang, Fudan University,FudanBioinspired Locomo
Jiangfeng Lu, Jiaming Liang, Dekuan Zhu, Dongkai Wan Tsinghua University,TenBioinspired Locomo
Sean Even, Holden Gordon, Hoeseok Yang, Yasemin University of Notre DamBioinspired Locomo
Takashi Ozaki, Norikazu Ohta, Tomohiko Jimbo, Kana Toyota Central R&D La Bioinspired Microro
Yi Xu, Yanzhou Jin, Weitao Zhang, Yunhao Si, Yulai ZhBeijing Institute of Tech Bioinspired Microro
Shashwat Singh, Zeynep Temel, Ryan St. Pierre Carnegie Mellon Universi Bioinspired Microro
Hang Gao, Sunghwan Jung, Farrell E. Helbling Cornell University Bioinspired Microro
Elijah Blankenship, Conor Trygstad, Francisco GoncalvWashington State Univer Bioinspired Microro
Michiel Vital M Firlefyn, Jesse Hagenaars, Guido De C TU Delft,Delft Universit Bioinspired Microro
Fangyuan Liu, Song Li, Jinwu Xiang, Daochun Li, ZhanBeihang University Bioinspired Microro
Yang Gu, Xianfa Cai, Khadga Thakuri, Wenyu Yang, YuNanjing University of P Bioinspired Microro
Zhe Chen, Xie Chen, Shingo Shimoda, Qiang Huang, QBeijing Institute of Te Bioinspired Microro
Ujjal Mavinkurve, Ayato Kanada, Amir Tafrishi, Koki Kyushu University,CardifBioinspired Robot Ab
Zhipeng He, Na Zhao, Yudong Luo, Sian Long, Xi Luo, Beijing Institute of Tec Bioinspired Robot Ab
Velin Kojouharov, Tianyu Wang, Matthew Fernandez, Georgia Institute of Tec Bioinspired Robot Ab
Xiangyu Chu, Shengzhi Wang, Raymond Ng, Chun YinThe F Chinese University Bioinspired Robot Ab
Xiao Chen, Junyi Han, Xin Jin, Shuhei Miyashita University of Sheffield,un
Bioinspired Robot Ab
Hoyoung Kim, Jungwon Yoon GIST,Gwangju Institutu Bioinspired Robot Ab
Sunil Kumar Rajendran, Qi Wei, Ningshi Yao, Feitian ZBSS Technologies Inc., Bioinspired Robot Ab
Linghao Chen, Yuzhe Qin, Xiaowei Zhou, Hao Su Zhejiang University,UC Calibration and Ident
Zhaotong Luo, Guohang Yan, Xinyu Cai, Botian Shi Tsinghua University,ShaCalibration and Ident
Hongbo Zhao, Yikang Zhang, Qijun Chen, Rui Fan Tongji University Calibration and Ident
Jaroslav Moravec, Radim Šára Faculty of Electrical En Calibration and Ident
Jaeho Shin, Seungsang Yun, Ayoung Kim SNU,Seoul National Univ Calibration and Ident
Zhanhong Huang, Xiao Zhang, Antony Garcia, Xinmin Worcester Polytechnic In Calibration and Ident
Jiahao Pi, Guohang Yan, Chengjie Wang, Xinyu Cai, BoShanghai AI Laboratory,Calibration and Ident
Zhipeng Lin, Zhi Gao, Xinyi Liu, Jialiang Wang, Weiw The Chinese University Calibration and Ident
Thomas Gossard, Andreas Ziegler, Levin Kolmar, JonaUniversity of Tübingen, Calibration and Ident
Yu-chen Lee, Kuan-Wen Chen National Yang Ming Chia Calibration and Ident
Xingyu Yang, Yixiong Du, Leihui Li, Zhengxue Zhou, X Aarhus University,UniverCalibration and Ident
Muhammet Yunus Seker, Oliver Kroemer Carnegie Mellon UniversCalibration and Ident
Xuzhong Hu, Zaipeng Duan, Junfeng Ding, Zhe Zhang,Huazhong University of Calibration and Ident
Yihao Liu, Jiaming Zhang, Zhangcong She, Amir Khe Johns Hopkins UniversitCalibration and Ident
Zizhou Lao, Yuanfeng Han, Yunshan Ma, Gregory ChiriNational University of S Calibration and Ident
Yuxuan Xiao, Yao Li, Chengzhen Meng, Xingchen Li, JUniversity of Science a Calibration and Ident
Wanli Xing, Shijie Lin, Lei Yang, Jia Pan The University of Hon Calibration and Ident
Jun Zhang, Zihan Yang, Fangwei Zhang, Zhenyu Wu, GNanyang Technological Calibration and Ident
Helei Duan, Bikram Pandit, Mohitvishnu S. Gadde, Ba Oregon State UniversityCognitive Robotics
Chong Zhang, Jin Jin, Jonas Frey, Nikita Rudin, Mati ETH Zurich,Tongji UniveCognitive Robotics
Naoki Yokoyama, Sehoon Ha, Dhruv Batra, Jiuguang WGeorgia Institute of Te Cognitive Robotics
Ajay Sridhar, Dhruv Shah, Catherine Glossop, Sergey University of California Cognitive Robotics
Shengchao Yan, Baohe Zhang, Yuan Zhang, Joschka BUniversity of Freiburg, Cognitive Robotics
Javier Muñoz Mendi, Peter Lehner, Luis Moreno, Alin Universidad Carlos III Collision Avoidance
Oscar De Groot, Anish Sridharan, Javier Alonso-Mora, Delft University of Tec Collision Avoidance
Ze Zhang, Hadi Hajieghrary, Emmanuel Dean-Leon, KnChalmers University of Collision Avoidance
Zeming Wu, Zhuping Wang, Hao Zhang Tongji University Collision Avoidance
Wei Zhu, Mitsuhiro Hayashibe Tohoku University Collision Avoidance
Joonhee Lim, Kibeom Lee, Dongsuk Kum KAIST,Gachon UniversitCollision Avoidance
Nikhil Chandak, Kenny Chour, Sivakumar Rathinam, R IIIT Hyderabad,Texas ACollision Avoidance
Yinchuan Wang, Nianfei Du, Yongsen Qin, Xiang ZhanShandong University,Sch Collision Avoidance
Jianwei Liu, Maria Stamatopoulou, Dimitrios Kanoulas University College LondCollision Avoidance
Daniel Keren, Amit Shahar, Roi Poranne University of Haifa Collision Avoidance
Wenhao Yu, Jie Peng, Quecheng Qiu, Hanyu Wang, LuUniversity of Science anCollision Avoidance
Luca Paparusso, Shreyas Kousik, Edward Schmerling, Politecnico di Milano,GeCollision Avoidance
Alexandre Amice, Peter Werner, Russ Tedrake MIT,Massachusetts InstiCollision Avoidance
Yankai Yin, Zhe Sun, Peiying Ruan, Feng Duan, RuidonNankai University,RIKE Collision Avoidance
Xi Huang, Gergely Sóti, Christoph Ledermann, Björn HeKarlsruhe Institute of T Collision Avoidance
Muhammad Alhaddad, Konstantin Mironov, Aleksei StaMoscow Institute of Ph Collision Avoidance
Senthil Hariharan Arul, Jong Jin Park, Vishnu Prem, University of Maryland, Collision Avoidance
Adarsh Jagan Sathyamoorthy, Kasun Weerakoon, Moha University of Maryland, Collision Avoidance
Abraham Abe Leininger, Mahmoud Ali, Hassan Jardali, Indiana University Collision Avoidance
Jingjing Li, Jialin Ji, Qianhao Wang, Huan Yu, Yu Pan, Zhejiang University Collision Avoidance
Minjae Song, Yeseung Kim, Min Jun Kim, Daehyung PaKAIST,Korea AdvancedCollisionI Avoidance
Zhuo Yao, Wei Wang, Jiadong Zhang, Yan Wang, JinjiaBeihang University,BeihCollision Avoidance
Kegan Strawn, Nora Ayanian, Lars Lindemann University of Southern CCollision Avoidance
Kasun Weerakoon, Adarsh Jagan Sathyamoorthy, JingUniversity
L of Maryland, Collision Avoidance
Hongcheng Ji, Haibo Xie, Cheng Wang, Huayong YangZhejiang University,ZheJCollision Avoidance
Shaghayegh Keyumarsi, Made Widhi Surya Atman, Azw Tampere University Collision Avoidance
Jensen Gao, Bidipta Sarkar, Fei Xia, Ted Xiao, Jiajun Stanford University,GooComputer Vision fo
Sicong Pan, Liren Jin, Hao Hu, Marija Popovic, Maren University of Bonn,FudaComputer Vision fo
Hao Hu, Sicong Pan, Liren Jin, Marija Popovic, Maren Fudan University,UniverComputer Vision fo
Mathijs Ruben Van Geerenstein, Felicia Ruppel, Klaus Delft University of Tec Computer Vision fo
Haowen Wang, Zhen Zhao, Zhao Jin, Zhengping Che, LBeijing University of P Computer Vision fo
Jintao Cheng, Kang Zeng, Zhuoxu Huang, Xiaoyu Tang, South China Normal Univ Computer Vision fo
Xinyi Yu, Liqin Lu, Rong Jintao, Guangkai Xu, Linlin OuZhejiang University of TComputer Vision fo
Junho Lee, Sang Min Kim, Yonghyeon Lee, Young MinSeoul National UniversitComputer Vision fo
Wenbo Wang, Gen Li, Miguel Zamora, Stelian Coros ETH Zurich Computer Vision fo
Maximilian Luz, Rohit Mohan, Ahmed Rida Sekkat, Oli University of Freiburg, Computer Vision for
Jannik Zuern, Ingmar Posner, Wolfram Burgard University of Freiburg, Computer Vision for Transportation
Shuaijun Wang, Rui Gao, Ruihua Han, Qi Hao Southern University Computer Vision for
Zhengqi Bai, Wenjun Shi, Dongchen Zhu, Hanlong Kang Shanghai Institute of M Computer Vision for
Yilun Wu, Federico Paredes-valles, Guido De Croon TU Delft,Delft Universit Computer Vision for
Suozhi Huang, Juexiao Zhang, Yiming Li, Chen Feng Tsinghua University,New Computer Vision for
Kazuto Nakashima, Ryo Kurazume Kyushu University Computer Vision for
Wonhyeok Choi, Mingyu Shin, Hyukzae Lee, Jaehoon DGIST,daegu gyeongbuk Computer Vision for
Zixiang Zhou, Dongqiangzi Ye, Weijia Chen, Yufei Xi University of Central F Computer Vision for
Genliang Chen, Shujie Tang, Shaoqiu Xu, Tong Guan, Shanghai Jiao Tong Univ Contact Modeling
Mingxuan Li, Yen Hang Zhou, Tiemin Li, Yao Jiang Tsinghua University,Tsi Contact Modeling
Laurent Yves Emile Ramos Cheret, Vinicius Prado Da FLakehead University,Mem Contact Modeling
Kevin Haninger, Kangwagye Samuel, Filippo Rozzi, SeFraunhofer IPK,DGIST,P Contact Modeling
Zhenyu Lu, Jialong Yang, Haoran Li, Yifan Li, Weiyon Bristol Robotics LaboratContact Modeling
Mahdi Saleh, Michael Sommersperger, Nassir Navab, Technical University MuContact Modeling
Yujie Tang, Hai Zhu, Susan Potters, Martijn Wisse, WeDelft University of Tech Contact Modeling
J. Zachary Woodruff, Kevin Lynch Northwestern UniversityContact Modeling
Zixi Liu, Robert D. Howe Harvard University Contact Modeling
Shuo Cheng, Danfei Xu Gatech,Georgia Institut Continual Learning
Junha Song, Kwanyong Park, Inkyu Shin, Sanghyun WKAIST Continual Learning
Soumya Banerjee, Vinay Kumar Verma, Avideep Mukher IIT Kanpur,AMAZON,Unive Continual Learning
Chao Zhao, Jie Xu, Ru Peng, Xingyu Chen, Kuizhi Mei Xi'an Jiaotong Universit Continual Learning
Pierrick Lorang, Helmut Horvath, Tobias Kietreiber, PatAIT Austrian Institute Continual Learning
Georgios Tziafas, Hamidreza Kasaei University of GroningenContinual Learning
Meenal Parakh, Alisha Fong, Anthony Simeonov, Tao CPrinceton University,MaContinual Learning
Senlin Fang, Yi Wen Liu, Chengliang Liu, Jingnan Wa Shenzhen Institute of A Continual Learning
Weikang Wan, Yifeng Zhu, Rutav Shah, Yuke Zhu Peking University,The UnContinual Learning
Song Li, Fangyuan Liu, Yuzhe Gao, Jinwu Xiang, ZhanBeihang University Cooperating Cellula
Zhengrong Xue, Han Zhang, Cheng Jingwen, Zhengmao Tsinghua University,TsinCooperating Cellula
Jonathan Külz, Matthias Althoff Technical University of Cooperating Cellula
Sudheera Akalanka, Harith Sandeepa, Manu Athauda PUniversity of Moratuwa,Cooperating Cellula
Burak Aksoy, John Wen Rensselaer Polytechnic Cooperating
I Cellula
Yun Chang, Luca Ballotta, Luca Carlone MIT,Delft University of Cooperating Cellula
Zhehan Li, Rui Mao, Nanhe Chen, Chao Xu, Fei Gao, Zhejiang University,SunCooperating Cellula
Chenghao Yu, Dengyu Zhang, Dr. Qingrui Zhang Sun Yat-sen University,Cooperating Cellula
Simon Marc Pannek, Shervin Dehghani, Michael Sommers TUM,Technical UniversitData Sets for Roboti
Alberto Bacchin, Leonardo Barcellona, Sepideh Shamsiz University of Padua,UniData Sets for Roboti
Santosh Thoduka, Nico Hochgeschwender, Juergen Gall Hochschule Bonn-Rhein- Data Sets for Roboti
Boris Meden, Emanuel Pablo Vega, Fabrice Mayran DeUniversité Paris Saclay,Data Sets for Roboti
Yuxiang Yan, Boda Liu, Jianfei Ai, Qinbu Li, Ru Wan, JFudan University,Moo AData Sets for Roboti
Kaixin Bai, Lei Zhang, Zhaopeng Chen, Fang Wan, Jia University of Hamburg,SData Sets for Roboti
Asim Unmesh, Rahul Jain, Jingyu Shi, V. K. Chaitha Purdue University Data Sets for Roboti
Maxime Vaidis, Mohsen Hassanzadeh Shahraji, Effie Da Université Laval Data Sets for Roboti
Xingyu Liu, Chenyangguang Zhang, Gu Wang, Ruida Zh Tsinghua University Data Sets for Roboti
Ivan Tang, Cong Tai, Fang-xing Chen, Wanting Zhang,Tsinghua University,Ts Datasets for Robot
Kushal Kedia, Atiksh Bhardwaj, Prithwish Dan, Sanjib Cornell University Datasets for Robot
Napat Karnchanachari, Dimitrios Panagiotis GeromichaMotional,Motional AD,De Datasets for Robot
Allan Wang, Daisuke Sato, Yasser Corzo, Sonya Simkin, Carnegie Mellon UniversDatasets for Robot
Pierre Sermanet, Tianli Ding, Jeffrey Zhao, Fei Xia, D Google,Google Inc,GooDatasets for Robot
George Drayson, Efimia Panagiotaki, Daniel Omeiza, LUniversity College LondDatasets for Robot
Hao-shu Fang, Hongjie Fang, Zhenyu Tang, Jirong Li Shanghai Jiao Tong Univ Datasets for Robot
Noriaki Hirose, Dhruv Shah, Ajay Sridhar, Sergey LevinUC Berkeley / TOYOTADatasets
Mo for Robot
Chi-hsi Kung, Pang-yuan Pao, Pin-lun Chen, Hsin-chenNational Tsing Hua UnivDatasets for Roboti
Dennis Griesser, Matthias Franz, Georg Umlauf University of Applied ScDatasets for Roboti
Chirag Parikh, Rohit Saluja, C.V. Jawahar, Ravi Kiran International Institute Datasets for Roboti
Jin Fang, Dingfu Zhou, Jingjing Zhao, Chenming Wu, Baidu,tusimple,Baidu Res Datasets for Roboti
Jingyi Tang, Zeyu Chen, Bowen Fu, Wenjie Lu, Shengqua Tsinghua University,HarDatasets for Roboti
Peter Mortimer, Raphael Hagmanns, Miguel Granero, Th Universität der BundeswDatasets for Roboti
Yeong Gwang Son, Tat Hieu Bui, Juyong Hong, Yong HSungKyunKwan UniversiDatasets for Roboti
Diwei Sheng, Anbang Yang, John-ross Rizzo, Chen FeNew York University,NYDatasets for Roboti
Derek Cheng, Fernando Cladera, Ankit Prabhu, Xu Liu,University of PennsylvanDatasets for Roboti
Oggyu Lee, Kyungdon Joo, Jae-young Sim Ulsan National Institut Deep Learning for V
Shaohu Wang, Yuchuang Tong, Xiuqin Shang, Zhengt Institute of Automation Deep Learning for V
Nan Ma, Mohan Wang, Yiheng Han, Yong-Jin Liu Beijing University of Te Deep Learning for V
Lian Fu, Ryoichi Ishikawa, Yoshihiro Sato, Takeshi OishThe University of TokyoDeep Learning for V
Aldi Piroli, Vinzenz Dallabetta, Johannes Kopp, Marc Universität Ulm,BMW Gro Deep Learning for V
Haodi Yao, Ning Hao, Chen Xie, Fenghua He Harbin Institute of Tech Deep Learning for V
Guangyan Chen, Meiling Wang, Yi Yang, Li Yuan, Yuf Beijing Institute of tech Deep Learning for V
Irving Fang, Yuzhong Chen, Yifan Wang, Jianghan Zhan New York University,Xi'aDeep Learning for V
Jiayi Ni, Senqiao Yang, Ran Xu, Jiaming Liu, Xiaoqi L Peking University,Harbi Deep Learning for V
Jin Lu, Bangyan Niu, Huan Ma, Zhu Jiafeng, Jingjing JiHuazhong University of Deep Learning for V
Sang Uk Lee Motional Deep Learning for V
Lennart Schulze, Hod Lipson Columbia University,ColDeep Learning for V
Toan Nguyen, Minh Nhat Vu, Baoru Huang, Tuan Van V FPT Software,TU Wien,Deep A Learning for V
Seunghyeon Lim, Youngjae Yoo, Jun Ki Lee, Byoung- Seoul National UniversitDeep Learning for V
Sreejani Chatterjee, Duc [email protected], Berk Calli Worcester Polytechnic In Deep Learning for V
Christopher Maxey, Jaehoon Choi, Hyungtae Lee, Di University of Maryland Deep Learning for Vi
Jiaxu Xing, Leonard Bauersfeld, Yunlong Song, ChunwUniversity of Zurich,Uni Deep Learning for Vi
Florian Achermann, Julian Andreas Haug, Tobias ZumsETH Zurich, ASL,ETH Zu Deep Learning for Vi
Luca Crupi, Alessandro Giusti, Daniele Palossi IDSIA USI-SUPSI,ETH Z Deep Learning for Vi
Linfeng Li, Na Zhao Nanyang Technological Deep Learning for Vi
Claus Smitt, Michael Halstead, Patrick Zimmer, ThomasUniversity of Bonn,BonnDeep Learning for Vi
Anqi Cheng, Zhiyuan Yang, Haiyue Zhu, Kezhi Mao Nanyang Technological Deep Learning for Vi
Yang Tian, Jiyao Zhang, Guowei Huang, Bin Wang, Pi Peking University,Huawe Deep Learning for Vi
Runkai Zhao, Yuwen Heng, Heng Wang, Yuanda Gao, University
S of Sydney,BaDeep Learning for Vi
Zhen Luo, Junyi Ma, Zijie Zhou, Guangming Xiong Beijing Institute of Tech Deep Learning I
Daniel Honerkamp, Tim Welschehold, Abhinav Valada Albert Ludwigs UniversitDeep Learning I
Hugues Thomas, Jian Zhang, Timothy Barfoot University of Toronto,PuDeep Learning I
Jan Ole Von Hartz, Eugenio Chisari, Tim Welschehold University of Freiburg, Deep Learning I
Sepideh Afshar, Nachiket Deo, Akshay Bhagat, Titas Motional,UC San Diego,Deep Learning I
Shilin Shan, Quang-Cuong Pham Nanyang Technological Deep Learning I
Daniel Lawson, Ahmed H. Qureshi Purdue University Deep Learning I
Honghui Wang, Weiming Zhi, Gustavo Batista, RohitasUniversity of New Sou Deep Learning II
Lesong Tao, Miao Kang, Jinpeng Dong, Songyi Zhang,Xi'an Jiaotong Universit Deep Learning II
Takuma Yoneda, Jiading Fang, Peng Li, Huanyu Zhang, Toyota Technological InDeep Learning II
Yiqian Gan, Hao Xiao, Yizhe Zhao, Ethan Zhang, Zhe HTusimple, Inc,TuSimple Deep Learning II
Alina Böhm, Tim Schneider, Boris Belousov, Alap KshirTU Darmstadt,TechnicalDeep Learning II
Monika Farsang, Mathias Lechner, David Lung, Ramin TU Wien,MassachusettsDeep I Learning II
Baiyu Pan, Jichao Jiao, Jian Xin Pang, Jun Cheng University of Macau,BeiDeep Learning II
Feifei Li, Panwen Hu, Qi Song, Rui Huang The Chinese University Deep Learning II
Gabriele Mario Caddeo, Andrea Maracani, Paolo DidierIstituto Italiano di Tecn Deep Learning II
Esteve Valls Mascaro, Yashuai Yan, Dongheui Lee Technische Universitat Deep Learning III
Liang Xie, Wei Gao, Huiming Zheng, Ge Li Peking University,Peki Deep Learning III
Jeremy Morgan, David Millard, Gaurav Sukhatme University of Southern CDeep Learning III
Wenli Xiao, Tairan He, John Dolan, Guanya Shi Carnegie Mellon UniversDeep Learning III
Martin Bensch, Tim-David Job, Tim-Lukas Habich, ThoLeibniz University HanoDeep Learning III
Piotr Kicki, Puze Liu, Davide Tateo, Haitham Bou Amma Poznan University of T Deep Learning III
Vivek Gupta, Prabhpreet Dhir, Jeegn Dani, Ahmed H. QPurdue University, WestDeep Learning III
Kazumi Kasaura, Shuwa Miura, Tadashi Kozuno, Ryo YOMRON SINIC X,Universi Deep Learning III
Yitian Shi, Philipp Schillinger, Miroslav Gabriel, Al Bosch Center for Artif Deep Learning in Gr
Huy Le, Philipp Schillinger, Miroslav Gabriel, Alexan TU Dormund,Bosch Cente Deep Learning in Gr
Xujun Han, Shaochen Wang, Xiucai Huang, Zhen Kan University of Science a Deep Learning in Gr
Michael Lanighan, Oscar Youngquist TRACLabs, Inc,UniversiDeep Learning in Gr
Valerija Holomjova, Andrew Joe Starkey, Bruno Yun, PUniversity of Aberdeen, Deep Learning in Gr
Xinghao Zhu, Devesh Jha, Diego Romeres, Lingfeng SUniversity of California Deep Learning in Gr
Michelle Guo, Ziang Liu, Stephen Tian, Zhaoming Xie, Stanford University,CornDeep Learning in Gr
Chahyon Ku, Carl Winge, Ryan Diaz, Wentao Yuan, Kar University of Minnesota,Deep Learning in Gr
Zixing Wang, Ahmed H. Qureshi Purdue University Deep Learning in Gr
Cihan Acar, Kuluhan Binici, Alp Tekırdag, Yan Wu Institute for Infocomm Deep Learning in Gr
Colin Kohler, Anuj Shrivatsav Srikanth, Eshan Arora, RoNortheastern UniversityDeep Learning in Gr
Yixuan Huang, Jialin Yuan, Chanho Kim, Pupul PradhaUniversity of Utah,OregoDeep Learning in Gr
Xinyi Zhang, Yukiyasu Domae, Weiwei Wan, Kensuke Osaka University,The Na Deep Learning in Gr
Robert Lee, Jad Chakra, Fangyi Zhang, Peter Corke Australian Centre for R Deep Learning in Gr
Kai-tai Song, Hsiang-Hsi Chen National Yang Ming Chia Deep Learning in Gr
Jiakang Zhong, Yew Wee Wong, Jiong Jin, Yong Song,Swinburne University ofDeep Learning in Gr
Bao Thach, Tanner Watts, Shing-hei Ho, Tucker HermaUniversity of Utah Deep Learning in Gr
Annie Xie, Lisa Lee, Ted Xiao, Chelsea Finn Stanford University,GooDeep Learning in Gr
Yuxuan Kuang, Qin Han, Danshi Li, Qiyu Dai, Lian Di Peking University,New YDeep Learning in Gr
Mel Vecerik, Carl Doersch, Yi Yang, Todor Bozhinov D DeepMind,Google DeepM Deep Learning in Gr
Shih-Min Yang, Martin Magnusson, Johannes A. Stork,Örebro University,OrebrDeep Learning in Gr
Matias Sivertsvik, Kirill Sumskiy, Ekrem Misimi Norwegian University o Deep Learning in Gr
Qiaojun Yu, Junbo Wang, Wenhai Liu, Ce Hao, Liu Liu SHANGHAI JIAO TONGDeep UNIVLearning in Gr
Yushi Liu, Alexander Qualmann, Zehao Yu, Miroslav Gab University Tübingen,RobDeep Learning in Gr
Erdi Sayar, Zhenshan Bing, Carlo D’Eramo, Ozgur S. Technical University of Deep Learning in Gr
Jun Shi, Yong A, Yixiang Jin, Dingzhe Li, Haoyu Niu, Samsung Research China Deep Learning in Gr
Qingxiuxiong Dong, Toshimitsu Kaneko, Masashi Sugi Toshiba Corporation,The Deep Learning in Gr
Boshi An, Yiran Geng, Kai Chen, Xiaoqi Li, Qi Dou, Ha Peking University,The CDeep Learning in G
Pengwei Xie, Rui Chen, Siang Chen, Yuzhe Qin, FanboTsinghua University,UC Deep Learning in G
Zhanpeng He, Matei Ciocarlie Columbia University Deep Learning in G
Thomas Lampe, Abbas Abdolmaleki, Sandy H. Huang, Google Sar UK Ltd.,DeepMin Deep Learning in G
Pietro Mazzaglia, Taco Cohen, Daniel Dijkman University of Gent,Qu Deep Learning in G
Jungwook Mun, Khang Truong Giang, Yunghee Lee, Na Korea Advanced Institu Deep Learning in G
Siang Chen, Wei Tang, Pengwei Xie, Wenming Yang, G Tsinghua University Deep Learning in G
Hongxiang Yu, Anzhe Chen, Kechun Xu, Zhongxiang ZZhejiang University,Alib Deep Learning in G
Sverre Herland, Kerstin Bach, Ekrem Misimi Norwegian University o Deep Learning in G
Kaustab Pal, Aditya Sharma, Avinash Sharma, MadhavInternational Institute Deep Learning Met
Xu Zheng, Yunhao Luo, Chong Fu, Kangcheng Liu, Li The Hong Kong UniversiDeep Learning Met
Zhen Yao, Jiawei Xu, Shuhang Hou, Mooi Choo ChuahLehigh University Deep Learning Met
Zhixing Hou, Yuzhang Shang, Yan Yan Nanjing University of ScDeep Learning Met
Kai Luan, Chenghao Shi, Neng Wang, Yuwei Cheng, Hu Intelligence Science an Deep Learning Met
Mohamadreza Faridghasemnia, Jennifer Renoux, Alessa Orebro University,ÖrebrDeep Learning Met
David Benjamin Adrian, Andras Kupcsik, Markus Spie Bosch Corporate Research Deep Learning Met
Mingrui Li, Jiaming He, Yangyang Wang, Hongyu Wan Dalian University of TecDeep Learning Met
Nikhil Mishra, Maximilian Sieb, Pieter Abbeel, Xi Chen UC Berkeley,covariant.aDeep Learning Met
Irmak Guzey, Yinlong Dai, Ben Evans, Soumith Chintala New York University,N Dexterous Manipulat
Kejia Chen, Zhenshan Bing, Yansong Wu, Fan Wu, LidTechnical University o Dexterous Manipulat
Yongliang Wang, Kamal Mokhtar, Cock Heemskerk, HaUniversity of Groninge Dexterous Manipulat
Shuai Wang, Yihao Huang, Wang Wei Lee, Tianliang LiTencent,The University Dexterous Manipulat
Hamidreza Kasaei, Mohammadreza Kasaei University of Groningen,Dexterous Manipulat
Joshua Grace, Podshara Chanrungmaneekul, Kaiyu Han Yale University,Rice UniDexterous Manipulat
Qingtao Liu, Qi Ye, Zhengnan Sun, Yu Cui, Gaofeng LiZhejiang University Dexterous Manipulat
Kei Ota, Devesh Jha, Krishna Murthy, Asako Kanezak Tokyo Institute of Tech Dexterous Manipulat
Ling Tang, Yan-Bin Jia, Yuechuan Xue Iowa State University, Dexterous Manipulat
Yuanyuan Jia, Jessica Qu, Tadahiro Taniguchi Ritsumeikan University Dexterous Manipulat
Yansong Wu, Fan Wu, Lingyun Chen, Kejia Chen, Samu Technische Universität Dexterous Manipulat
Zhaole Sun, Jihong Zhu, Robert Fisher Tsinghua University, theDexterous Manipulat
Ruben Castro Ornelas, Tomas Cantu, Isabel Sperandio,Massachusetts InstituteDexterous Manipulat
Tomohiro Chaki, Tomohiro Kawakami Honda R&D Co., Ltd. Dexterous Manipulat
Bohan Yang, Bo Lu, Wei Chen, Fangxun Zhong, YunhuiThe Chinese University Dexterous Manipulat
Tao Pang, Hyung Ju Terry Suh, Lujie Yang, Russ Tedr Boston Dynamics AI InstDexterous Manipulat
Xianyi Cheng, Sarvesh Bipin Patil, Zeynep Temel, Oli Carnegie Mellon UniversDexterous Manipulat
Won Kyung Do, Bianca Aumann, Camille Chungyoun, Stanford University Dexterous Manipulat
Xin Lu, Zegui Wu, Ruqing Zhao, Fusheng Li University of Electronic Distributed Robot
Kaiyuan Eric Chen, Michael Wang, Marcus Gualtieri, NaUniversity of Californ Distributed Robot
Sadie Cutler, Kirstin Hagelskjaer Petersen Cornell University Distributed Robot
Augustinos Saravanos, Yuichiro Aoyama, Hongchang Georgia Institute of Tec Distributed Robot
Junfeng Chen, Zili Tang, Meng Guo Peking University Distributed Robot
Xinyi Wang, Jiwen Xu, Chuanxiang Gao, Yizhou Chen,The Chinese University Distributed Robot
Kai Matsuka, Soon-Jo Chung California Institute of Distributed Robot
Yan Cao, Teng Long, Jingliang Sun, Zhu Wang, GuangBeijing Institute of Tec Distributed Robot
Tianzhu Gao, Yudong Luo, Chao Lv, Weirong Luo, XianDalian Maritime UniversDynamics
Na Zhao, Yudong Luo, Chaojun Qin, Xi Luo, Rong CheDalian Maritime UniversDynamics
Ruijia Liu, Shaoyuan Li, Xiang Yin Shanghai Jiao Tong UniDynamics
Siqi Wang, Shaoyuan Li, Li Yin, Xiang Yin Shanghai Jiao Tong UniDynamics
Marcel Gabriel Lahoud, Gabriele Marchello, MariapaolaItalian Institute of Tech Dynamics
Frédéric Boyer, Andrea Gotelli, Philipp Tempel, Vincen IMT atlantique,École CeDynamics
Ajay Suresha Sathya, Herman Bruyninckx, Wilm Decré,Inria,KU Leuven,KatholiDynamics
Xu Dai, Ali Raoofian, Jozsef Kovecses, Marek TeichmaMcGill University,CMLabDynamics
Sriram Siva, Maggie Wigness Army Research Laborato Dynamics
Xiang Zhang, Masayoshi Tomizuka, Hui Li University of California Factory/Assembly A
Qi Zhang, Zhengtao Hu, Weiwei Wan, Kensuke HaradaOsaka University Factory/Assembly A
Dong Li, Quan Mu, Yilin Yuan, Shiwei Wu, Hualin Hong,Chongqing University,FoFactory/Assembly A
Yunsheng Tian, Karl Willis, Bassel Al Omari, Jieliang MIT,Autodesk,Universit Factory/Assembly A
Tena Žužek, Rok Vrabič, Andrej Zdesar, Gašper Škulj, U I niversity of Ljubljana,FFactory/Assembly A
Zhenyu Wu, Wei Wang, Chunyang Zhao, Yufeng Yue, Nanyang Technological Factory/Assembly
Un A
Carlos Diaz Alvarenga, Nicola Basilico, Stefano Carpin University of California Factory/Assembly A
Carlos Diaz Alvarenga, Nicola Basilico, Stefano Carpin University of California Factory/Assembly A
James Maier, Prasanna Sriganesh, Matthew Travers Carnegie Mellon UniversFactory/Assembly A
Haixin Zhao, Ruifeng Li, Quan Quan Beihang University Failure Detection
Benn Proper, Alexander Andreas Kurdas, Saeed AbdolEindhoven University o Failure Detection
Jan Thieß Brockmann, Marco Rudolph, Bodo Rosenhahvoraus robotik GmbH,Leib Failure Detection
Arda Inceoglu, Eren Erdal Aksoy, Sanem Sariel istanbul technical unive Failure Detection
Zeren Luo, Erdong Xiao, Peng Lu The University of Hong Failure Detection
Renzo Caballero, Piotr PiÄ™kos, Eric Feron, Jürgen King Abdullah Universit Failure Detection
Jasper Van Beers, Prashant Solanki, Coen De Visser Delft University of Tech Failure Detection
Felipe Toledo, Trey Woodlief, Sebastian Elbaum, Matt University of Virginia Failure Detection
Jeon Ho Kang, Neel Dhanaraj, Siddhant Ravindra Wada University of Southern CFailure Detection
Boseong Kim, Hyunki Seong, David Hyunchul Shim Korea Advanced Institu Field Robot System
Angus Lynch, Corey Duguid, Joao Buzzatto, Minas LiarUniversity of Auckland,TField Robot System
Kshitij Goel, Wennie Tabib Carnegie Mellon UniversField Robot System
Zhengzhe Xu, Yanbo Chen, Zhuozhu Jian, Junbo Tan, Harbin Institute of Tech Field Robot System
Turcan Tuna, Julian Nubert, Yoshua Alfredo Nava ChocETH Zurich, Robotic Sy Field Robot System
Giacomo Picardi, Anna Astolfi, Marcello Calisti Instituto de Ciencias d Field Robot System
Finn Lukas Busch, Nathalie Bauschmann, Sami Haddadi Hamburg University of TField Robot System
Weimin Qi, Qinbo Sun, Huihuan Qian The Chinese University Field Robot System
Zhefan Xu, Baihan Chen, Xiaoyang Zhan, Yumeng Xiu,Carnegie Mellon UniversField Robot System
Bryan Busby, Shifei Duan, Marcus Thompson, Minas LiThe University of Auckl Field Robotics and
Miguel Angel Pérez Cutiño, Jesus Capitan, José-migueUniversidad de Sevilla,UField Robotics and
Juan Heredia, Robin Kirschner, Christian Schlette, S University of Southern Field Robotics and
Alexander Moortgat-pick, Marie Schwahn, Anna AdamcTechnical University of Field Robotics and
Rohit Vishwajit Nanavati, Callum Rhodes, Matthew CooLoughborough UniversitField Robotics and
Kizito Masaba, Monika Roznere, Mingi Jeong, Alberto Qu Dartmouth College Field Robotics and
Felix Esser, Gereon Tombrink, André Cornelißen, LassUniversity of Bonn Field Robotics and
Nathaniel Bristow, Nikolas Pardoe, Jiarong Hong University of Minnesot Field Robotics and
Mateus Valverde Gasparino, Arun Narenthiran SivakumUniversity of Illinois a Field Robotics and
Wen Fan, Haoran Li, Weiyong Si, Shan Luo, Nathan L University of Bristol,Un Force and Tactile S
Wendong Zheng, Kun Liu, Di Guo, Wuqiang Yang, Jun Tsinghua University,Ins Force and Tactile S
George Brayshaw, Benjamin Ward-Cherrier, Martin Pe University of Bristol,Bri Force and Tactile S
Yuni Fuchioka, Masashi Hamaya University of British Force and Tactile S
Max Yang, Yijiong Lin, Alex Church, John Lloyd, Dand University of Bristol,Ca Force and Tactile S
Nural Yilmaz, Ugur Tumerdem Marmara University Force and Tactile S
Wanlin Li, Meng Wang, Jiarui Li, Yao Su, Devesh Jha, Beijing Institute for General Artificial
Force Intelligence
and Tactile S (BIGAI),Bei
Yuxiang Ma, Alan Zhao, Edward Adelson Massachusetts InstituteForce and Tactile S
Megha Tippur, Edward Adelson Massachusetts InstituteForce and Tactile S
Yiyue Luo, Murphy Wonsick, Jessica Hodgins, Brian O Massachusetts InstituteForce and Tactile S
Osher Azulay, Nimrod Curtis, Rotem Sokolovsky, Guy Lev Tel Aviv University,Tel-AForce and Tactile S
Changyi Lin, Han Zhang, Jikai Xu, Lei Wu, Huazhe Xu Carnegie Mellon UniversForce and Tactile S
Zhonglin Lin, Jiaquan Zhuang, Yufeng Li, Xianyu Wu, Fuzhou University,FUZ Force and Tactile S
Etienne Roberge, Guillaume Fornes, Jean-Philippe Ro École de technologie Force and Tactile S
Eric T. Chang, Runsheng Wang, Peter Ballentine, JingxiColumbia University,NAForce and Tactile S
Giulia Pagnanelli, Simone Ciotti, Nathan Lepora, AntoniUniversity of Pisa,UniverForce and Tactile S
Yijiong Lin, Alex Church, Max Yang, Haoran Li, John University of Bristol,Ca Force and Tactile S
Amir Hossein Shahid Zadeh, Seong Jong Yoo, Pavan Ma University of Maryland,UForce and Tactile S
Xiao Liu, Hongjin Chen, Haoyao Chen Harbin Institute of Tec Force and Tactile Se
Yun Liu, Xiaomeng Xu, Weihang Chen, Haocheng Yuan, Tsinghua University,Stan Force and Tactile Se
Remko Proesmans, Andreas Verleysen, Francis wyffel Ghent University Force and Tactile Se
Jing-chen Peng, Shaoxiong Yao, Kris Hauser University of Illinois @ Force and Tactile Se
Wenjie Lai, Jiajun Liu, Bing Rui Sim, Ming Rui Joel Nanyang Technological Force
U and Tactile Se
Ziwu Song, Ran Yu, Xuan Zhang, Kit Wa Sou, Shilong Tsinghua University,SheForce and Tactile Se
Bo Ying Su, Yuchen Wu, Chengtao Wen, Changliu Liu Carnegie Mellon Univer Force and Tactile Se
Boxin Xu, Luoyan Zhong, Grace Zhang, Xiaoyu Liang, Cornell University Force and Tactile Se
Osher Azulay, Alon Mizrahi, Nimrod Curtis, Avishai Sin Tel Aviv University,Tel-AForce and Tactile Se
Zhiyuan Zhang, Huan Ma, Yulin Zhou, Jingjing Ji, Hua Huazhong University of Force and Tactile S
Won Dong Kim, Sanghoon Yang, Woojong Kim, Jeong-Korea Advanced InstitutForce and Tactile S
Antonia Bronars, Sangwoon Kim, Parag Patre, Alberto MIT,Massachusetts InstiForce and Tactile S
Xinran Liu, Hao Chen, Hongbin Liu university of chinese a Force and Tactile S
Dane Brouwer, Joshua Citron, Hojung Choi, Marion LepStanford University Force and Tactile S
Funing Hou, Gang Li, Chenxing Mu, Mengqi Shi, JixiaoFudan University,Hebei Force and Tactile S
Fan Zhang, Yiannis Demiris Honda Research InstituForce and Tactile S
Vedant Dave, Fotios Lygerakis, Elmar Rueckert Montanuniversität LeobeForce and Tactile S
Shuo Jiang, Lawson L.S. Wong Northeastern UniversityForce and Tactile S
Fangxun Zhong, Bin Li, Wei Chen, Yunhui Liu The Chinese University Force Control and
Woosik Lee, Chuchu Chen, Guoquan Huang University of Delaware Force Control and
Hanwen Zhang, Zeyu Lu, Wenyu Liang, Haoyong Yu, Institute of optics and Force Control and
Takamasa Kawahara, Toshiaki Tsuji Saitama University Force Control and
Guanglin Cao, Mingcong Chen, Jian Hu, Hongbin Liu Institute of Automation Force Control and
Zhengguo Zhu, Guoteng Zhang, Zhongkai Sun, Teng Ch Shandong University,Zhe Force Control and
Yanran Wei, Shangke Lyu, Wenshuo Li, Xiang Yu, Lei Beihang University,NanyForce Control and
Kübra Karacan, Robin Kirschner, Hamid Sadeghian, F Technical University of Formal Methods in
David Gundana, Hadas Kress-Gazit Cornell University Formal Methods in
Qi Heng Ho, Zachary Sunberg, Morteza Lahijanian University of Colorado BFormal Methods in
Lipsy Gupta, Jahid Chowdhury Choton, Pavithra PrabhKansas State UniversityFormal Methods in
Yuanfei Lin, Haoxuan Li, Matthias Althoff Technical University of Formal Methods in
Noémie Jaquier, Leonel Rozo, Tamim Asfour Karlsruhe Institute of TeFormal Methods in
Disha Kamale, Cristian Ioan Vasile Lehigh University Formal Methods in
Kaier Liang, Gustavo A. Cardona, Cristian Ioan Vasile Lehigh University Formal Methods in
Sai Shankar Narasimhan, Sharachchandra Bhat, SandeThe University of Texas Formal Methods in
Alexis Linard, Anna Gautier, Daniel Duberg, Jana Tum KTH Royal Institute of Formal Methods in
Hao Zhang, Hao Wang, Zhen Kan University of Science a Formal Methods in
Karan Muvvala, Andrew Wells, Morteza Lahijanian, LydUniversity of Colorado BFormal Methods in
Ziyang Chen, Zhangli Zhou, Lin Li, Zhen Kan University of Science a Formal Methods in
Lin Li, Ziyang Chen, Hao Wang, Zhen Kan University of Science a Formal Methods in
Xinyu Liu, Ankit Jayesh Shah, Eric Rosen, Mingxi Jia, GBrown University,Massac Formal Methods in
Genichiro Tanaka, Yoshinobu Takahashi, Hiroyasu IwaWaseda university,WaseFormal Methods in
Andrew Benton, Eugen Solowjow, Prithvi Akella Siemens,Siemens Corpora Formal Methods in
Dylan Shell, Jason O'kane Texas A&M University Formal Methods in
Alessandro Palleschi, Franco Angelini, Chiara GabellierFloating Robotics AG,Uni Grasping I
Shoujie Li, Haixin Yu, Wenbo Ding, Houde Liu, Linqi Tsinghua Shenzhen Inter Grasping I
Jérôme Bastien, Lionel Birglen Polytechnique Montréal,Grasping I
Ahmet Tekden, Marc Peter Deisenroth, Yasemin BekiroChalmers University of Grasping I
Michael Noseworthy, Seiji Shaw, Chad C. Kessens, Ni Massachusetts InstituteGrasping I
Haonan Duan, Yiming Li, Daheng Li, Wei Wei, Yayu H Institute of Automation Grasping I
Hanzhi Chen, Binbin Xu, Stefan Leutenegger Technical University of Grasping I
Naif Alharthi, Martim Brandao King's College London Grasping I
Chaitanya Mitash, Mostafa Hussein, Jeroen Vanbaar, ViAmazon Robotics,UnviriGrasping I
Jiayuan Luo, Yaxin Liu, Han Wang, Mengyuan Ding, X Xi'an Jiaotong Universit Grasping II
S. Vishnu Rajendran, Willow Mandil, Kiyanoush Nazar University of Lincoln,UniGrasping II
Johann Huber, François Hélénon, Hippolyte Christian ISIR, Sorbonne UniversitGrasping II
Julia Starke, Tamim Asfour Karlsruhe Institute of T Grasping II
Pedro Piacenza, Jiacheng Yuan, Jinwook Huh, Volkan S I amsung Research Amer Grasping II
Malcolm Tisdale, Joel Burdick The California Institute Grasping II
René Zurbrügg, Yifan Liu, Francis Engelmann, Suryansh ETH Zürich,ETH Zurich Grasping II
Kyle Dufrene, Keegan Nave, Josh Campbell, Ravi Bal Oregon State UniversityGrasping II
Huihan Liu, Shivin Dass, Roberto Martín-martín, Yuke University of Texas, AusGrasping II
Yu Cui, Qi Ye, Qingtao Liu, Anjun Chen, Gaofeng Li, J Zhejiang University Grasping III
Patrick Lancaster, Nicklas Hansen, Aravind Rajeswara Meta AI,University of CaGrasping III
Ho Jin Choi, Nadia Figueroa University of Pennsylva Grasping III
Hengxu Yan, Hao-shu Fang, Cewu Lu Shanghai Jiao Tong Univ Grasping III
Fuqiang Zhao, Dzmitry Tsetserukou, Qian Liu Dalian University of Te Grasping III
Jungpyo Lee, Zheng Sun, Zhipeng Dong, Fei Chen, HaUniversity of Californi Grasping III
Diego Xavier Hidalgo Carvajal, Hanzhi Chen, Gemma Car Technical University of Grasping III
Brice Denoun, Miles Hansard, Beatriz Leon, Lorenzo The Shadow Robot Comp Grasping III
Digby Chappell, Fernando Bello, Petar Kormushev, NicImperial College LondonGrasping III
Toshihiro Nishimura, Tetsuyou Watanabe Kanazawa University Grippers and Other
Yi-shian Tsai, Pin-chun Yeh, Chun-hung Huang, I-che National Cheng Kung Un Grippers and Other
Chih-Hsing Liu, Sy-yeu Yang, Yi-chieh Shih National Cheng Kung Un Grippers and Other
Frederik Ostyn, Bram Vanderborght, Guillaume CrevecGhent University,VUB Grippers and Other
Connor Yako, Jerome Nowak, Shenli Yuan, Kenneth SaStanford University,SRI Grippers and Other
Xiangbo Wang, Tianran Zhang, Hongze Yu, Zhenwei Wen College of Quality and TGrippers and Other
Mengde Li, Fuqiang Zhao, Xiangli Li, Mingchang Li, SheThe Institute of Techno Grippers and Other
Philip Gümbel, Klaus Dröder Technische Universität Grippers and Other
Malte Hansjosten, Jan Baumgärtner, Jürgen Fleischer Karlsruhe Institute of T Grippers and Other
Joel Burdick, Malcolm Tisdale California Institute of T Grippers and Other
Leonardo Franco, Enrico Turco, Valerio Bo, Maria PozziUniversity of Siena,Istit Grippers and Other
Apoorv Vaish, Oliver Brock TU Berlin,Technische Uni Grippers and Other
Stanley Wang, Duyi Kuang, Sebastian Lee, Robert FullUniversity of California Grippers and Other
Austin Allison, Nathaniel Hanson, Sebastian Wicke, TasNortheastern UniversityGrippers and Other
Raunaq Mahesh Bhirangi, Abigail Defranco, Jacob AdkiCarnegie Mellon UniversGrippers and Other
Koki Yamane, Yuki Saigusa, Sho Sakaino, Toshiaki TsuUniversity of Tsukuba,SGrippers and Other
Amar Hajj-ahmad, Lukas Kaul, Carolyn Chen, Mark CuStanford University,ToyoGrippers and Other
Xueyang Yao, Yinghan Chen, Bryan Patrick Tripp University of Waterloo Grippers and Other
Jiarui Hu, Yanmin Zhou, Zhipeng Wang, Xin Li, YongkaTongji University Haptics and Haptic
Yuan Su, Gaofeng Li, Yongsheng Deng, Ioannis SarakoNortheastern University,Haptics and Haptic
Alessia Silvia Ivani, Federica Barontini, Manuel GiuseppFondazione Istituto Itali Haptics and Haptic
Xiaosa Li, Runze Zhao, Chengyue Lu, Xiao Xiao, WenbTsinghua University,Tsi Haptics and Haptic
Alberto Rota, Ke Fan, Elena De Momi Politecnico di Milano Haptics and Haptic
Ayush Giri, Robert Bloom, Tania K. Morimoto University of California Haptics and Haptic
Jace Rozsa, Sarah Costrell, Melisa Orta Martinez, Gar Carnegie Mellon UniversHaptics and Haptic
Jake Ketchum, Ahalya Prabhakar, Todd Murphey Northwestern University,Haptics and Haptic
Wing-sum Adrienne Law, Sofia Wyetzner, Raymond ZhStanford University Haptics and Haptic
Han Gao, Yating Liu, Fang Cao, Hao Wu, Fengyuan X National Key Lab for NoHuman Factors and
Lukas Cha, Annika Guez, Chih-yu Chen, Sion Kim, ZheTechnical University of Human Factors and
Beril Yalçinkaya, Micael Couceiro, Lucas Pina, Salvia Ingeniarius Lda,UniversHuman Factors and
Roberto Meattini, Alessio Caporali, Alessandra BernardinUniversity of Bologna Human Factors and
Marta Lagomarsino, Marta Lorenzini, Merryn Dale ConsIstituto Italiano di Tec Human Factors and
Qiuhao Li, Shenghai Yuan Northeastern Univer Human Factors and
Siddharth Singi, Zhanpeng He, Alvin Pan, SandipkumarColumbia University,A Human Factors and
Siya Kunde, Brittany Duncan University of Nebraska,UHuman Factors and
Aya Yaacoub, Vincent Thomas, Francis Colas, PaulineLORIA-CNRS,LORIA - Uni Human Factors and
Daniel Marta, Simon Holk, Christian Pek, Iolanda Leite KTH Royal Institute of THuman Factors and
Ifueko Igbinedion, Sertac Karaman Massachusetts InstituteHuman Factors and
Kenta Mukoya, Erica Weng, Rohan Choudhury, Kris KitCarnegie Mellon UniversHuman Factors and
Heng Tan, Yukun Yuan, Hua Yan, Shuxin Zhong, Yu YLehigh University,Univ Human Factors and
Yuhang Zhong, Guangyu Zhao, Qianhao Wang, GuangtZhejiang Unviersity,ZhejHuman Factors and
Emir Mobedi, Sebastian Hjorth, Wansoo Kim, Elena DeIstituto Italiano di Tecn Human Factors and
Junnan Huang, Xuefeng Wang, Chongkun Xia, Houde Li Tsinghua University,PekHuman Factors and
Sebastian Hjorth, Johannes Lachner, Arash Ajoudani, Istituto italiano di Tech Human Factors and
Malte Riechmann, André Kirsch, Matthias König, Jan Re Bielefeld University of Human-Aware Motio
Rashmi Bhaskara, Hrishikesh Viswanath, Aniket Bera Purdue University Human-Aware Motio
Sara Pohland, Alvin Tan, Prabal Dutta, Claire Tomlin University of California Human-Aware Motio
Aamodh Suresh, Angelique Taylor, Laurel Riek, Sonia US Army Research Labor Human-Aware Motio
Songyang Han, Apaar Sadhwani, Tushar Agarwal, Hamid Sony AI,Stanford UniverHuman-Aware Motio
Hojun Chung, Jeongwoo Oh, Jae Seok Heo, Gunmin LSeoul National UniversitHuman-Aware Motio
Yuxiao Chen, Sushant Veer, Peter Karkus, Marco Pav Nvidia research,NVIDIA,Human-Aware Motio
Qianyi Zhang, Shichao Wu, Yuhang Jia, Yuang Xu, JingNankai University Human-Aware Motio
Qishuai Liu, Martim Brandao University of Nebraska- Human-Aware Motio
Xiwen Dengxiong, Xueting Wang, Shi Bai, Yunbo ZhanRochester Institute of Human-Robot Collab
Caiyue Xu, Changming Zhang, Yanmin Zhou, Zhipeng W Tongji University Human-Robot Collab
Behnam Khojasteh, Friedrich Solowjow, Sebastian TrimMax Planck Institute fo Human-Robot Collab
Elena Vella, Airlie Chapman, Nir Lipovetzky The University of MelboHuman-Robot Collab
Cedric Goubard, Yiannis Demiris Imperial College LondonHuman-Robot Collab
Zhen Zhang, Anran Lin, Chun Wai Wong, Xiangyu Chu,The Chinese University Human-Robot Collab
Eion Tyacke, Kunal Gupta, Jay Patel, Raghav Katoch, New York University,NeHuman-Robot Collab
Mukund Mitra, Gyanig Kumar, Partha Pratim Chakrabart IISc Bangalore,Indian Ins
Human-Robot Collab
Chuang Yu, Baris Serhan, Angelo Cangelosi University College LondHuman-Robot Collab
Chuanxiong Zheng, Lei Zhang, Hui Wang, Randy Gomez, ocean university of chi Human-Robot Collab
Zhenyu Liao, Jose Victorio Salazar Luces, Ankit RavanTohoku University Human-Robot Collab
Tadashi Sumioka, Kazushi Akimoto, Takuya Tsujimura Honda R&D Co., Ltd.,Hon Human-Robot Collab
Ruolin Ye, Yifei Hu, Yuhan (anjelica) Bian, Luke Kulm Cornell University Human-Robot Collab
John Atkins, Hyunglae Lee Arizona State UniversityHuman-Robot Collab
Hideki Deguchi, Kazuki Shibata, Shun Taguchi Toyota Central R&D labs Human-Robot Collab
Hunter Ray, Zakariya Laouar, Zachary Sunberg, Nisar University of Colorado BHuman-Robot Collab
Sunwoong Choi, Zaid Al-Sabbag, Sriram Narasimhan, University of California Human-Robot Collab
Warnakulasuriya Chandima Fernando, Stuart Campbell,Brookhaven
D National LaHuman-Robot Collab
Kasper Hald, Matthias Rehm Aalborg University Human-Robot Collabo
Suzanne Oliver, Peter Paik, Xingyuan Zhou, S. Farokh New York University,NeHuman-Robot Collabo
Weihan Chen, Hanwen Ren, Ahmed H. Qureshi Purdue University Human-Robot Collabo
Hannah Dinovitzer, Mohammad Shushtari, Arash AramiUniversity of Waterloo Human-Robot Collabo
Nithesh Kumar, Hsin-ming Chao, Bruno Dantas Da Silva Clemson,Cornell UniversHuman-Robot Collabo
Matthias Rehm, Ioannis Pontikis, Kasper Hald Aalborg University,A Human-Robot Collabo
Jianwei Peng, Zhelin Liao, Zefan Su, Hanchen Yao, Y University of Chinese A Human-Robot Collabo
Anne Van Der Horst, Bastiaan Guillermo Lorenzo Meere Eindhoven University ofHuman-Robot Collabo
Pian Yu, Shuyang Dong, Shili Sheng, Lu Feng, Marta University of Oxford,Univ Human-Robot Collab
Ravi Pandya, Zhuoyuan Wang, Yorie Nakahira, ChangliCarnegie Mellon Univer Human-Robot Collab
Yuening Zhang, Paul Robertson, Tianmin Shu, Sungkwe Massachusetts InstituteHuman-Robot Collab
Shichen Cao, Jing Xiao Worcester Polytechnic In Human-Robot Collab
Malak Sayour, Mohammad Karim Yassine, Nadim Dib, AmericanIm University of BHuman-Robot Collab
Matija Mavsar, Jun Morimoto, Aleš Ude Jozef Stefan Institute, Human-Robot Collab
Aayush Jain, Philip Long, Valeria Villani, John D. Kelle Irish Manufacturing Rese Human-Robot Collab
Andrea Pupa, Marco Minelli, Cristian Secchi University of Modena anHuman-Robot Collab
Zhaoyang Jacopo Hu, Ziwei Wang, Yanpei Huang, AranImperial College LondonHuman-Robot Collab
Michael Fulton, Junaed Sattar, Rafa Absar University of Minnesota,Human-Robot Collab
Shail Jadav, Johannes Heidersberger, Christian Ott, D IIT Gandhinagar,TechniHuman-Robot Collab
Takayuki Osa, Tatsuya Harada University of Tokyo,The Human-Robot Collab
Longrui Chen, Zhaoyang Jacopo Hu, Yanpei Huang, EtImperial college LondonHuman-Robot Collab
Lihan Zha, Yuchen Cui, Li-heng Lin, Minae Kwon, MontStanford University,Go Human-Robot Collab
Fan Shao, Fanny Ficuciello University of Naples Fede Human-Robot Collab
Zhiwei Liao, Marta Lorenzini, Mattia Leonori, Fei Zhao Xi'an Jiaotong Universit Human-Robot Collab
Haoyi Zheng, Zhaoyang Jacopo Hu, Yanpei Huang, XiaImperial College LondonHuman-Robot Collab
Keyhan Kouhkiloui Babarahmati, Mohammadreza Kasaei, University of Edinburgh,Human-Robot Collab
Silvia Izquierdo-badiola, Gerard Canal, Carlos Rizzo, GEurecat,King's College Human-Robot Collab
Ravi Pandya, Michelle Zhao, Changliu Liu, Reid Simm Carnegie Mellon UniversHuman-Robot Collab
Andrea Pupa, Cristian Secchi University of Modena anHuman-Robot Collab
Kaige Shi, Guoqiang Hu Nanyang Technological Human-Robot
U Collab
Andrea Pupa, Marco Minelli, Cristian Secchi University of Modena anHuman-Robot Collab
Yuxing Long, Xiaoqi Li, Wenzhe Cai, Hao Dong Peking University,SouthHuman-Robot Collab
Alice Nardelli, Antonio Sgorbissa, Carmine Tommaso RUniversity of Genoa,UniHuman-Robot Collab
Simon Holk, Daniel Marta, Iolanda Leite KTH Royal Institute of Human-Robot Intera
Peter Schaldenbrand, Gaurav Parmar, Jun-yan Zhu, Carnegie Mellon UniversHuman-Robot Intera
Junwei Zheng, Jiaming Zhang, Kailun Yang, Kunyu Peng Karlsruhe Institute of T Human-Robot Intera
Pietro Balatti, Idil Özdamar, Doganay Sirintuna, Luca FIstituto Italiano di Tecn Human-Robot Intera
Tilman Daab, Noémie Jaquier, Christian R. G. Dreher, Karlsruhe Institute of T Human-Robot Intera
Erik Ballesteros, Sang-Yoep Lee, Kalind Carpenter, HaMassachusetts InstituteHuman-Robot Intera
Kavya Puthuveetil, Sasha Wald, Atharva Pusalkar, PratCarnegie Mellon Univer Human-Robot Interac
Matthew Luebbers, Aaquib Tabrez, Kanaka Samagna Ta University of Colorado Human-Robot Interac
Yufei Zhu, Han Fan, Andrey Rudenko, Martin Magnusson, Örebro University,RobeHuman-Robot Interac
Ali Ayub, Chrystopher L. Nehaniv, Kerstin Dautenhahn University of Waterloo Human-Robot Interac
Fei Gao, Dai Lingna, Jingjie Zhu, Mei Du, Zhang Yiyu Xidian University,AiSke Human-Robot Interac
Alexander Thomas, Jianan Chen, Anna Hella-szabo, Mer University College LondHuman-Robot Interac
Yan Wang, Shuli Lou, Kai Wang, Xiaohu Yuan, HuapingYantai University,YantaiHuman-Robot Interac
Hao Jin, Minghui Lian, Shicheng Qiu, Xuxu Han, Xizhi Northwest A&F UniversitHuman-Robot Interac
Jiaqi Chen, Boyang Sun, Hermann Blum, Marc PollefeyETH Zurich Human-Robot Interac
Cheng-yu Kuo, Hirofumi Shin, Takamitsu Matsubara Nara Institute of Scien Humanoid and Bipe
Xunjie Chen, Anikode Aditya, Jingang Yi, Tao Liu Rutgers University,ZhejiHumanoid and Bipe
lishen pu, Yixuan Liu, Zheng Aiqun, Bofeng Qi, Chunq Tongji University,ShangHumanoid and Bipe
Annan Tang, Takuma Hiraoka, Naoki Hiraoka, Fan Shi,The University of TokyoHumanoid and Bipe
Xiang Meng, Zhangguo Yu, Xuechao Chen, Zelin HuanBeijing Institute of Tech Humanoid and Bipe
Fengyi Wang, Julio Rogelio Guadarrama-Olvera, Nitis Technical University of Humanoid and Bipe
Margaret Eva Mungai, J.W Grizzle, Gokul PrabhakaranUniversity of Michigan,UHumanoid and Bipe
Juncai Long, Jituo Li, Yiwen Lu, Chengdi Zhou, Guodo ZheJiang University,ZhejHumanoid and Bipe
Ines Sorrentino, Giulio Romualdi, Daniele Pucci Istituto Italiano di TecnoHumanoid and Bipe
Jinlong Li, Runsheng Xu, Xinyu Liu, Baolu Li, Qin Zou, cleveland state universi Image-Based Naviga
Jiahui Fu, Chen Gao, Zitian Wang, Lirong Yang, XiaofeBeihang University,MeitImage-Based Naviga
Zhe Wang, Siqi Fan, Xiaoliang Huo, Tongda Xu, Yan Wa Institute for AI Industr Image-Based Naviga
Marvin Klemp, Royden Wagner, Kevin Rösch, Martin Laue Karlsruhe Institute of T Image-Based Naviga
Zihua Liu, Yizhou Li, Masatoshi Okutomi Tokyo Institute of Tech Image-Based Naviga
Jongmin Yu, Chi Bene Chen, Sebastiano Fichera, PaolKing's College London,Un Image-Based Naviga
Yuhang Liu, Sun Boyi, Yuke Li, Yuzheng Hu, Feiyue WChinese Academy of Scie Image-Based Naviga
Jiawei Hou, Xiaoyan Li, Wenhao Guan, Gang Zhang, DFudan University,Chine Image-Based Naviga
Dongqiangzi Ye, Yufei Xie, Weijia Chen, Zixiang Zhou, Tusimple,N/A,UniversityImage-Based Naviga
Yupeng Zheng, Xiang Li, Pengfei Li, Yuhang Zheng, BuSchool of Artificial In Image-Based Naviga
Tara Sadjadpour, Jie Li, Rares Ambrus, Jeannette BohStanford University,ToyoImage-Based Naviga
Jinlong Li, Baolu Li, Xinyu Liu, Runsheng Xu, Jiaqi Ma cleveland state universi Image-Based Naviga
Jinlong Li, Baolu Li, Xinyu Liu, Jianwu Fang, Felix Jue cleveland state univers Image-Based Naviga
Yihong Xu, Loick Chambon, Eloi Zablocki, Mickaël Chen Valeo.ai,Valeo,valeo,EPImage-Based Naviga
Siqi Fan, Haibao Yu, Wenxian Yang, Jirui Yuan, Zaiqin Tsinghua University,TheImage-Based Naviga
Jie Zhan, Yucong Duan, Junfeng Ding, Xuzhong Hu, XiHuazhong University of Image-Based Naviga
Linyi Zhou, Zhongxue Gan, Jiayuan Fan Fudan University Image-Based Naviga
Hsu-kuang Chiu, Chien-yi Wang, Min-hung Chen, StephCarnegie Mellon Univer Image-Based Naviga
Mathias Lechner, Ramin Hasani, Alexander Amini, Ts Massachusetts InstituteImitation Learning
Zhenyang Lin, Yurou Chen, Zhiyong Liu University of Chinese Imitation Learning
Yang Liu, Marius Hofert The University of Hong Imitation Learning
Norman Di Palo, Edward Johns Imperial College LondonImitation Learning
Daniel Yang, Davin Tjia, Jacob Herman Berg, Dima DaMassachusetts Institute Imitation Learning
Evan Ellis, Gaurav Ghosal, Stuart Jonathan Russell, A UC Berkeley,Carnegie Mel Imitation Learning
Tianyu Fu, Yunfeng Bai, Cheng Li, Fengming Li, Chao Shandong University,ShImitation Learning
Sergio Paniego Blanco, Nikhil Paliwal, José M. Cañas Universidad Rey Juan CImitation Learning
Yifei Dong, Florian T. Pokorny KTH,KTH Royal Institut In-Hand Manipulati
Hinano Ichikura, Mitsuru Higashimori Osaka University In-Hand Manipulati
Guotao Li, Xu Liang, Guotao Li, Tingting Su, Zhijie Li Institute of Automation In-Hand Manipulati
Lingfeng Tao, Jiucai Zhang, Qiaojie Zheng, Xiaoli ZhanOklahoma State UniversIn-Hand Manipulati
Karl Van Wyk, Ankur Handa, Viktor Makoviichuk, Yijie Gu NVIDIA,NVidia,UniversitIn-Hand Manipulati
Boyuan Liang, Kei Ota, Masayoshi Tomizuka, Devesh University of California In-Hand Manipulati
Gagan Khandate, Cameron Mehlman, Xingsheng Wei, ColumbiaMa University In-Hand Manipulati
Ying Yuan, Haichuan Che, Yuzhe Qin, Binghao Huang,Tsinghua University,UnivIn-Hand Manipulati
Farshad Khadivar, Aude G. Billard EPFL In-Hand Manipulati
Abdur Rosyid, Bashar El-khasawneh Khalifa University Industrial Robotics
Lei Zhang, Kaixin Bai, Qiang Li, Zhaopeng Chen, Jian University of Hamburg, Industrial Robotics
Michael Schwegel, Andreas Kugi TU Wien Industrial Robotics
Benjamin Alt, Florian Stöckl, Silvan Müller, Christop ArtiMinds Robotics,DHBW Industrial Robotics
Ruiqing Chen, Wenbin Song, Weiqin Zu, Zixin Dong, Z ShanghaiTech University Industrial Robotics
Miao Zhang, Yiqing Shen, Shenghui Zhong, Guofeng PTsinghua Shenzhen Inter Industrial Robotics
Hung-ching Lin, Murilo Marinho, Kanako Harada University of Tokyo,TheIndustrial Robotics
Yohan John, Connor Hughes, Gilberto Diaz-garcia, JasUC Santa Barbara,Univer Industrial Robotics
Carlos Suarez Zapico, Y. R. Petillot, Mustafa Suphi Er Edinburgh Centre for RoIndustrial Robotics
Adam Polevoy, Marin Kobilarov, Joseph Moore Johns Hopkins UniversitIntegrated Planning
Sourav Biswas, Sergio Casas Romero, Quin Sykora, BWaabi, University of To Integrated Planning
Inkyu Jang, Sunwoo Hwang, Jeonghyun Byun, H. Jin KSeoul National UniversitIntegrated Planning
Jiaping Ren, Jiahao Xiang, Hongfei Gao, Jinchuan ZhaInceptio Technology,TonIntegrated Planning
Shuhao Zhang, Mathias Bos, Bastiaan Vandewal, WilmKU D Leuven,Katholieke Un Integrated Planning
Mohamed-khalil Bouzidi, Yue Yao, Joerg Reichardt, DaContinental, FU Berlin,FIntegrated Planning
Zikang Xiong, Daniel Lawson, Joe Kurian Eappen, AhmPurdue University Integrated Planning
Leon Yan, Santosh Devasia University of WashingtoIntegrated Planning
Anuj Pasricha, Alessandro Roncone University of Colorado Integrated Planning
Panagiotis Rousseas, Charalampos Bechlioulis, KostasNational Technical UniveIntegrated Planning
Alex Lagrassa, Moonyoung Lee, Oliver Kroemer Carnegie Mellon UniversIntegrated Planning
Lukas Lao Beyer, Gilhyun Ryou, Patrick Spieler, Sert Massachusetts InstituteIntegrated Planning
Jonathan Styrud, Matthias Mayr, Erik Hellsten, Volker ABB,Lund University,KTH Integrated Planning
Dibyendu Das, Yuanjie Lu, Erion Plaku, Xuesu Xiao George Mason UniversitIntegrated Planning
Jiayu Guo, Mingyue Feng, Pengfei Zhu, Jinsheng Dou, Fudan University,Mogo IAntegrated Planning
Yinghan Wang, Jianping He, Xiaoming Duan Shanghai Jiao Tong Univ Integrated Planning
Zixuan Huang, Yating Lin, Fan Yang, Dmitry Berenson University of Michigan Integrated Planning
Marco Faroni, Dmitry Berenson Politecnico di Milano,UnIntegrated Planning
Rasmus Haugaard, Yitaek Kim, Thorbjørn Mosekjær IvUniversity of Southern Intelligent and Flex
Victor Vantilborgh, Tom Lefebvre, Guillaume Crevecoe Ghent University Intelligent and Flex
Peter So, Andriy Sarabakha, Fan Wu, Utku Culha, Far Technical University of Intelligent and Flex
Sadeq Yaqubi, Seyedmohammad Tahamipourzarandi, Tampere Jo University Intelligent and Flex
Subyeong Ku, HyunWoong Choi, Ho-young Kim, Yong-Seoul National UniversitIntelligent and Flex
Stefan Schneyer, Arne Sachtler, Thomas Eiband, KorbiGerman Aerospace Cente Intelligent and Flex
Kshitij Bhatta, Qing Chang University of Virginia Intelligent and Flex
Florian Mirus, Frederik Pasch, Kay-ulrich Scholl Intel Labs,Intel Intelligent and Flex
Yan Ding, Yijun Mao, Chongshan Jiao, Pengju Ren Xi'an Jiaotong Universit Intelligent Transpor
Tuong Do, Binh Nguyen, Quang Tran, Hien Nguyen, ErAIOZ,National Tsing Hua Intelligent Transpor
Yun Tang, Dhanush Raj, Xingyu Zhao, Xizhe Zhang, Ant University of Warwick,WIntelligent Transpor
Cameron Hickert, Sirui Li, Cathy Wu Massachusetts InstituteIntelligent Transpor
Nathan Ludlow, Yiwei Lyu, John Dolan Brigham Young Universit Intelligent Transpor
Rezaul Karim, Soheil Mohamad Alizadeh Shabestary, AYork University,Huawei Intelligent Transpor
Zhenmin Huang, Haichao Liu, Shaojie Shen, Jun Ma The Hong Kong UniversiIntelligent Transpor
Wei-jer Chang, Chen Tang, Chenran Li, Yeping Hu, M University of California, Intelligent Transpor
Junyu Diao, Renzhi Tang, Yi Gu, Sen Tian, Zhihao Jia Shanghaitech Universit Intelligent Transpor
Maria Lyssenko, Piyush Pimplikar, Maarten Bieshaar, FRobert Bosch GmbH, Un Intelligent Transpor
Xian-feng Han, Huixian Cheng, Hang Jiang, Dehong HeSouthwest University Intelligent Transpor
Longlong Jing, Ruichi Yu, Xu Chen, Zhengli Zhao, Shi Waymo,UCI,Johns Hopki Intelligent Transpor
Yuang Zhang, Haonan An, Zhengru Fang, Guowen Xu,Tsinghua University,NanIntelligent Transpor
Andrew Holliday, Gregory Dudek McGill University Intelligent Transpor
Yanhu Wang, Gan Yu, Wei Xie, Weidong Zhang, CarlosShanghai Jiao Tong Univ Intelligent Transpor
Fabio Maresca, Filippo Grazioli, Antonio Albanese, Vin NEC Laboratories EuropIntelligent Transpor
Zhijun Pan, Fangqiang Ding, Hantao Zhong, Chris Xia Royal College of Art,UniIntelligent Transpor
Michael Villarreal, Bibek Poudel, Jia Pan, Weizi Li University of TennesseeIntelligent Transpor
Yiqun Duan, Qiang Zhang, Renjing Xu University of Technolg Intelligent Transpor
Zhiyu Huang, Peter Karkus, Boris Ivanovic, Yuxiao Ch Nanyang Technological Intelligent
U Transpor
Vidyaa Krishnan Nivash, Ahmed H. Qureshi Purdue University Intelligent Transpor
Liting Sun, Rebecca Roelofs, Ben Caine, Khaled RefaatUniversity of Californ Intelligent Transpor
Truls Nyberg, Jonne Van Haastregt, Jana Tumova KTH Royal Institute of Intelligent Transpor
Vindula Jayawardana, Sirui Li, Cathy Wu, Yashar FaridMassachusetts Institut Intelligent Transpor
Daniel Garces, Sushmita Bhattacharya, Dimitri BertsekaHarvard University,MIT Intelligent Transpor
Haoyi Niu, Yizhou Xu, Xingjian Jiang, Jianming Hu Tsinghua University Intelligent Transpor
Yu Shen, Laura Zheng, Tianyi Zhou, Ming C. Lin University of Maryland,UIntelligent Transpor
Chen Feng, Hangning Zhou, Huadong Lin, Zhigang Zhan Hong Kong University ofIntelligent Transpo
Niklas Kochdumper, Stanley Bak Stony Brook University Intelligent Transpo
Peining Shen, Jianwu Fang, Hongkai Yu, Jianru Xue Chang'an University,Xian Intelligent Transpo
Chenran Li, Aolin Xu, Enna Sachdeva, Teruhisa Misu, University of Californi Intelligent Transpo
Jacob Knaup, Jovin D'sa, Behdad Chalaki, Tyler Naes, Georgia Institute of Te Intelligent Transpo
Daniel Bogdoll, Jing Qin, Moritz Nekolla, Ahmed AboueFZI Research Center forIntelligent Transpo
Yubin Wang, Yulin Li, Zengqi Peng, Hakim Ghazzai, J The Hong Kong UniversiIntelligent Transpo
Liao Haicheng, Shangqian Liu, Yong Kang Li, Zhennin University of Macau,UniIntelligent Transpo
Pei Xu, Jean-Bernard Hayet, Ioannis Karamouzas Stanford University,CI Intelligent Transpo
He Li, Shuxiang Guo, Dongdong Bu, Hanze Wang, MasBeijing Institute of Tec Intention Recogniti
Pinhao Song, Pengteng Li, Erwin Aertbelien, Renaud DKU Leuven,Shenzhen Un Intention Recogniti
Ho Jin Choi, Satyajeet Das, Shaoting Peng, Ruzena BaUniversity of PennsylvanIntention Recogniti
Yaqi Song, Tong Chen, Shigang Li, Jianfeng Li Southwest University,HiIntention Recogniti
Yung-chi Kung, Arthur King Zhang, Junmin Wang, Joy The University of Texas Intention Recogniti
Dmitriy Rivkin, Nikhil Rajiv Kakodkar, Francois Hoga None,McGill University,M Intention Recogniti
Long Wang, Xiaoling Li, Zhangyi Chen, Zhipeng Sun, JXi'an Jiaotong Universit Intention Recognition
Amir Rasouli Huawei Technologies CIntention Recogniti
Maleen Jayasuriya, Gibson Hu, Dinh Dang Khoa Le, KaUniversity of Technolog Intention Recogniti
Amir Firouzeh, Jongeun Lee, Hyunsoo Yang, Dongjun LEPFL,Seoul National Univ Joint Mechanism
Clemens Claudio Christoph, Amirhossein Kazemipour, ETH Zürich,ETH Zurich Joint Mechanism
Matthew Auer, Suhrud Parag Joglekar, Hyunglae Lee Arizona State UniversityJoint Mechanism
Noah Harmatz, Alina Zahra, Amir Abdelmalak, Shivam Rutgers University,Rutge Joint Mechanism
Yicheng Dai, Zuan Li, Xin Wang, Han Yuan Harbin Institute of Tec Joint Mechanism
Yanan Li, Weining Lu, Yu Liu, Deshan Meng, Xueqian Tsinghua University,Depa Joint Mechanism
Quentin Peyron, Jessica Burgner-kahrs Inria and CRIStAL UMRJointCN Mechanism
Zane Zake, Stephane Caro IRT Jules Verne,CNRS/Joint Mechanism
Daniel Pérez-suay, Yu Li, Hamid Sadeghian, AbdeldjallTechnical University of Joint Mechanism
Nikolas Jakob Wilhelm, Sami Haddadin, Rainer Burgkart Technical University o Kinematics
Jihong Yan, Xin Zhang, Kai Yang, Jie Zhao Harbin Institute of Tech Kinematics
Jonas Kiemel, Torsten Kroeger Karlsruhe Institute of T Kinematics
Yucheng Li, David H. Myszka, Andrew Murray University of Dayton Kinematics
Victor Montano, Nikhil V. Navkar, Aaron T. Becker University of Houston, Kinematics
Keerthi Sagar Somenedi Nageswara Rao, Stephane Caro Irish Manufacturing ResKinematics
Inbar Meir, Avital Bechar, Avishai Sintov Tel Aviv University,Agri Kinematics
Arno Verduyn, Maxim Vochten, Joris De Schutter KU Leuven Kinematics
Yi Xu, Gregory Chirikjian National University of S Kinematics
Harshit Khurana, Aude G. Billard EPFL Kinematics and Dy
Richard Black, Marco Caballero, Andromachi ChatzieleMicrosoft,Microsoft Re Kinematics and Dy
Jie Zhao, Guilin Yang, Haoyu Shi, Silu Chen, Chin-Yin Chinese Academy of Scie Kinematics and Dy
Sébastien Briot, Jean-Pierre Merlet LS,N,INRIA Kinematics and Dy
Adolfo Suarez Roos, Zane Zake, Tahir Rasheed, Nico IRT Jules Verne,IRCCy Kinematics and Dy
Ngo Foon Chan, Wai Yi Lam, Darwin Lau The Chinese University Kinematics and Dy
Grama Srinivas Shourie Grama Srinivas Shouri, Mahdi Deutsches Forschungszen Kinematics and Dy
Feng Han, Jingang Yi Rutgers University Kinematics and Dy
Haotian Zheng, Haitong Ma, Sifa Zheng, Shengbo Li, Tsinghua University,HarLearning
Alaa Eddine Chriat, Chuangchuang Sun Mississippi State UniverLearning
Abdullah Cihan Ak, Eren Erdal Aksoy, Sanem Sariel Istanbul Technical UniveLearning
Moritz A. Graule, Volkan Isler Harvard University,Unive Learning
Hideyuki Ichiwara, Hiroshi Ito, Kenjiro Yamamoto, Hiro Hitachi, Ltd. / Waseda ULearning
Kento Kawaharazuka, Kei Okada, Masayuki Inaba The University of TokyoLearning
Maximilian Diehl, Karinne Ramirez-Amaro Chalmers University of Learning
Bryon Tjanaka, Matthew Fontaine, David H. Lee, AnirudUniversity of Souther Learning
Kento Kawaharazuka, Kei Okada, Masayuki Inaba The University of TokyoLearning
Rodrigo Pérez-Dattari, Jens Kober Delft University of Tech Learning from Demon
Xuanhui Xu, You Mingyu, Zhou Hongjun, Zhifeng Qian,TongJi University,TongjiLearning from Demon
Francesco Di Felice, Salvatore D'avella, Alberto Remus,Mechanical Intelligence Learning from Demon
Keito Sugawara, Sho Sakaino, Toshiaki Tsuji Saitama University,UnivLearning from Demon
Rosa Enna Sophia Maessen, Joseph Micah Prendergast Delft University of Tech Learning from Demon
Marina Y Aoyama, Joao Moura, Namiko Saito, Sethu VThe University of EdinbuLearning from Demon
Fei Meng, Liangliang Chen, Han Ma, Jiankun Wang, The Chinese University Learning from Demon
Yue Meng, Sai Vemprala, Rogerio Bonatti, Chuchu FanMassachusetts InstituteLearning from Demon
Siddhant Sharma, Shreshth Tuli, Rohan Paul Indian Institute of Tech Learning from Demon
Muhammad Arifin, Yuta Kage, Yuchen Yang, AlexanderWaseda University Learning from Demon
Mara Levy, Nirat Saini, Abhinav Shrivastava University of Maryland, Learning from Demon
Artūras Straižys, Michael Burke, Subramanian RamamoUniversity of Edinburgh Learning from Demon
Freek Stulp, Adrià Colomé, Carme Torras DLR - Deutsches Zentrum Learning from Demon
Hongjie Fang, Hao-shu Fang, Yiming Wang, Jieji Ren, Shanghai Jiao Tong Univ Learning from Demon
Dimitris Papadimitriou, Daniel Brown UC Berkeley,University Learning from Demon
Weiming Zhi, Tianyi Zhang, Matthew Johnson-RobersoCarnegie Mellon UniversLearning from Demon
Jenny Wang, Octavian Donca, David Held Carnegie Mellon UniversLearning from Demon
Amin Abyaneh, Mariana Sosa Guzmán, Hsiu-chin Lin McGill University,Unive Learning from Demon
Martin Schonger, Hugo Tadashi Kussaba, Lingyun Chen, Technical University of Learning from Demon
Weiyao Wang, Gregory Hager The Johns Hopkins Unive Learning from Demon
Farhad Nawaz Savvas Sadiq Ali, Tianyu Li, Nikolai MatUniversity of Pennsylva Learning from Demon
Yu Zhang, Yongxiang Zou, Haoyu Zhang, Xiuze Xia, L University of Chinese Learning from Demon
Zhehao Jin, Weiyong Si, Andong Liu, Wen-an Zhang, Zhejiang University of TLearning from Demon
Jaehwi Jang, Minjae Song, Daehyung Park Korea Advanced Institu Learning from Demon
Ryota Takamido, Jun Ota Research into Artifacts Learning from Demon
Rohan Chitnis, Yingchen Xu, Bobak Hashemi, Lucas Leh Massachusetts InstituteLearning in Control
David Emukpere, Bingbing Wu, Julien Perez, Jean-micNaver Labs Europe,N Learning in Control
Jamie Hathaway, Alireza Rastegarpanah, Rustam StolkDepartment of Metallurgy Learning in Control
Jesse Zhang, Karl Pertsch, Jiahui Zhang, Joseph Lim University of Southern Learning in Control
Jiawei Zheng, Yonghong Song Xi'an Jiaotong Universit Learning in Control
Chong Zhang, Jiapeng Sheng, Tingguang Li, He ZhangTencent Robotics X,ShaLearning in Control
Jun Yamada, Marc Rigter, Jack Collins, Ingmar PosnerUniversity of Oxford,OxfLearning in Control
Andrea Bajcsy, Antonio Loquercio, Ashish Kumar, JitenCarnegie Mellon UniversLearning in Control
Xiao Ma, Wu-jun Li Nanjing University Learning in Control
Yuyou Zhang, Yaru Niu, Xingyu Liu, Ding Zhao Carnegie Mellon UniversLearning in Control
Nilaksh Nilaksh, Abhishek Ranjan, Shreenabh Agrawal,Indian Institue of Techn Learning in Control
Xin Yu, Yongkai Tian, Li Wang, Pu Feng, Wenjun Wu, Beihang University Learning in Control
Shashank Hegde, Zhehui Huang, Gaurav Sukhatme University of Southern CLearning in Control
Laura Smith, Yunhao Cao, Sergey Levine UC Berkeley Learning in Control
Donghyeon Kim, Glen Berseth, Mathew Schwartz, Jae Graduate School of Conv Learning in Control
Yanjiang Guo, Zheyuan Jiang, Yen-jen Wang, Jingyue Tsinghua university,Tsi Learning in Control
Matthew Sivaprakasam, Parv Maheshwari, Mateo Guama Carnegie Mellon UniversiLearning in Field R
Junjie Wen, Jinqiang Cui, Benyun Zhao, Bingxin Han, The Chinese University Learning in Field R
Chen Hua, Chunmao Jiang, Runxin Niu, Biao Yu, Hui Zh University of Science a Learning in Field R
Samuel Triest, David D. Fan, Sebastian Scherer, Ali Carnegie Mellon Universi Learning in Field R
Laijian Li, Yukai Ma, Kai Tang, Xiangrui Zhao, Chao ChZhejiang University,zhejLearning in Field R
Matthias Eder, Gerald Steinbauer-Wagner Graz University of TechLearning in Field R
Andong Yang, Wei Li, Yu Hu Institute of Computing Learning in Field R
Zhuo Xu, Rui Zhou, Yida Yin, Huidong Gao, MasayoshiUC Berkeley,University Learning of in Field R
Lingfeng Sun, Haichao Zhang, Wei Xu, Masayoshi TomUniversity of California, Learning in Graspin
Zhifeng Qian, You Mingyu, Zhou Hongjun, Xuanhui Xu,Tongji University,Tongji Learning in Graspin
Alireza Rezazadeh, Changhyun Choi University of Minnesota,Learning in Graspin
Eleftherios Triantafyllidis, Filippos Christianos, Zhibin LiThe University of Edinb Learning in Graspin
Naoki Morihira, Pranav Deo, Manoj Bhadu, Akinobu HaHonda R&D, Ltd.,HondaLearning in Graspin
Yuhong Deng, Kai Mo, Chongkun Xia, Xueqian Wang National University of S Learning in Graspin
Malte Mosbach, Sven Behnke University of Bonn Learning in Graspin
Ben Abbatematteo, Eric Rosen, Rachael Thompson, Me University of Texas at ALearning in Graspin
Kai Lu, Jia-xing Zhong, Bo Yang, Bing Wang, Andrew University of Oxford,Th Learning in Graspin
Hien Bui, Michael Posa University of Pennsylva Learning in Graspin
Haofu Qian, Haoyang Zhang, Jun Shao, Jiatao Zhang, Zhejiang University,Dal Learning in Graspin
Haohong Lin, Radu Ioan Corcodel, Ding Zhao Carnegie Mellon UniversiLearning in Graspin
Entong Su, Chengzhe Jia, Yuzhe Qin, Wenxuan Zhou, University of California Learning in Graspin
Wenhao Li, Shishun Zhang, Sisi Dai, Hui Huang, Ruiz National University of Learning in Graspin
Jingyun Yang, Congyue Deng, Jimmy Wu, Rika Antonov Stanford University,StanLearning in Graspin
Jeffrey Huang, Ho Jin Choi, Nadia Figueroa University of Pennsylva Learning in Graspin
Mélodie Daniel, Anh Trinh Quoc, Miguel Aranda, LaureLaBRI - Université de Learning in Graspin
Abulikemu Abuduweili, Shiqi Zhao, Peng Yin, Changliu Carnegie Mellon Universi Learning in Localiz
Yan Gao, Jing Wu, Xintong Yang, Ze Ji Cardiff University,CA Learning in Localiz
Jin Cheng, Marin Vlastelica, Pavel Kolev, Chenhao Li, ETH Zurich,Max Planck LearningIn in Localiz
Shuo Jiang, Adarsh Salagame, Alireza Ramezani, LawNortheastern UniversityLearning in Localiz
Abhijeet Nayak, Daniele Cattaneo, Abhinav Valada University of Freiburg Learning in Localiz
Zhiyao Zhang, Yunzhou Zhang, You Shen, Lei Rong, SNortheastern UniversityLearning in Localiz
Mingsheng Yin, Tao Li, Haozhe Lei, Yaqi Hu, Sundee New York University,N Learning in Localiz
Pengbin Chen, Qi Liu, Yanjie Li, Shuaikang Ma Harbin Institute of Tec Learning in Localiz
Vishnu Sharma, Anukriti Singh, Pratap Tokekar University of Maryland Learning in Localiz
Jakob Thumm, Felix Trost, Matthias Althoff Technical University of Learning in Plannin
Wen Zheng Terence Ng, Jianda Chen, Sinno Jialin Pannanyang technological uLearning in Plannin
Vinal Asodia, Zhenhua Feng, Saber Fallah University of Surrey Learning in Plannin
Shijun Lin, Hao Wang, Ziyang Chen, Zhen Kan University of Science a Learning in Plannin
Mayank Mittal, Nikita Rudin, Victor Klemm, Arthur AllshETH Zurich,ETH Zurich,Learning in Plannin
Shishun Zhang, Qijin She, Wenhao Li, Chenyang Zhu, National University of Learning in Plannin
Jinning Li, Xinyi Liu, Banghua Zhu, Jiantao Jiao, Mas University of California, Learning in Plannin
Thomas Bi, Raffaello D'Andrea ETH Zurich,ETHZ Learning in Plannin
Xinyi Yang, Yuxiang Yang, Chao Yu, Jiayu Chen, Yu Tsinghua University,MeiLearning in Plannin
Anqiao Li, Chenyu Yang, Jonas Frey, Joonho Lee, CesETH Zurich Legged Robots and
Huang Huang, Antonio Loquercio, Ashish Kumar, Neerja University of California Legged Robots and
Francesco Vezzi, Jiatao Ding, Antonin Raffin, Jens KobTechnical University of Legged Robots and
Alireza Sarmadi, Farshad Khorrami, Prashanth KrishnaNew York University,New Legged Robots and
Guilherme Christmann, Ying-sheng Luo, Wei-chao CheInventec Corporation,InvLegged Robots and
Yunfei Li, Jinhan Li, Wei Fu, Yi Wu Tsinghua University Legged Robots and
Shusheng Xu, Huaijie Wang, Yutao Ouyang, Jiaxuan Ga Tsinghua University,XiaLegged Robots and
Mincheol Kim, Ukcheol Shin, Jung-Yup Kim Seoul National Universi Legged Robots and
Taixian Hou, Jiaxin Tu, Xiaofei Gao, Zhiyan Dong, Pen FuDan University,Beijin Legged Robots and
Niranjan Kumar Kannabiran, Irfan Essa, Sehoon Ha Georgia Institute of Tec Legged Robots and
Adrian Hartmann, Dongho Kang, Fatemeh Zargarbashi,ETH Zürich,ETH Zurich Legged Robots and
Dong Hoon Youm, Hyunyoung Jung, Hyeongjun Kim, Korea Advanced InstitutLegged Robots and
Yiyu Chen, Quan Nguyen University of Southern CLegged Robots and
Xuxin Cheng, Kexin Shi, Ananye Agarwal, Deepak Pat University of California Legged Robots and
Lukas Schneider, Jonas Frey, Takahiro Miki, Marco HutETH Zurich Legged Robots and
Jiyuan Shi, Chenjia Bai, Haoran He, Lei Han, Dong Wan Tsinghua University,Shan Legged Robots and
Srinath Mahankali, Chi-chang Lee, Gabriel Margolis, Z Massachusetts InstituteLegged Robots and
Zifan Xu, Amir Hossain Raj, Xuesu Xiao, Peter Stone University of Texas at Legged Robots and
Zhaoyuan Gu, Rongming Guo, William Yates, Yipu CheGeorgia Institute of Te Legged Robots I
Jonghun Choe, Joon-Ha Kim, Seungwoo Hong, Jinoh KAIST,Korea AdvancedLegged I Robots I
Maegan Tucker, Kejun Li, Aaron Ames Georgia Institute of Tec Legged Robots I
Stefan Fasano, James Paul Foster, Sylvain Bertrand, Ch Florida Institute for H Legged Robots I
George Mesesan, Robert Schuller, Johannes Englsberger German Aerospace Cent Legged Robots I
Guillermo Castillo, Bowen Weng, Wei Zhang, Ayonga HThe Ohio State UniversiLegged Robots I
Qingrui Zhao, Mingyuan Li, Yongliang Shi, Xuechao C Beijing Institute of Tech Legged Robots I
Giorgio Valsecchi, Nikita Rudin, Lennart Nachtigall, K Robotic System Lab, ETLegged Robots I
Yi Wang, Jiarong Kang, Zhiheng Chen, Xiaobin Xiong Columbia University,UniLegged Robots II
Jenny Zhang, Steve Heim, Se Hwan Jeon, Sangbae KiMassachusetts InstituteLegged Robots II
Joseph Elliot Humphreys, Jun Li, Yuhui Wan, Haibo G University of Leeds,HarbLegged Robots II
Yuhong Huang, Zhenshan Bing, Zitao Zhang, GenghanTechnische Universität Legged Robots II
John Zhang, Shuo Yang, Gengshan Yang, Arun Bisho Carnegie Mellon UniversLegged Robots II
Victor Barasuol, Sinan Emre, Vivian Suzano Medeiros, Istituto Italiano di Tecn Legged Robots II
Alexander Spiridonov, Fabio Buehler, Moriz Berclaz, V ETH Zurich,ETH Zürich Legged Robots II
Milad Shafiee, Guillaume Bellegarda, Auke Ijspeert EPFL Legged Robots II
Zhongjin Ju, Ke Wei, Lei Jin, Yundou Xu Yanshan University,YanS Legged Robots III
Philip Arm, Mayank Mittal, Hendrik Kolvenbach, Marco ETH Zurich Legged Robots III
Shi Haojie, Tingguang Li, Qingxu Zhu, Jiapeng Sheng Chinese University of Legged Robots III
Xuanqi Zeng, Hongbo Zhang, Linzhu Yue, Zhitao Song,Chinese University of Legged Robots III
Jeeseop Kim, Randall Fawcett, Vinaykarthik Kamidi, Caltech,Virginia Polytech Legged Robots III
Dongho Kang, Jin Cheng, Miguel Zamora, Fatemeh Zarg ETH Zurich Legged Robots III
Fumihiko Asano, Taiki Sedoguchi, Cong Yan Japan Advanced InstitutLegged Robots III
Kuan-lun Lu, I-Chia Chang, Wei-shun Yu, Pei-Chun LinNational Taiwan UniversLegged Robots III
Jiatao Ding, Mees Alexander Van Löben Sels, Franco ADelft University of Tech Legged Robots IV
Jiawei Chen, Ripeng Qin, Longfei Huang, Zongbo He, KBeihang university,InnerLegged Robots IV
Bhavyansh Mishra, Duncan Calvert, Sylvain Bertrand, Jer Institute of Human and Legged Robots IV
James Zhu, J. Joe Payne, Aaron Johnson Carnegie Mellon UniversLegged Robots IV
Zhe Xu, Mohamed Al-khulaqui, Hanxin Ma, Jiajun WangBeijing Institute of Te Legged Robots IV
Maxwell Asselmeier, Evgeniia Ivanova, Ziyi Zhou, PatriGeorgia Institute of Te Legged Robots IV
Qingfeng Yao, Linghan Meng, Qifeng Zhang, Jing Zhao,Shenyang Institute of A Legged Robots IV
Simon Chamorro, Victor Klemm, Miguel De La Iglesia Va Université de Sherbroo Legged Robots IV
Yuxuan Xiang, Yanqiu Zheng, Fumihiko Asano Japan Advanced InstitutLegged Robots IV
Jiaxin Zhang, Chen Shiyuan, Haoran Yin, Ruohong MeiSoochow University,NortLocalization and Ma
Leonardo Brizi, Emanuele Giacomini, Luca Di GiammariSapienza University of Localization and Ma
Linna Song, Dianxi Shi, Jianqiang Xia, Qianying Ouya National University of D Localization and Ma
Benjamin Ramtoula, Daniele De Martini, Matthew Gad University of Oxford,OxfLocalization and Ma
Beichen Xiang, Yuxin Sun, Zhongqu Xie, Xiaolong YanNanjing University of S Localization and Mapping I
Maria Leyva-vallina, Nicola Strisciuglio, Nicolai Petkov University of Groningen,Localization and Ma
Suji Jang, Ue-Hwan Kim Gwang-ju Institute of S Localization and Ma
Zhuo Wang, Yunzhou Zhang, Xinge Zhao, Jian Ning, De Northeastern UniversityLocalization and Ma
Xiaoshuai Hao, Hui Zhang, Yifan Yang, Yi Zhou, Sangi Samsung Research China Localization and Ma
Yuxiang Peng, Chuchu Chen, Guoquan Huang University of Delaware Localization and Ma
Heng Li, Yifan Duan, Xinran Zhang, Haiyi Liu, Jianmin University of Science a Localization and Ma
Xuan Yu, Yili Liu, Sitong Mao, Shunbo Zhou, Rong XionZhejiang University,Sh Localization and Ma
Madhu Vankadari, Samuel Hodgson, Sangyun Shin, KaiUniversity of Oxford,OxfLocalization and Ma
Yifu Tao, Yash Sanjay Bhalgat, Lanke Frank Tarimo FuUniversity of Oxford Localization and Ma
Liu Mingrui, Xinyang Tang, Yeqiang Qian, Jiming Chen,Zhejiang University,Shan Localization and Ma
Rongguang Wu, Chenglin Pang, Xuankang Wu, ZhengNortheastern UniversityLocalization and Ma
Xiao Han, Houxuan Liu, Yunchao Ding, Lu Yang University of Electroni Localization and Ma
Heajung Min, Kyungmin Han, Young Jun Kim Ewha Womans UniversitLocalization and Ma
Chun Qing, Rongxiang Zeng, Xuan Wu, Yongliang Shi Shenzhen Technology Un Localization and Na
Sanghun Jung, Joonho Lee, Xiangyun Meng, Byron Boo University of WashingtoLocalization and Na
Yurim Jeon, E-in Son, Seung-Woo Seo Seoul National UniversitLocalization and Na
Robin Karlsson, Alexander Carballo, Francisco Lepe-saNagoya University,Gifu Localization and Na
Yehui Shen, Mingmin Liu, Huimin Lu, Xieyuanli Chen NorthEast University,S Localization and Na
Koki Aoki, Kenji Koide, Shuji Oishi, Masashi Yokozuka Meijo University,NationaLocalization and Na
Huanfeng Zhao, Meibao Yao, Xueming Xiao, Bo ZhengJilin University,Changc Localization and Na
Hongcheng Wang, Yuxuan Wang, Zfw Fangwei Zhong,Peking University,PekinLocalization and Na
George Tang, Krishna Murthy, Antonio Torralba MIT Localization and Na
Jooyong Park, Jungwoo Lee, Euncheol Choi, Younggu Inha University,Inha univLocalization I
Yixiao Feng, Zhou Jiang, Yongliang Shi, Yunlong Fen University of New SouthLocalization I
Haifeng Li, Jiajun Guo, Dezhen Song Civil Aviation Universit Localization I
Chunge Bai, Ruijie Fu, Xiang Gao AgiBot Technology Co. L ocalization I
Patrick Pfreundschuh, Helen Oleynikova, Cesar D. Ca ETH Zurich,KTH Royal In Localization I
Kenji Koide, Shuji Oishi, Masashi Yokozuka, Atsuhiko National Institute of A Localization I
Kenji Koide, Shuji Oishi, Masashi Yokozuka, Atsuhiko National Institute of A Localization I
Spencer Carmichael, Rahul Agrawal, Ram Vasudevan, University
K of Michigan Localization I
Liming Han, Zhaoxiang Liu, Shiguo Lian China Unicom Localization II
Alejandro Fontán Villacampa, Javier Civera, Michael MiQueensland University Localization II
Xiyue Guo, Haocheng Peng, Junjie Hu, Hujun Bao, Gu Zhejiang University,TheLocalization II
Changhao Wang, Guanwen Zhang, Wei Zhou Northwestern Polytechnic Localization II
Alberto García-hernández, Riccardo Giubilato, Klaus H.Universidad de Zarago Localization II
Zhiqiang` Chen, Hongbo Chen, Yuhua Qi, Shipeng ZhoSUN YAT-SEN UNIVERSITY Localization II
Zhenyu Xu, Ren Ziliang, Qieshi Zhang, Lou Jie, Dache CAS Key Laboratory of Localization II
Hailiang Tang, Tisheng Zhang, Xiaoji Niu, Liqiang WangWuhan University Localization II
Jun-Jee Chao, Kazim Selim Engin, Nikhil Chavan-dafle,University of Minnesot Localization II
Obin Kwon, Dongki Jung, Youngji Kim, Soohyun Ryu, Seoul Natl University, Localization III
Chengyu Qiao, Zhiyu Xiang, Yuangang Fan, Tingming Ba Zhejiang University,ChinaLocalization III
Song Wenzheng, Ran Yan, Boshu Lei, Takayuki Okatan Tohoku University,MegviLocalization III
Le Chen, Weirong Chen, Rui Wang, Marc Pollefeys Max Planck Institute for Localization III
Gabriele Berton, Gabriele Trivigno, Barbara Caputo, C Politecnico di Torino,PoLocalization III
Jiafeng Cui, Xieyuanli Chen Tongji University,NationLocalization III
Alexander Schperberg, Yusuke Tanaka, Saviz Mowlavi,University of Californi Localization III
Milad Ramezani, Liang Wang, Joshua Barton Knights, CSIRO,University of QuLocalization III
Javier Morlana, Juan D. Tardos, Jose M M Montiel Universidad de Zaragoza Localization III
Denis Von Arx, Cedric Fischer, Harun Torlakcik, Salva ETH Zurich,ETHZ Localization III
Chencan Fu, Lin Li, Jianbiao Mei, Yukai Ma, Linpeng PZhejiang University,zhejLocalization IV
Xinrui Wu, Jianbo Xu, Puyuan Hu, Guangming Wang, Shanghai Jiao Tong UniLocalization IV
Louis Wiesmann, Tiziano Guadagnino, Ignacio Vizzo, NUniversity of Bonn,DexoLocalization IV
Pengyu Zhang, Shuaifeng Zhi, Yuelin Yuan, Beizhen BiNational University of Localization IV
Yanhao Zhang, Yujiao Shi, Shan Wang, Ankit Vora, AkhUniversity of Technolog Localization IV
Zhendong Xiao, Changhao Chen, Yang Shan, Wu Wei South China University Localization IV
Jiaming Cui, Jiming Chen, Liang Li Zhejiang University,ZhejLocalization IV
Changkun Liu, Shuai Chen, Yukun Zhao, Huajian Huang, The Hong Kong UniversiLocalization IV
Mohammad Altillawi, Shile Li, Sai Manoj Prakhya, ZiyuaHuawei, Autonomous Uni Localization IV
Joo Hyung Kim, Gunhee Koo, Heewon Park, Nakju Do Korea University,SamsuLocalization V
Xudong Cai, Yongcai Wang, Zhe Huang, Yu Shao, DeyiRenmin University of ChLocalization V
Adam D. Hines, Peter Stratton, Michael Milford, TobiasQueensland University oLocalization V
Chen Xie, Rui Xing, Ning Hao, Fenghua He Harbin Institute of Tech Localization V
Nikhil Varma Keetha, Avneesh Mishra, Jay Karhade, K Carnegie Mellon Universi Localization V
Aaron Wilhelm, Nils Napp Cornell University Localization V
Saimouli Katragadda, Woosik Lee, Yuxiang Peng, Pat University of Delaware, Localization V
Tianxiao Gao, Mingle Zhao, Cheng-zhong Xu, Hui Kon University of Macao,UniLocalization V
Junlin Song, Pedro J Sanchez-cuevas, Antoine RichardUniversity of Luxembourg Localization VI
Nicky Zimmerman, Hanna Müller, Michele Magno, LucaUniversity of Lugano,ETLocalization VI
Ilayda Yaman, Guoda Tian, Martin Larsson, Patrik PerssLund University,LTH, LuLocalization VI
Wenqian Lai, Ruonan Guo, Kejian Wu XREAL Localization VI
Jinhao He, Huaiyang Huang, Shuyang Zhang, Jianhao The Ji Hong Kong UniversiLocalization VI
Jessica Paterson, Bruno Vilhena Adorno, Barry LennoxUniversity of ManchesteLocalization VI
Abhishek Goudar, Zhao Wenda, Angela P. Schoellig University of Toronto,T Localization VI
Alessandro Fornasier, Yixiao Ge, Pieter Van Goor, MarUniversity of Klagenfurt Localization VI
Fan Jiang, David Caruso, Ashutosh Dhekne, Qi Qu, Ja Georgia Institute of Te Localization VI
Zhaoguang Yi, Xiangyu Wen, Qiyue Xia, Peize Li, FrancUniversity of Edinburg Localization VII
Shigemichi Matsuzaki, Takuma Sugino, Kazuhito Tanaka Toyota Motor Corpora Localization VII
Taro Suzuki Chiba Institute of TechnLocalization VII
Fujing Xie, Soeren Schwertfeger Shanghaitech Universit Localization VII
Qingyu Xiao, Zulfiqar Zaidi, Matthew Gombolay Georgia Institute of Tec Localization VII
Andres Galeote-luque, Vladimir Kubelka, Martin Magnus Universidad de Málaga,Ö Localization VII
Vladimir Kubelka, Emil Fritz, Martin Magnusson Örebro University Localization VII
Pengyu Yin, Cao Haozhi, Thien-Minh Nguyen, ShenghaNanyang Technological Localization VII
Chuchu Chen, Yuxiang Peng, Guoquan Huang University of Delaware Localization VII
Gojko Perovic, Francesco Iori, Angela Mazzeo, Marco CoScuola Superiore Sant' Machine Learning fo
Xuyang Zhang, Ziyang Feng, Quecheng Qiu, Yu'an Chen University of Science anMachine Learning fo
Yu Zhang, Long Wen, Xiangtong Yao, Zhenshan Bing, Technical University of Machine Learning fo
Chenchang Li, Zihao Ai, Tong Wu, Xiaosa Li, Wenbo DTsinghua University Machine Learning fo
Angel Romero, Yunlong Song, Davide Scaramuzza University of Zurich Machine Learning fo
Adam Villaflor, Brian Yang, Huangyuan Su, Aikaterini FCMU,University of CalifoMachine Learning fo
So Kuroki, Jiaxian Guo, Tatsuya Matsushima, Takuya The university of Tokyo,Machine Learning fo
Jeremy Dao, Helei Duan, Alan Fern Oregon State UniversityMachine Learning fo
Abdullah Altawaitan, Jason Stanley, Sambaran Ghosal,University of California Machine Learning fo
Jacob Sacks, Rwik Rana, Kevin Huang, Alexander SpitUniversity of WashingtoMachine Learning fo
Sean J. Wang, Honghao Zhu, Aaron Johnson Carnegie Mellon Univer Machine Learning fo
Jianlan Luo, Zheyuan Hu, Charles Xu, You Liang Tan, UC Berkeley,University M o achine Learning fo
Yating Lin, Glen Chou, Dmitry Berenson University of Michigan, Machine Learning fo
Chethan Bhateja, Derek Guo, Dibya Ghosh, Anikait Si Stanford University,UC Machine Learning fo
Deniz Kerimoglu, Daniel Soto, Malone Lincoln Hemsle Georgia Institute of Te Machine Learning fo
xiao zhang, Hai Zhang, Hongtu Zhou, Chang Huang, DiTongji University,TongJiMachine Learning fo
Baichuan Huang, Xujia Zhang, Jingjin Yu Rutgers University,SoutManipulation Plann
Utkarsh Aashu Mishra, Yongxin Chen Georgia Institute of Tec Manipulation Plann
Jeremy Collins, Cody Houff, You Liang Tan, Charlie K Georgia Institute of Tec Manipulation Plann
Xinyuan Zhao, Wenyu Liang, Xiaoshi Zhang, Chee Me Agency for Science, TecManipulation Plann
Peter Mitrano, Dmitry Berenson University of Michigan Manipulation Plann
Suhan Ling, Yian Wang, Ruihai Wu, Shiguang Wu, Yuzh Peking University,UmasManipulation Plann
Dhruv Saxena, Maxim Likhachev The Robotics Institute, Manipulation Plann
Ramkumar Natarajan, Hanlan Yang, Qintong Xie, Yas Robotics Institute, Car Manipulation Plann
Katie Luo, Xinshuo Weng, Yan Wang, Shuang Wu, Jie C L ornell University,NVIDMapping I
Shuo Sun, Malcolm Mielle, Achim J. Lilienthal, Martin Orebro University,SchinMapping I
Jianheng Liu, Haoyao Chen Harbin Institute of Tec Mapping I
Shiyao Xu, Caiyun Liu, Yuantao Chen, Zhenxin Zhu, Z Institute for AI Industr Mapping I
Ji Wu, Huai Yu, Wen Yang, Gui-Song Xia Wuhan University Mapping I
Jiabao Zhang, Yu Zhang Zhejiang University Mapping I
Chenxing Jiang, Hanwen Zhang, Peize Liu, Zehuan Yu,The Hong Kong Universit Mapping I
Ziwei Liao, Jun Yang, Jingxing Qian, Angela P. Schoel University of Toronto,T Mapping I
Sourav Garg, Krishan Rana, Mehdi Hosseinzadeh, Lach University of Adelaide,QMapping I
Joel Loo, David Hsu National University of S Mapping II
Gang Chen, Wei Dong, Peng Peng, Javier Alonso-MoraDelft University of Tec Mapping II
Benedikt Mersch, Tiziano Guadagnino, Xieyuanli Chen,University
I of Bonn,Nati Mapping II
Igor Bogoslavskyi, Konstantinos Zampogiannis, Raym Magic Leap Mapping II
Bonan Liu, Guoyang Zhao, Jianhao Jiao, Guang Cai, CHKUST(GZ),University C Mapping II
Qianqian Zou, Claus Brenner, Monika Sester Leibniz University Hanno Mapping II
Yixi Cai, Fanze Kong, Yunfan Ren, Fangcheng Zhu, JiaUniversity of Hong KongMapping II
Zihong Yan, Xiaoyi Wu, Zhuozhu Jian, Bin Lan, Xueqi Tsinghua University,Harb Mapping II
Luca Di Giammarino, Emanuele Giacomini, Leonardo Bri Sapienza University of Mapping II
Pascal Spino, Daniela Rus Massachusetts InstituteMarine Robotics I
Xu Chao, Imran Hameed, D. Navarro-Alarcon, Xingjian The Hong Kong Polytechn Marine Robotics I
Cora Maria Sourkounis, Tom Kwasnitschka, Annika RaLeibniz University Han Marine Robotics I
Marcus Rosette, Hannah Kolano, Chris Holm, GeoffreyOregon State UniversityMarine Robotics I
Yulin Bai, Yufei Jin, Chunhu Liu, Zheng Zeng, Lian LianShanghai Jiao Tong Univ Marine Robotics I
Timothy Ju Kin Ng, Nan Chen, Fu Zhang The University of Hong Marine Robotics I
Yaohui Xu, Hanlin Li, Furui Yu, Qiyang Zuo, Fengran XiShenzhen Institute of Marine Robotics I
Anna Astolfi, Mrudul Chellapurath, Giacomo Picardi, Mart
Scuola Superiore Sant'AMarine Robotics I
Lepeng Chen, Rongxin Cui, Weisheng Yan, Hui Xu, ShNorthwestern PolytechniMarine Robotics I
Herman Biørn Amundsen, Torben Falleth Olsen, MariosNTNU,SINTEF
X Ocean AMarine Robotics II
Luca Ebner, Gideon Billings, Stefan Bernard Williams ETH Zurich,University ofMarine Robotics II
Davide Sapienza, Elena Govi, Sara Aldhaheri, BertgonaUnimore,TII,TechnologyMarine Robotics II
Samiran Gode, Akshay Hinduja, Michael Kaess Carnegie Mellon UniversMarine Robotics II
Amin Ullah, Taiqing Yan, Fuxin Li Oregon State UniversityMarine Robotics II
Tianyi Zhang, Matthew Johnson-Roberson Carnegie Mellon UniversMarine Robotics II
Adnan Abdullah, Titon Barua, Reagan Tibbetts, Zijie CheUniversity of Florida,UniMarine Robotics II
Seth McCammon, Stewart Jamieson, T. Aran Mooney, Woods Hole Oceanographi Marine Robotics II
Mingyue Cui, Mingjian Feng, Junhua Long, Daosong HuSun Yat-sen Universit Marine Robotics II
Martin Filliung, Juliette Drupt, Charly Peraud, Claire CNRS LIS, COSMER Labor Marine Robotics III
James Hunt, Louis Whitcomb Johns Hopkins UniversitMarine Robotics III
Yuanbo Bi, Yufei Jin, Hexiong Zhou, Yulin Bai, ChenxinShanghai jiao tong UnivMarine Robotics III
Kevin Macauley, Levi Cai, Peter G. Adamczyk, YogeshUniversity of Wisconsin Marine Robotics III
Wei Wang, Wei Xiao, Alejandro Gonzalez-Garcia, Jan SUniversity of Wisconsin Marine Robotics III
Xinmeng Ma, Gang Wang, Liu Kaixin Harbin Engineering UnivMarine Robotics III
Monika Roznere, Adithya Pediredla, Samuel Lensgraf, Y Dartmouth College,Wood Marine Robotics III
Li Ling, Jun Zhang, Nils Bore, John Folkesson, Anna WKTH Royal Institute of Marine Robotics III
Evan Palmer, Chris Holm, Geoffrey Hollinger Oregon State UniversityMarine Robotics III
Tyler Paine, Michael Benjamin Massachusetts InstituteMarine Robotics IV
Junwoo Jang, Sangli Teng, Maani Ghaffari University of Michigan,UMarine Robotics IV
Priti Singh, Geoffrey Hollinger Oregon State UniversityMarine Robotics IV
Xi Lin, Yewei Huang, Fanfei Chen, Brendan Englot Stevens Institute of Te Marine Robotics IV
Ivana Collado-Gonzalez, John Frederic Mcconnell, Jin Stevens Institute of Te Marine Robotics IV
Zuoquan Zhao, Yu Zhai, Chuanxiang Gao, Wendi Ding,The Chinese University Marine Robotics IV
Advaith Venkatramanan Sethuraman, Philip Baldoni, K University of Michigan, Marine Robotics IV
Juzheng Mao, Guangming Song, Shuang Hao, MingquaSoutheast University Marine Robotics IV
Tim Hansen, Andreas Birk Constructor University Marine Robotics V
Aldo Terán Espinoza, antonio teran, John Folkesson, PKTH Royal Institute of Marine Robotics V
Carl Schiller, Deran Maas, Bruno Arsenali, Jukka Pelto ABB Corporate Research Marine Robotics V
Corey Knutson, Zhipeng Cao, Junaed Sattar University of Minnesota Marine Robotics V
Jonghwi Kim, Changyu Lee, Dongha Chung, Jinwhan KKAIST Marine Robotics V
Mohit Singh, Mihir Rahul Dharmadhikari, Kostas Alexis NTNU: Norwegian Univer Marine Robotics V
Bharat Joshi, Ioannis Rekleitis University of South CaroMarine Robotics V
Jiangbo Song, Wanqing Li, Xiangwei Zhu Sun Yat-sen University Marine Robotics V
António José Oliveira, Bruno Ferreira, Nuno Cruz INESC TEC,University oMarine Robotics V
Jaeu Choe, Jeongseob Lee, Hyunsoo Yang, Hai-nguyeSeoul National UniversitMechanics and Cont
Shawndy Michael Lee, Jingmin Liu, Jer Luen Chien, WeSINGAPORE UNIVERSITY Mechanics and Cont
Yuxia Yuan, Markus Ryll Technical University of Mechanics and Cont
Wooyong Park, Xiangyu Wu, Dongjae Lee, Seung Jae Seoul National UniversitMechanics and Cont
Xiaofeng Guo, Guanqi He, Mohammadreza Mousaei, Ju Carnegie Mellon UniveriMechanics and Cont
Wei Yan, Genliang Chen, Zhuang Zhang, Hao Wang Shanghai Jiaotong Unive Mechanics and Cont
Dongjae Lee, Sunwoo Hwang, Jeonghyun Byun, SeungSeoul National Universi Mechanics and Cont
Guanrui Li, Liu Xinyang, Giuseppe Loianno New York University Mechanics and Cont
Bruno Chapdelaine, Mathis Celce, Charles Vidal, LionelNational Research CounMechanics and Cont
Jiawei Sun, Xiang Zhou, Taoze Ban, Jiannan Zhao, F Guangxi University,Co.,Mechanics and Contr
Chiara Gabellieri, Marco Tognon, Dario Sanalitro, Anto University of Twente,InrMechanics and Contr
Tian Lan, Luca Romanello, Mirko Kovac, Sophie FranziTechnical University of Mechanics and Contr
Meng Wang, Zeshuai Chen, Kexin Guo, Xiang Yu, YouBeihang University,ConcMechanics and Contr
Lucas Waelti, Alcherio Martinoli EPFL Mechanics and Contr
Krishna Murali, Elena Ponce Moreno, Leandro LustosaISAE-SUPAERO,ISAE-SMechanics and Contr
Grant Nations, Charles Luke Nelson, Daniel S. Drew University of Utah Mechanics and Contr
Deepak Prakash Kumar, Swaroop Darbha, SatyanarayTexas A&M University,TMechanics and Contr
Aliaksei Pilko, Andy Oakey, Mario Ferraro, James ScanUniversity of SouthamptMechanics and Contr
Sensen Liu, Zhaoying Wang, Xinjun Sheng, Wei Dong ShanghaiJiaotong Univer Mechanics and Contr
Emmanuel Tang, Wei Jun Ang, Kian Wee Tan, ShaohuSingapore University of Mechanics and Contr
Tong Hui, Eugenio Cuniato, Michael Pantic, Marco TogTechnical University of Mechanics and Contr
Jiali Sun, Kaidi Wang, Chuanbeibei Shi, Xiujia Li, Xia Beijing Institute of Tec Mechanics and Contr
Hee-deok Jang, Jae-Hyeon Park, Dong Eui Chang Korea Advanced Institu Mechanics and Contr
Robin Ferede, Christophe De Wagter, Dario Izzo, Guid TU Delft,Delft Universi Mechanics and Contr
Huaiyuan Jia, Runze Ding, Kaixu Dong, Songnan Bai, City University of H Mechanics and Contr
Fabio Bergonti, Gabriele Nava, Valentin Wueest, Antone Istituto Italiano di Tec Mechanics and Contr
Zujian Chen, Mo Shaolin, Botao Zhang, Hui Cheng, JiyuShenzhen University,Sun Mechanics and Contr
Francisco Goncalves, Ryan Bena, Konstantin Matveev,Washington State Univer Mechanics and Cont
Yanhui Zhang, Zheyu Tong, Yifan Zhang, Song Chen, Zhejiang University,zhejMechanics and Control IV
Tong Hui, Manuel Jesús Fernández González, Matteo Technical University of Mechanics and Cont
Nicolas Gorlo, Mario Sven Müller, Samuel Bamert, TimETH Zurich,ETH Zürich,Mechanics and Cont
Ying Wu, Zida Zhou, Mingxin Wei, Hui Cheng Sun Yat-sen Universi Mechanics and Cont
Akash Kopparam Sreedhara, Deepesh Padala, Shashan Vellore Institute of Tec Mechanics and Cont
Mengguang Li, Kai Cui, Heinz Koeppl Technische Universität Mechanics and Cont
Hemjyoti Das, Minh Nhat Vu, Tobias Egle, Christian OttTechnical University of Mechanics and Cont
Louis Petit, Alexis Lussier Desbiens McGill University,Unive Mechanics and Cont
Hangyeol Song, Edgar Lee, Hyung-tae Seo, SeokhwanSogang University,Kyong Mechanism Design
Timothy Amish, Jeffrey Auletta, Chad C. Kessens, Joshu University of Washingt Mechanism Design
Shengjie Xiao, Kai Hu, Yuhong Sun, Yun Wang, Bo QinBeijing University of A Mechanism Design
Lukáš Supik, Kateřina Stránská, Miroslav Kulich, Libor Czech Technical Universi Mechanism Design
Paul Baisamy, Adam A. Stokes, Francesco Giorgio-SerThe University of EdinbuMechanism Design
Dongdong Du, Emanuela Del Dottore, Alessio Mondini,Zhejiang University,IstituMechanism Design
Carlos Adrian Perez Diaz, Ignacio Muñoz Planelles, LuArquimea Research Cente Mechanism Design
Linliang Zheng, Qingcong Wu, Yanghui Zhu, Qiang Zh Nanjing University of AeMechanism Design
Moses Gladson Selvamuthu, Kazuki Abe, Kenjiro TadaYamagata University,ToMechanism Design
Seunghwan Um, Heeyeon Jeong, Chunsoo Kim, Issac SungKyunKwan UniversiMechanism Design I
Tasuku Makabe, Kei Okada, Masayuki Inaba The University of TokyoMechanism Design I
Yunong Li, Hailin Huang, Bing Li Harbin Institute of Tec Mechanism Design I
Haoran Li, Yongzhong Bu, Yongjian Bu, Shixin Mao, YiGuangdong University ofMechanism Design I
Do-yun Kim, Seong-Ho Yun, Joongkyung Lee, Jongju KOREATECH University,K Mechanism Design I
Cheng Liang, Bairun Lin, Huihuan Qian Shenzhen Institute of ArMechanism Design I
Youngrae Kim, Seunghyun Lim, Hanmin Lee, SeokchanDaegu Gyeongbuk Institu Mechanism Design I
Marco Bolignari, Gianluca Rizzello, Luca Zaccarian, M University of Trento,Sa Mechanism Design I
Zirui Fan, Hongying Zhang National University of S Mechanism Design I
Yuto Miyazaki, Mitsuru Higashimori Graduate School of EngiMechanisms and De
Haochen Yong, Fukang Xu, Chenfei Li, Han Ding, Zhi Huazhong University of Mechanisms and De
Eric Sihite, Alireza Ramezani, Gharib Morteza California Institute of Mechanisms and De
Lianchao Yang, Chuang Zhang, Ruiqian Wang, Yiwei Zh Shenyang Institute of A Mechanisms and De
Lorenzo Vianello, Emek Baris Kucuktabak, Matthew Sho Shirley Ryan Ability LabMedical Robotics
Zhengyang Li, Magejiang Yeerbulati, Qingsong Xu University of Macau Medical Robotics
Yongyi Jia, Shu Miao, Junjian Zhou, Niandong Jiao, LiaTsinghua University,SheMedical Robotics
Runfeng Zhu, Xilong Hou, Wei Huang, Lei Du, Zhong WThe Hong Kong Polytechn Medical Robotics
Shervin Dehghani, Michael Sommersperger, Mahdi Saleh, TUM,Technical UniversitMedical Robotics
Ryan Posh, Jonathan Allen Tittle, David Kelly, Jim Sch University of Notre DamMedical Robotics
Michael Sommersperger, Shervin Dehghani, Philipp MaTechnical University of Medical Robots I
Shujing He, Yujie Zhang, Baoyi Huang, Jie Lin, Chaoy Southern University of Medical Robots I
Zejian Cui, João Cartucho, Stamatia Giannarou, Ferd Imperial College LondonMedical Robots I
Luca Beber, Edoardo Lamon, Davide Nardi, Daniele Font University of Trento Medical Robots I
Jichen Li, Shuxin Wang, Zhiqiang Zhang, Chaoyang ShTianjin University,UniverMedical Robots I
Yordanka Velikova, Mohammad Farid Azampour, Walter TU Munich,Technical UnMedical Robots I
Junjie Yang, Zhihao Zhao, Mathias Maier, Kai Huang, NTUM,Technische Univers Medical Robots I
Zhongliang Jiang, Xuesong Li, Chenyu Zhang, Yuan Bi,Technical University o Medical Robots I
Yuuki Shida, Souto Kumagai, Ryosuke Tsumura, HiroyWaseda University,NatioMedical Robots I
Zhongliang Jiang, Felix Duelmer, Nassir Navab Technical University of Medical Robots II
Jiayu Zhu, Runzhe Han, Mengning Yuan, Bimeng Jie, Beihang University,PekiMedical Robots II
Tianliang Li, Chen Zhao, Yuhang Wen, Fayin Chen, Y Wuhan University of Te Medical Robots II
Ning Wang, Xiaodong Zhang, Danail Stoyanov, Hongbing Xi'an Jiaotong Universit Medical Robots II
Chen Qian, Zhen Li, Qiang Ye, Pei Cong Ge, Jizong ZhInstitute of Automation, Medical Robots II
Muneaki Miyasaka, Pepijn Van Esch, Atsushi MorikawaRiverfield Inc.,Riverfiel Medical Robots II
Hisashi Ishida, Manish Sahu, Adnan Munawar, NimeshJohns N Hopkins UniversitMedical Robots II
Yunfei Luan, Muhang He, Yudong Tian, Chengjie Lin, Shanghai Jiao Tong Univ Medical Robots II
Ziyang Chen, Ke Fan, Laura Cruciani, Matteo Fontana,Politecnico di Milano,EuMedical Robots II
Jin Fang, Ling Li, Xiaojian Li, Hangjie Mo, Pengxin GuoHefei University of Tec Medical Robots III
Tao Zhang, Huxin Gao, Hongliang Ren Chinese University of Medical Robots III
Yonggan Yan, Shuxiang Guo, Chuqiao Lyu, Jian Guo, JBeijing Institute of Technology,Kagawa
Medical Robots University,Shenzhen
III
Hiroyuki Suzuki Sony Computer ScienceMedical
L Robots III
Jintaek Im, Eunsil Jang, Cheol Song DGIST,Daegu Gyeongbuk Medical Robots III
Junnan Xue, Moqiu Zhang, Xurui Liu, Jiaqi Zhu, Yanfei The Chinese University Medical Robots III
Yanding Qin, Zhuocong Cai, Jianda Han Nankai University Medical Robots III
Jian Chen, Mingcong Chen, Qing Xiang Zhao, Shuai WUniversity of Chinese A Medical Robots III
Alexander Schwarz, Arian Mehrfard, Golchehr AmirkhaJohns Hopkins UniversitMedical Robots III
Reinhard Grassmann, Anastasiia Senyk, Jessica BurgnUniversity of Toronto,UkMedical Robots IV
Jiajia Zhang, Yang Liu, Jiyuan Tian, Dibin Zhu, Shyam University of Exeter,Ro Medical Robots IV
Premal Gorroochurn, Charles Hong, Carter Klebuc, YitoColumbia University,Geo Medical Robots IV
Beili Dong, Junhong Chen, Zeyu Wang, Kaizhong Deng, Imperial College LondonMedical Robots IV
Kenneth Afebu, Jiyuan Tian, Yang Liu, Evangelos Pap University of Exeter,Ro Medical Robots IV
Alex Ranne, Yordanka Velikova, Nassir Navab, Ferdin Imperial College Londo Medical Robots IV
Hanyi Zhang, Guocai Yao, Feifan Zhang, Fanchuan LinImperial College LondonMedical Robots IV
Alireza Alikhani, Satoshi Inagaki, Shervin Dehghani, MaAugen- klinik und Polik Medical Robots IV
Mojtaba Esfandiari, Ji Woong Kim, Botao Zhao, Golcheh Johns Hopkins UniversitMedical Robots IV
Peiyang Jiang, Wei Li, Yifan Li, Dandan Zhang university of Bristol,Imp Medical Robots IV
Sishen Yuan, Guang Li, Baijia Liang, Lailu Li, Qingzh The Chinese University Medical Robots IV
Hui Liu, Ning Li, Shuai Li, Gregory Mancini, Jindong TaUniversity of TennesseeMedical Robots V
Sishen Yuan, Baijia Liang, Po Wa Wong, Mingjing Xu, The Chinese University Medical Robots V
Chuqiao Lyu, Shuxiang Guo, Yonggan Yan, Yongxin ZhBeijing Institute of Tec Medical Robots V
Yiang Lu, Bin Li, Wei Chen, Junyan Yan, Shing Shin ChThe Chinese University Medical Robots V
Jiawei Ge, Michael Kam, Justin Opfermann, Hamed Sae Johns Hopkins UniversitMedical Robots V
Nicholas Strohmeyer, Ji Hwan Park, Braden Murphy, FaUniversity of Texas at AMedical Robots V
Jenireth Torrealba Molina, Toqa Abubaker, Yanpei Huan Imperial College LondonMedical Robots V
Satoshi Inagaki, Alireza Alikhani, Nassir Navab, MathiaNSK.Ltd,Augen- klinik uMedical Robots V
Christoff Heunis, Bruno Silva, Giulia Sereni, Martijn University of Twente Medical Robots V
Jing Dai, Bohao Zhu, Xiaomei Wang, Zhiyi Jiang, Men The University of Hon Medical Robots VI
Shuai Wang, Qing Xiang Zhao, Jian Chen, Mingcong Ch HKPU(The Hong Kong Poly Medical Robots VI
Naner Li, Yiwei Wang, Haoyuan Cheng, Huan Zhao, HaHuazhong University of Medical Robots VI
Nicotra Emanuele, Nguyen Chi Cong, James J. Davies,UNSW Sydney,University Medical Robots VI
Yang Liu, Hansoul Kim, Yash Kulkarni, Farshid AlambeiThe University of Texas Medical Robots VI
Yanzhou Wang, Lidia Al-zogbi, Guanyun Liu, Jiawei Liu,Johns Hopkins UniversitMedical Robots VI
Tian Huanyu, Martin Huber, Christopher Edwin Mower,Beijing Institution of Medical Robots VI
Yue Zhao, Xiaoming Liu, Ruixi Wang, Dan Liu, Masaru Beijing Institute of Tec Micro/Nano Robots
Sukjun Kim, Sarah Bergbreiter University of California Micro/Nano Robots
Zichen Xu, Qingsong Xu, Hon Ho Yu University of Macau,KiaMicro/Nano Robots
Aaron C. Davis, Emmett Freeman, Dave Cappelleri Purdue University Micro/Nano Robots
Toshiro Yamanaka, Fumihito Arai The University of TokyoMicro/Nano Robots
Lorenzo Noseda, Addison Liu, Lucio Pancaldi, MahmutEPFL,Harvard UniversitMicro/Nano Robots
Hongyu Zhao, Min Ye, Bradley Nelson, Xiaopu Wang Shenzhen Institute of ArtMicro/Nano Robots
Ece Ozelci, Erfan Etesami, Laurel Rohde, Andrew Oa EPFL Micro/Nano Robots
Yujian An, Jianxin Yang, Jinkai Li, Bingze He, Yao G Shanghai Jiao Tong Univ Micro/Nano Robots
Yujie Jiang, Xiang Fu, Chengxi Zhong, Teng Li, HaojianShanghaiTech University Micro/Nano Robots
Han Yang, Yufei Jin, Guanqiao Shan, Yibin Wang, YonThe Chinese University Micro/Nano Robots
Ao Wang, Chunyuan Gan, Haocheng Han, Hongyi Xiong, Beihang University,BeihMicro/Nano Robots
Youchao Zhang, Xiangyu Guo, Qingyu Wang, FanghaoZhejiangW University Micro/Nano Robots
Haoqi Han, Xiaoming Liu, Yan Chen, Hao Pang, XiaoqiShanghai Jiao Tong Unive Micro/Nano Robots
Yide Liu, Bo Feng, Tianlun Cheng, Yanhong Chen, Xiy zhejiang university,Zhej Micro/Nano Robots
Yuxuan Xiang, Jiachen Zhang City University of Hong Micro/Nano Robots
Wei Dai, Zixuan Wu, Jiaqi Wang, Rui Liu, Min Wang, T City University of Hon Microrobotics for Bi
Yameng Zhang, Ao Xu, Yuhan Chen, Max Qing Hu Meng The Chinese University Microrobotics for Bi
Kazusa Otani, Hirotaka Sugiura, Shiro Watanabe, Bilal The University of TokyoMicrorobotics for Bi
Freddy Romero Leiro, Stéphane Régnier, Frederic De Sorbonne Université - I Microrobotics for Bi
Haocong Song, Wenyuan Chen, Changsheng Dai, Guanq University of Toronto,u Microrobotics for Bi
Wenyuan Chen, Haocong Song, Changsheng Dai, Aojun university
J of Toronto,U Microrobotics for Bi
Yating Luo, Yuxuan Liu, Jiasheng Zhou, Sung-liang Shanghai Jiao Tong Univ Microrobotics for Bi
Zhaofeng Zuo, Xiaoming Liu, Zhuo Chen, Yuyang Li, XiBeijing institute of tec Microrobotics for Bi
Leendert-Jan Wouter Ligtenberg, Nicole Christina AntoiUniversity of Twente,UnMicrorobotics for Bi
Juncheng Li, Dave Cappelleri Purdue University Mobile Manipulatio
Max Merlin, Shane Parr, Neev Parikh, Sergio Orozco, Brown University,UniverMobile Manipulatio
Chengkai Wu, Ruilin Wang, Mianzhi Song, Fei Gao, Ji Harbin Institute of Te Mobile Manipulatio
Fabian Schmalstieg, Daniel Honerkamp, Tim WelschehUniversity of Freiburg,AMobile Manipulatio
Adam Heins, Angela P. Schoellig University of Toronto,T Mobile Manipulatio
Dimitrios Rakovitis, Dennis Mronga DFKI,University of Breme Mobile Manipulatio
Jiazhao Zhang, Gireesh Nandiraju, Jaylon Wang, Xia Peking University,IIIT HMobile Manipulatio
Qikai Li, Qinchen Meng, Yuxing Qin, Jiawei Chen, Xilu Beihang University,Beiha Mobile Manipulatio
Snehal Jauhri, Sophie C. Lueth, Georgia Chalvatzaki TU Darmstadt,TechnicalMobile Manipulatio
Alessandro Saviolo, Jonathan Frey, Abhishek Rathod, New York University,Univ Model Learning for
Alison Bartsch, Charlotte Avra, Amir Barati Farimani Carnegie Mellon UniversModel Learning for
David Gomes, Miguel Botto, Pedro U. Lima Instituto Superior Técni Model Learning for
Hojin Lee, Taekyung Kim, Jungwi Mun, Wonsuk Lee Agency for Defense De Model Learning for
Siming Deng, Noah J. Cowan, Brian Bittner Johns Hopkins Univers Model Learning for
Namhoon Cho, Taeyoon Lee, Hyo-Sang Shin Cranfield University,NavModel Learning for
Tyler Toner, Vahidreza Molazadeh, Miguel Saez, DawnUniversity
T of Michigan, Model Learning for
Gil Serrano, Marcelo Jacinto, Jose Ribeiro-gomes, JoaoInstituto Superior Técni Model Learning for
Ananya Trivedi, Mark Zolotas, Adeeb Abbas, Sarvesh Pr Northeastern UniversityModel Learning for
Andre Meixner, Mischa Carl, Franziska Krebs, Noémie Karlsruhe Institute of T Modeling and Simu
Alice Mylaeus, Stephanie Vogt, Berken Utku Demirel, METH Zurich,ETH Zürich Modeling and Simu
Dingrui Wang, Lai Zheyuan, Yuda Li, Yi Wu, Yuexin MaTechnical University of Modeling and Simu
David Schneider, Marco Keller, Zeyun Zhong, Kunyu Pen Karlsruhe Institute of T Modeling and Simu
Zhenyu Lu, Lu Chen, Hengtai Dai, Haoran Li, Zhou Z Bristol Robotics LaboratModeling and Simu
Jacques Zhong, Vincent Weistroffer, Jean-Baptiste MouCEA-List,CEA LIST,Inri Modeling and Simu
Kaibo He, Chenhui Zuo, Jing Shao, Yanan Sui Tsinghua University Modeling and Simu
Yuhang Chen, Chenxing Wang Southeast University Modeling and Simu
Amedeo Ceglia, François Bailly, Mickael Begon University of Montreal,I Modeling and Simu
Bowen Yi, Yeman Fan, Dikai Liu Polytechnique Montreal,Modeling, Control,
Panagiotis Vartholomeos, Zicong Wu, S.m.hadi Sadati,University of Thessaly, Modeling, Control,
Junyi Shen, Tetsuro Miyazaki, Shingo Ohno, Maina So The University of TokyoModeling, Control,
Zhanchi Wang, Vector G.T.Wang, Xiaoping Chen, NikolUniversity of Science a Modeling, Control,
Hanna Zhang, Matthew Giamou, Filip Maric, Jonathan K University of Toronto,UnModeling, Control,
David Null, William Edwards, Dohun Jeong, Teodor Tc University of Illinois U Modeling, Control,
Jiaming Zhang, Zhaomeng Zhang, Yihao Liu, Yaqian Johns Hopkins UniversitModeling, Control,
Jingting Zhang, Xiaotian Chen, Paolo Stegagno, MingxUniversity of Electronic Modeling, Control,
Michele Pierallini, Francesco Stella, Franco Angelini, Centro di Ricerca E. PiaModeling, Control,
Yixin Wang, Xin-Jun Liu, Huichan Zhao Tsinghua University Modeling, Control, a
Phillip Maximilian Mehl, Max Bartholdt, Simon Ehlers, Leibniz University Hanno Modeling, Control, a
Diego Bianchi, Michele Gabrio Ernesto Antonelli, Cecili Scuola Superiore Sant'An Modeling, Control, a
James Bern, William May, Austin Osborn, Francesco StWilliams College,EPFL,U Modeling, Control, a
Etienne Ménager, Quentin Peyron, Christian Duriez Univ. Lille, Inria, CNRS Modeling, Control, a
Zicong Wu, S.m.hadi Sadati, Kawal Rhode, Christos BeKing's College London Modeling, Control, a
Uksang Yoo, Ziven Lopez, Jeffrey Ichnowski, Jean Oh Carnegie Mellon UniversModeling, Control, a
Siming Deng, Junning Liu, Bibekananda Datta, AishwarJohns Hopkins Universi Modeling, Control, a
Zach Patterson, Cosimo Della Santina, Daniela Rus MIT,TU Delft Modeling, Control, a
Nikolas Jakob Wilhelm, Claudio Glowalla, Sami Haddadi Technical University o Motion Analysis an
Jiacheng Weng, Ehsan Hashemi, Arash Arami University of Waterloo,UMotion Analysis an
Paul Nadan, Spencer Backus, Aaron Johnson Carnegie Mellon UniversMotion Analysis an
Jichun Xiao, Lina Hao, Hongzhi Xu, Xu Zhang, Xing Li, Northeastern University,Motion Analysis an
Brian Bittner, Jason Reid, Kevin C. Wolfe JHUAPL,Jet Propulsion Motion
L Analysis an
Yanhao Yang, Capprin Bass, Ross Hatton Oregon State UniversityMotion Analysis an
Alessio De Luca, Luca Muratore, Nikos Tsagarakis Istituto Italiano di TecnoMotion and Path Pla
Yue Meng, Chuchu Fan Massachusetts InstituteMotion and Path Pla
Ju Heon Lee, Chanyeol Yoo, Stuart Anstee, Robert Fit University of Technolo Motion and Path Pla
Matej Novosad, Robert Pěnička, Vojtech Vonasek Faculty of Electrical En Motion and Path Pla
Kohei Honda, Naoki Akai, Kosuke Suzuki, Mizuho Aoki,Nagoya University Motion and Path Pla
Ibrahim Ibrahim, Joris Gillis, Wilm Decré, Jan Swevers KU Leuven,Katholieke Un Motion and Path Pla
Minhyeong Lee, Dongjun Lee Seoul National UniversitMotion and Path Pla
Teng Guo, Jingjin Yu Rutgers University Motion and Path Pla
Adam Seewald, Cameron Lerch, Marvin Chancán, Aaron Yale University Motion and Path Pla
Fang Chen, Yu Zheng, Zheng Wang, Wanchao Chi, SicSouthern University of Motion and Path Pla
Yifan Wang, Yue Chen Georgia Institute of Tec Motion and Path Pla
Javier Garcia Gonzalez, Mike Yanuzzi, Peter Kramer, CUniversity of Houston, Motion and Path Pla
Zhe Huang, Hongyu Chen, John Pohovey, Katherine DrUniversity of Illinois a Motion and Path Pla
Wil Thomason, Zachary Kingston, Lydia Kavraki Rice University Motion and Path Pla
Brennan Brodt, Alyssa Pierson Boston University Motion and Path Pla
Apan Dastider, Hao Fang, Lin Mingjie University of Central FloMotion and Path Pla
Vishnu S. Chipade, Rahul Kumar, Sze Zheng Yong University of Michigan,NMotion and Path Pla
Thomas Manzini, Robin Murphy Texas A&M Motion and Path Pla
Itamar Mishani, Hayden Feddock, Maxim Likhachev Carnegie Mellon Universi Motion and Path Pla
Kasun Weerakoon, Adarsh Jagan Sathyamoorthy, MohUniversity of Maryland, Motion and Path Pla
Kallol Saha, Vishal Reddy Mandadi, Jayaram Gurram, A International Instititut Motion and Path Pla
Peter Werner, Alexandre Amice, Tobia Marcucci, Dani Massachusetts InstituteMotion and Path Pla
Nicholas Stiffler, Jason O'kane University of Dayton,Te Motion and Path Pla
Yong Li, Hui Cheng Guangzhou Shiyuan Elect Motion and Path Pla
Sihui Li, Neil Dantam Colorado School of Min Motion and Path Pla
Thibault Noël, Antoine Lehuger Lehuger, Eric Marchan INRIA Rennes,Groupe Cré Motion and Path Pla
Changan Chen, Jonas Frey, Philip Arm, Marco Hutter ETH Zurich Motion and Path Pla
Tilemahos Mitroudas, Vasiliki Balaska, Athanasios PsoDemocritus University oMotion Control and
Chao Qin, Maxime Simon Joseph Michet, Jingxiang ChUniversity of Toronto Motion Control and
Yifei Shao, Yuwei Wu, Laura Jarin-Lipschitz, Pratik Ch University of Pennsylva Motion Control and
Haojia Li, Haokun Wang, Chen Feng, Fei Gao, Boyu ZhThe Hong Kong Universit Motion Control and
Huan Yu, Jie Tu, Pengqin Wang, Zhi Zheng, Kewen ZhZhejiang University,TheMotion Control and
Fotis Panetsos, George Karras, Kostas Kyriakopoulos National Technical UniveMotion Control and
Frantisek Nekovar, Jan Faigl, Martin Saska Czech Technical UniversMotion Control and
Pratik Prajapati, Vineet Vashista Indian Institute of Tec Motion Control and
Marios-nektarios Stamatopoulos, Avijit Banerjee, Geor Luleå University of Tec Motion Control and
Jiqian Xu, Huaizhen Wang, Qiankun Zhao, Yue Gao, YiNortheastern UniversityMotion Control I
Baozhe Zhang, Xinwei Chen, Zhehan Li, Giovanni Belt The Chinese University Motion Control I
Chenxu Ke, Kai-yuan Cai, Quan Quan Beihang University,BeijiMotion Control I
Jon Arrizabalaga, Lukas Pries, Riddhiman Laha, RunkaTechnical University of Motion Control I
Christopher Rother, Zhaodong Zhou, Jun Chen Oakland University,oaklMotion Control I
Han Wang, Chao Ning, Longyan Li, Weidong Zhang Shanghai Jiao Tong Univ Motion Control I
Vincent Fortineau, Vincent Padois, David Daney Inria, Talence, France,InMotion Control I
Wuichung Cheng, Ngo Foon Chan, Darwin Lau The Chinese University Motion Control I
Guanwei He, Guodong Feng, Beichen Ding Shenzhen Campus of Sun Motion Control I
Jeeseop Kim, Jaemin Lee, Aaron Ames Caltech,California Insti Motion Control II
Jaemin Lee, Jeeseop Kim, Wyatt Ubellacker, Tamas GCalifornia Institute of Motion Control II
Long Jin, Kun Liu, Mei Liu Lanzhou University Motion Control II
Christopher König, Miks Ozols, Anastasiia Makarova, EInspire AG,ETH Zurich,ZMotion Control II
Zhongcheng Wang, Shiwei Yuan, Manfen Dou, JianhuaNorthwestern PolytechniMotion Control II
Xinyu Jia, Jun Yang, Lu Kaixin, Yongping Pan, HaoyonNational University of S Motion Control II
Yazan Al-rawashdeh, Mohammad Al Saaideh, Marcel HMemorial University of Motion Control II
Jiwan Han, Daehyung Park, Min Jun Kim Korea Advanced Institu Motion Control II
Xilin Xiao, Xiaojian Li, Shi Yudong, Jin Fang, Ling Li, Hefei University of Tec Motion Control II
Maged Iskandar, Christian Ott, Alin Albu-Schäffer, BrunGerman Aerospace Cente Motion Control III
Diletta Sacerdoti, Federico Benzi, Cristian Secchi University of Modena anMotion Control III
Wenshuo Li, Wenfu Xu, Peisheng Huang, Boyang Lin, B Harbin Institute of Tech Motion Control III
Yu-Ming Chen, Hien Bui, Michael Posa University of Pennsylva Motion Control III
Zehao Wang, Han Zhang, Jingchuan Wang ShanghaiJiaoTong Univer Motion Control III
Takumi Hachimine, Jun Morimoto, Takamitsu MatsubarNara Institute of Scie Motion Control III
Wei-cheng Huang, Alp Aydinoglu, Wanxin Jin, Michael University of Pennsylva Motion Control III
Dominic Baril, Simon-pierre Deschênes, Luc Coupal, Cyr Université Laval,EPFL Motion Control III
Alin Albu-Schäffer, Arne Sachtler DLR - German Aerospace Motion Control III
Servet Bora Bayraktar, Andreas Orthey, Zachary KingstTU Berlin,Realtime Robot Motion Planning I
Michael Lanighan, Oscar Youngquist TRACLabs, Inc,UniversiMotion Planning I
Kjell Keune, Aaron T. Becker Technische Universität Motion Planning I
Mingrui Luo, Yunong Tian, Yinghua Cao, Minghao ChenInstitute of Automation Motion Planning I
Alexandre Miranda Anon, Sangjae Bae, Manish Saroya,Honda Research InstitutMotion Planning I
Yeping Wang, Carter Sifferman, Michael Gleicher University of Wisconsin Motion Planning I
Halil Utku Unlu, Vinicius Mariano Gonçalves, Dimitris New York University,New Motion Planning I
Ruiqi Ni, Ahmed H. Qureshi Purdue University Motion Planning I
Himani Sinhmar, Marcus Carl Greiff, Stefano Di Cairan Cornell University,MitsuMotion Planning I
Wenxi Wu, Fabio Pierazzi, Yali Du, Martim Brandao King's College London Motion Planning II
Kaleb Ben Naveed, Devansh Agrawal, Christopher Verm University of Michigan, Motion Planning II
Dror Livnat, Michael Moshe Bilevich, Dan Halperin Tel Aviv University Motion Planning II
Jiangpin Liu, Tong Yang, Wangtao Lu, Yue Wang, RonZhejiang University Motion Planning II
Mingxin Yu, Chenning Yu, Mohammad Mahdi Naddaf SMassachusetts Institut Motion Planning II
Quang-Nam Nguyen, Quang-Cuong Pham Nanyang Technological Motion Planning II
Jacob Johnson, Ahmed H. Qureshi, Michael Yip UCSD,Purdue University, Motion Planning II
Edo Jelavic, Kaixian Qu, Farbod Farshidian, Marco HutSwiss Federal Institute Motion Planning II
Rongchuan Wang, Mengyin Fu, Jing Yu, Yi Yang, WenjBeijing Institute of Tech Motion Planning II
Mengyu Yang, Patrick Grady, Samarth Brahmbhatt, Ar Georgia Institute of Tec Multi-Modal Percept
Simone Arreghini, Gabriele Abbate, Alessandro Giusti, IADSIA USI-SUPSI,Istituto Multi-Modal Percept
Alexandra Esperança Gonçalves, Plinio Moreno, Jodi Fo Instituto Superior Técni Multi-Modal Percept
Salih Ertug Ovur, Yiannis Demiris Imperial College LondonMulti-Modal Percept
Weiqin Zu, Wenbin Song, Ruiqing Chen, Ze Guo, FanglShanghaiTech University Multi-Modal Percept
Shilong Mu, Runze Zhao, Zenan Zenanlin, Yan Huang, Tsinghua University,Wuh Multi-Modal Percept
Haowen Liu, Fengxian Wu, Bin Zhong, Yijun Zhao, Jia Southern University of SMulti-Modal Perception for HRI I
Guwen Han, Qi Ye, Anjun Chen, Jiming Chen Zhejiang university,ZhejMulti-Modal Percept
Syed Waleed Hyder, Muhammad Usama, Anas Zafar, Mu Retrocausal,Retrocausal, Multi-Modal Percept
Christopher Yee Wong, Lucas Vergez, Wael Suleiman McGill University,Arts e Multi-Modal Percepti
Jie Xu, Hanbo Zhang, Qingyi Si, Yifeng Li, Xuguang L Xi'an Jiaotong Univers Multi-Modal Percepti
Yinpei Dai, Run Peng, Sikai Li, Joyce Chai University of Michigan,UMulti-Modal Percepti
Gi-cheon Kang, Junghyun Kim, Jaein Kim, Byoung-TakSeoul National UniversitMulti-Modal Percepti
Haofeng Chen, Xuanxuan Yang, Gang Ma, Yucheng Wa University of Science a Multi-Modal Percepti
Jeongeun Park, Seungwon Lim, Joonhyung Lee, SangbKorea University,YonseiMulti-Modal Percepti
Fei Tang, Chuanxiong Zheng, Hongqi Yu, Lei Zhang, EOcean University of ChiMulti-Modal Percepti
Olmo Alonso Moreno Franco, Raajshekhar Parameswari Istituto
N Italiano di TecnoMulti-Modal Percepti
Zhuqing Zhang, Dongxuan Li, Jiayao Ma, Yijia He, PanZhejiang University,PekMulti-Modal Percepti
Yingjun Fan, Chuanbeibei Shi, Ganghua Lai, Ruiheng Beijing Institute of Tec Multi-Robot SLAM
Yaojie Zhang, Haowen Luo, Weijun Wang, Wei Feng Shenzhen Institute of Multi-Robot SLAM
Pierre-yves Lajoie, Giovanni Beltrame École Polytechnique de Multi-Robot SLAM
Yizhu Zhang, Shuhui Bu, Yifei Dong, Zhang Yu, Kun Li,Northwestern PolytechniMulti-Robot SLAM
Shipeng Zhong, Hongbo Chen, Yuhua Qi, Dapeng Feng, SUN YAT-SEN UNIVERSITY Multi-Robot SLAM
Jinxin Liu, Guoqiang Hu Nanyang Technological Multi-Robot
U SLAM
Minwoo Jung, Sangwoo Jung, Ayoung Kim Seoul National UniversitMulti-Robot SLAM
Peng Yin, Shiqi Zhao, Haowen Lai, Ruohai Ge, Ji ZhanCarnegie Mellon Universi Multi-Robot SLAM
Shaohang Xu, Yi'an Wang, Wentao Zhang, Chin Pang H uazhong University of Multi-Robot System
Williard Joshua Jose, Hao Zhang University of MassachusMulti-Robot System
Shuaikang Wang, Yiannis Kantaros, Meng Guo Peking University,Washin Multi-Robot System
Jonas Kuckling, Robin Luckey, Viktor Avrutin, Andrew University of Konstanz,IMulti-Robot System
Daniel Mcgann, Kyle Lassak, Michael Kaess Carnegie Mellon Universi Multi-Robot System
Jingcheng Jiang, Haiyin Piao, Yu Fu, Yihang Hao, ChuaDalian University of Te Multi-Robot Systems
Feiqiang Lin, Changyun Wei, Raphael Grech, Ze Ji Cardiff University,Hoha Multi-Robot Systems
Mandi Zhao, Shreeya Jain, Shuran Song Stanford University,Col Multi-Robot Systems
Zhehui Huang, Zhaojing Yang, Rahul Krupani, Baskin University of Southern Multi-Robot Systems
Hongda Jia, Zijian Gao, Cheng Yang, Bo Ding, YuanzhNational University of Multi-Robot Systems
Andrea Henshall, Sertac Karaman MIT,Massachusetts InstiMulti-Robot Systems
Kangyao Huang, Di Guo, Xinyu Zhang, Xiangyang Ji, HTsinghua University,BeiMulti-Robot Systems
Haruyuki Nakagawa, Asako Kanezaki Tokyo Institute of Tech Multi-Robot Systems
Mattia Catellani, Lorenzo Sabattini University of Modena anMulti-Robot Systems
Mela Coffey, Alyssa Pierson Boston University Multi-Robot Systems
Riku Murai, Joseph Ortiz, Sajad Saeedi, Paul H J Kell Imperial College LondonMulti-Robot Systems
Siji Chen, Yanshen Sun, Peihan Li, Lifeng Zhou, ChangVirginia Tech,Drexel UniMulti-Robot Systems
Xiangyong Wen, Yingjian Wang, Xi Zheng, Kaiwei WanZhejiang University,TheMulti-Robot Systems
Fernando Cladera, Zachary Ravichandran, Ian DouglasUniversity
M of Pennsylva Multi-Robot Systems
Nare Karapetyan, Ahmad Bilal Asghar, Amisha BhaskaWoods Hole Oceanographic Multi-Robot Systems
Tyler Hummer, Sam Kriegman Northwestern UniversityMulti-Robot Systems
Ming Cheng, Hao Liu, Deyuan Liu, Haibo Gu, Xiangke Beihang University,Nati Multi-Robot Systems
Georg Heppner, David Oberacker, Arne Roennau, RüdiFZI ForschungszentrumMulti-Robot I Systems
Viet-anh Le, Vaishnav Tadiparthi, Behdad Chalaki, HosUniversity of Delaware, Multi-Robot Systems
Yixiao Zhang, Victor Nan Fernandez-Ayala, Dimos V. KTH Royal Institute of Multi-Robot Systems
James Motes, Tan Chen, Timothy Bretl, Marco Morale University of Illinois U Multi-Robot Systems
Mark Mercier, David Curtis, Clark Taylor Air Force Institute of T Multi-Robot Systems
Akseli Kangaslahti, Alberto Candela, Jason Swope, Qi University of Michigan,NMulti-Robot Systems
Nikolaos Stathoulopoulos, Mario Alberto Valdes Sauce Luleå University of Tec Multi-Robot Systems
Koen de Vos, Elena Torta, Herman Bruyninckx, Cesar Eindhoven University o Multi-Robot System
Junfeng Chen, Yuan Gao, Junjie Hu, Fuqin Deng, Tin Peking University,Shenzh Multi-Robot System
Siyuan Wu, Gang Chen, Moji Shi, Javier Alonso-Mora Delft University of Tech Multi-Robot System
Rhett Hull, Diluka Prasanjith Moratuwage, Emily Schei University of Technolog Multi-Robot System
Kamran Vakil, Mela Coffey, Alyssa Pierson Boston University Multi-Robot System
Arundhati Banerjee, Jeff Schneider Carnegie Mellon UniversMulti-Robot System
Yewei Huang, Xi Lin, Brendan Englot Stevens Institute of Te Multi-Robot System
Álvaro Calvo, Jesus Capitan University of Seville Multi-Robot System
Steve Paul, Nathan Maurer, Souma Chowdhury University of ConnecticutMulti-Robot System
Raymond Leung, Andrew John Hill, Arman MelkumyanThe University of SydneMulti-Robot System
Federico Pratissoli, Riccardo Brugioni, Nicola Battilani, Università degli studi Multi-Robot System
Liwang Zhang, Minglong Li, Wenjing Yang, Shaowu YaNational University of Multi-Robot System
Lorenzo Balandi, Nicola De Carli, Paolo Robuffo Giord Centre Inria de l'Univ Multi-Robot System
Julian Erskine, Sébastien Briot, Isabelle Fantoni, Abde Ecole Centrale de Nant Multi-Robot System
Hongliang Guo, Qi Kang, Wei Yun Yau, Marcelo Ang, DAgency for Science Tech Multi-Robot System
Jesús Bautista, Hector Garcia De Marina University of Granada, Multi-Robot System
Darren Chiu, Radhika Nagpal, Bahar Haghighat University of Southern CMulti-Robot System
Shambhavi Singh, Zejian Huang, Akshaya Ks, GeordanCarnegie Mellon UniversMulti-Robot System
Wei Xu, Weichao Guo, Xu Shi, Xinjun Sheng, Xiangya Shanghai Jiao Tong Univ Multifingered Hand
Che-ming Chang, Felipe Sanches, Geng Gao, Minas LiUniversity of Auckland, Multifingered Hand
Sean Taylor, Kyungseo Park, Sankalp Yamsani, JoohyUniversity of Illinois Multifingered Hand
Haoming Li, Qi Ye, Yuchi Huo, Qingtao Liu, Shijian Ji Zhejiang University Multifingered Hand
Dehao Wei, Huazhe Xu Tsinghua University Multifingered Hand
Aoi Nakane, Iori Yanokura, Shun Hasegawa, Naoya YaThe Univeersity of TokyMultifingered Hand
Jana Egli, Benedek Forrai, Thomas Jakob Konrad BuchETHZ,ETH Zürich,ETH Zu Multifingered Hand
Qiang Wang, Robert Mccarthy, David Cordova Bulens, University College DubliMultifingered Hand
Kohei Shibata, Hiroki Dobashi Wakayama University Multifingered Hand
Jianing Ding, Hai-Tao Zhang, Binbin Hu Huazhong University of Networked and Coo
Nicola De Carli, Paolo Salaris, Paolo Robuffo GiordanoCNRS,University of Pis Networked and Coo
Manao Machida, Masumi Ichien NEC,NEC Corporation Networked and Coo
Aalok Patwardhan, Andrew J Davison Imperial College LondonNetworked and Coo
Gerasimos Damigos, Akshit Saradagi, Sara Sandberg, Ericsson AB - Ericsson Networked and Coo
Muhammad Zahid, Florian T. Pokorny KTH Royal Institute of Networked and Coo
Kaiyuan Eric Chen, Kush Hari, Rohil Khare, Charlotte LUniversity of California Networked and Coo
Daniel Mox, Kashish Garg, Alejandro Ribeiro, Vijay Ku University of Pennsylva Networked and Coo
Fabian Jakob, Xiao Chen, Hamid Sadeghian, Sami HadTechnical University of Networked and Coo
Matteo Bortolon, Theodore Tsesmelis, Stuart James, Fa Istituto Italiano di Tec Object Detection a
Tao Ma, Zhiwei Zheng, Hongbin Zhou, Xinyu Cai, XuemShanghai AI Laboratory,Object Detection a
Yinbao Li, Songshan Yu, Wang Dongfeng, Jingen Jiao Jiaxing Joospeed ElectrObject Detection a
Chaoqun Wang, Yiran Qin, Zijian Kang, Ningning Ma, The Chinese University Object Detection a
Minsik Jeon, Junwon Seo, Jihong Min Agency for Defense De Object Detection a
Fabian Duffhauss, Sebastian Koch, Hanna Ziesche, N Bosch Center for Artifi Object Detection a
Alex Christopher Stutts, Danilo Erricolo, Sathya Ravi, University of Illinois Ch Object Detection a
Jiaxin Wei, Xibin Song, Weizhe Liu, Laurent Kneip, HonTechnical University of Object Detection a
Jia Li, Li Jin, Xibin Song, Yeheng Chen, Nan Li, Xueyi Shandong University,ShObject Detection a
Chihiro Noguchi, Toshiaki Ohgushi, Masao Yamanaka Toyota Motor Corpo Object Detection I
Zhaoqi Leng, Pei Sun, Tong He, Dragomir Anguelov, MWaymo LLC,Waymo,WaObject Detection I
Lei Li, Alexander Liniger, Mario Millhaeusler, Vagia Ts ETH Zurich,Huawei Zur Object Detection I
Manli Shu, Le Xue, Ning Yu, Roberto Martín-martín, C University of Maryland, Object Detection I
Xirui Li, Feng Wang, Naiyan Wang, Chao Ma Shanghai Jiao Tong UniObject Detection I
Konyul Park, Yecheol Kim, Junho Koh, Byungwoo ParkHanyang University,Seou Object Detection I
Jiaxin Cai, Qi Li, Yulin Shen, Jia Pan, Wenxi Liu Fuzhou University,UniveObject Detection I
Chen Zhili, Trung Kien Pham, Maosheng Ye, Zhiqiang Hong Kong University oObject Detection I
Xinyang Huang, Guanghui Zhang, Dongchen Zhu, Yunpen Shanghai Institute of M Object Detection I
Jianning Deng, King Wah Gabriel Chan, Hantao Zhong,University of Edinburgh Object Detection II
Nicolas Chapman, Feras Dayoub, Will Browne, ChristoQueensland University oObject Detection II
Qi Jiang, Hao Sun Shanghai Jiaotong Unive Object Detection II
Seokha Moon, Hongbeen Park, Jaekoo Lee, Jinkyu Ki Korea University,KookmiObject Detection II
Shiyi Tang, Shu Zhang, Yini Fang Ocean University of ChiObject Detection II
Chen Jiang, Yuchen Yang, Martin Jagersand University of Alberta Object Detection II
Sairam Rebbapragada, Pranoy Panda, Vineeth Balasu IIT Hyderabad,Fujitsu RObject Detection II
Yurong You, Cheng Perng Phoo, Carlos Diaz-Ruiz, KatiCornell University Object Detection II
Meghna Gummadi, Cassandra Kent, Karl Schmeckpeper University of PennsylvanObject Detection II
Howard H. Qian, Yangxiao Lu, Kejia Ren, Gaotian WanRice University,the UnivObject Detection III
Kota Kondo, Claudius Taroon Tewari, Mason B. PetersMassachusetts InstituteObject Detection III
Tuan Van Vo, Minh Nhat Vu, Baoru Huang, Toan Nguye FPT Software,TU Wien,Object
A Detection III
Helbert Paat, Qing Lian, Weilong Yao, Tong Zhang HKUST,Autowise.AI,Hong Object Detection III
Shutong Zhang, Yi-ling Qiao, Guanglei Zhu, Eric HeidenUniversity of Toronto,UnObject Detection III
Max Peter Ronecker, Markus Schratter, Lukas KuschniSETLabs Research GmbH Object Detection III
Shaoming Zhang, Tangjun Yao, Jianmei Wang, Tianti Tongji University Object Detection III
Lukas Maximilian Schmid, Olov Andersson, Aurelio SulsMassachusetts InstituteObject Detection III
Lin Duan, Tim Scargill, Ying Chen, Maria Gorlatova Duke University Object Detection III
Shiqi Tan, Hamidreza Fazlali, Richard Xu, Yuan Ren, BUniversity of Toronto, Object Detection IV
Matthias Zeller, Daniel Casado Herraez, Jens Behley, Mi CARIAD SE,University oObject Detection IV
Hongwei Liu, Jian Yang, Zhengyu Li, Ke Li, Jianzhan Fuzhou University,Infor Object Detection IV
Qibin He, Zhongyang Xiao, Ze Huang, Hongyuan Yuan,University of Chinese A Object Detection IV
David Williams, Matthew Gadd, Paul Newman, Daniele University of Oxford,OxfObject Detection IV
Shengkai Wu, Liangliang Ren, Linfeng Gao, Yupeng Li CVTE,Huazhong Universi Object Detection IV
Xunjie He, Yiming Li, Te Cui, Meiling Wang, Tong Liu, Beijing Institute of Tech Object Detection IV
Ruoyu Wang, Yanfei Xue, Bharath Surianarayanan, DoNew York University,InteObject Detection IV
Jaehyeok Lee, Jae-Hyeon Park, Dong Eui Chang Korea Advanced Institu Object Detection IV
Tsung Lin Tsou, Tsung-han Wu, Winston Hsu National Taiwan UniversObject Detection V
Maciej Wozniak, Viktor Karefjard, Marko Thiel, Patric JeKTH Royal Institute of Object Detection V
Bilal Hassan, Arjun Sharma, Nadya Abdel Madjid, MajidKhalifa University, Abu Object Detection V
Jisong Kim, Minjae Seong, Geonho Bang, Dongsuk KuHanyang university,Hany Object Detection V
Jing Gu Expleo Germany GmbHObject Detection V
Jingyu Song, Lingjun Zhao, Katherine A. Skinner University of Michigan,UObject Detection V
Ozan Unal, Dengxin Dai, Ali Tamer Unal, Luc Van GoolETH Zurich,Bosphorus U Object Detection V
Jiarong Wei, Yancong Lin, Holger Caesar Delft University of Tech Object Detection V
Jinge Si, Bin Li, Yongkang Xu, Liang Wang, Chench Beijing institute of tech Object Detection V
Ngoc Thinh Nguyen, Pranav Tej Gangavarapu, NicolasUniversity
M of Luebeck,UOptimization and Op
Jee-Eun Lee, Andrew Bylard, Zhouwen Sun, Luis SentiThe University of Texas Optimization and Op
Xueyi Bu, Brian Plancher Fu Foundation School ofOptimization and Op
Emre Adabag, Miloni Atal, William Gerard, Brian Planc Columbia University,BarOptimization and Op
Ziyu Zhou, Gang Wang, Sun Jian, Jikai Wang, Jie Che Beijing Institute of Tech Optimization and Op
Maxim Vochten, Ali Mousavi, Arno Verduyn, Tinne De La KU Leuven,Department Optimization
o and Op
Gerry Chen, Frank Dellaert, Seth Hutchinson Georgia Institute of Te Optimization and Op
Francesco Tassi, Arash Ajoudani Istituto Italiano di TecnoOptimization and Op
Thomas Lew, Riccardo Bonalli, Marco Pavone Stanford University,Lab Optimization and Op
Dennis Ossadnik, Vasilije Rakcevic, Mehmet Can Yild Technical University of Optimization and Op
Saroj Prasad Chhatoi, Michele Pierallini, Franco Angeli University of Pisa,CentroOptimization and Op
Armand Jordana, Sebastien Kleff, Justin Carpentier, NiNew York University,I Optimization and Op
Tobias Löw, Sylvain Calinon Idiap Research InstituteOptimization and Op
Michael Enqi Cao, Samuel Coogan Georgia Institute of Te Optimization and Op
Oswin So, Zachary Serlin, Makai Mann, Jake GonzalesMassachusetts InstituteOptimization and Op
Gennaro Notomista University of Waterloo Optimization and Op
Lujie Yang, Hongkai Dai, Alexandre Amice, Russ Tedr MIT,Toyota Research Ins Optimization and Op
Seo Wook Han, Maged Iskandar, Jinoh Lee, Min Jun KKAIST,German Aerospac Optimization and Op
Yuichiro Aoyama, Amin Haeri, Evangelos Theodorou Georgia Institute of Tec Optimization and Opt
Gianluigi Grandesso, Elisa Alboni, Gastone Pietro RosaUniversity of Trento,UniOptimization and Opt
Haoru Xue, Edward Zhu, John Dolan, Francesco BorrellCarnegie Mellon Universi Optimization and Opt
Hongyu Zhou, Yichen Song, Vasileios Tzoumas University of Michigan,BOptimization and Opt
Hae Yeon Park, Jung Hoon Kim POSTECH,Pohang Univer Optimization and Opt
Omar Qasem, Weinan Gao American International UOptimization and Opt
Federico Girlanda, Shivesh Kumar, Lasse Maywald, FraUniversity of Padua,DFKOptimization and Opt
Cem Bilaloglu, Tobias Löw, Sylvain Calinon Idiap Research InstituteOptimization and Opt
Arun Bishop, John Zhang, Swaminathan Gurumurthy, KCarnegie Mellon UniversOptimization and Opt
Sara Ayoubi, Ilija Hadzic, Lou Salaun, Antonio MassaroNokia Bell Labs Path Planning for M
Yuhan Zhao, Quanyan Zhu New York University Path Planning for M
Abhay Singh Bhadoriya, Christopher Montez, Sivaku Texas A&M University,TPath Planning for M
Ananya Rao, Guillaume Sartoretti, Howie Choset Carnegie Mellon UniversPath Planning for M
Weizheng Wang, Le Mao, Ruiqi Wang, Byung-Cheol MiPurdue University,Beiji Path Planning for M
Zhongxia Yan, Han Zheng, Cathy Wu Massachusetts InstitutePath Planning for M
Evangelos Psomiadis, Dipankar Maity, Panagiotis TsiotGeorgia Institute of Tec Path Planning for M
Minhyuk Park, Tsz-Chiu Au UNIST,Ulsan National InPath Planning for M
Kalvik Jakkala, Srinivas Akella University of North CarolPath Planning for M
Lukas Heuer, Luigi Palmieri, Anna Mannucci, Sven Ko Örebro University, Rob Path Planning for M
Jaesung Ryu, Youngjoon Kwon, Sangho Yoon, Kyungj Chung-Ang University Path Planning for M
Vivek Kantilal Adajania, Siqi Zhou, Arun Singh, Angela University of Toronto,TePath Planning for M
Ardalan Tajbakhsh, Lorenz Biegler, Aaron Johnson Carnegie Mellon UniversPath Planning for M
Akmaral Moldagalieva, Joaquim Ortiz De Haro, Marc T Technical University of Path Planning for M
Chengyang He, Tianze Yang, Tanishq Harish Duhan, Yu National University SingPath Planning for M
Ratijit Mitra, Indranil Saha IIT Kanpur Path Planning for M
Courtney McBeth, James Motes, Diane Uwacu, Marco University of Illinois U Path Planning for M
Jingtao Tang, Hang Ma Simon Fraser UniversityPath Planning for M
Martin Jacquet, Kostas Alexis NTNU,NTNU - Norwegian Perception and Aut
Fotis Panetsos, George Karras, Kostas Kyriakopoulos National Technical UnivePerception and Aut
Viet Duong Hoang, Frederik Falk Nyboe, Nicolaj Malle University of Southern Perception and Aut
Lu Junjie, Bailing Tian, Hongming Shen, Xuewei Zhang,Tianjin University,Nany Perception and Aut
Ivan David Changoluisa Caiza, Ana Milas, Marco AntonUniversity of Zagreb,UnPerception and Aut
QIRANUL SAADIYEAN, Samprithi S P, Suresh SundarIndian Institute of Scie Perception and Aut
Chuanbeibei Shi, Ganghua Lai, Yushu Yu, Mauro Bellon Univeristy of Bristol,Be Perception and Aut
Ashish Kumar, Jaesik Park, Laxmidhar Behera Indian Institute of Tech Perception and Aut
Ruoyu Wang, Zixuan Guo, Yizhou Chen, Xinyi Wang, The Chinese University Perception and Aut
Ryan Bena, Chongbo Zhao, Quan Nguyen University of Southern CPerception and Aut
Kshitij Goel, Wennie Tabib Carnegie Mellon UniversPerception and Aut
Yuezhan Tao, Eran Iceland, Beiming Li, Elchanan ZwecUniversity of Pennsylva Perception and Aut
Chen Feng, Haojia Li, Mingjie Zhang, Xinyi Chen, Boy Hong Kong University ofPerception and Aut
Seungchan Kim, Micah Corah, John Keller, Graeme Bes Carnegie Mellon UniversPerception and Aut
Luca Bartolomei, Lucas Teixeira, Margarita Chli ETH Zurich,ETH Zurich Perception
& and Aut
Mihir Kulkarni, Kostas Alexis NTNU: Norwegian Univer Perception and Aut
Yuhan Xie, Minghao Lu, Rui Peng, Peng Lu The University of Hong Perception and Aut
Siqi Zhou, Sotiris Papatheodorou, Stefan Leutenegger, Technical University of Perception and Aut
Teresa Kent, Sarah Bergbreiter Carnegie Mellon UniversPerception and Aut
Zhefan Xu, Xiaoyang Zhan, Yumeng Xiu, Christopher SCarnegie Mellon UniversPerception and Aut
Xinyi Chen, Yichen Zhang, Boyu Zhou, Shaojie Shen The Hong Kong UniversiPerception and Aut
Marsela Polic, Barbara Arbanas, Jelena Vuletic, Matko University of Zagreb,UniPerception and Aut
Christian Gall, Walter Fichter, Aamir Ahmad University of Stuttgart Perception and Aut
Xunkuai Zhou, Benyun Zhao, Guidong Yang, Jihan Zhan Tongji University,The C Perception and Aut
Nathaniel Glaser, Rajashree Ravi, Zsolt Kira Georgia Institute of Te Perception and Aut
Lei Zhou, Haozhe Wang, Zhengshen Zhang, Zhiyang LiNational University of Perception for Gras
Peng Lu, Jiaming Liang, Bidan Huang, Sicheng Yang, Tencent Perception for Gras
Emmanouil Papadakis, Markos Sigalas, Michail VangosFoundation for ResearcPerception for Gras
Haoxiang Ma, Ran Qin, Modi Shi, Boyang Gao, Di HuaBeihang University,Geom Perception for Gras
Bart Duisterhof, Yuemin Mao, Si Heng Teng, Jeffrey Ic Carnegie Mellon Univer Perception for Gras
Hyelim Choi, Minji Lee, Jiseock Kang, Dongjun Lee Seoul National UniversitPerception for Gras
Hung-Jui Huang, Jingyi Xiang, Wenzhen Yuan Carnegie Mellon Universi Perception for Gras
Hongsuk Choi, Nikhil Chavan-dafle, Jiacheng Yuan, VolSamsung AI Center, New Perception for Gras
Luuk Van Den Bent, Tomás Coleman, Robert BabuskaTechnical University DelPerception for Gras
Hongyu Li, Snehal Dikhale, Soshi Iba, Nawid Jamali Brown University,HondaPerception for Grasping and Manipu
Mingdi Niu, Zhenyu Lu, Lu Chen, Jing Yang, Chengua Shanxi University,BristoPerception for Gras
Bing Wu, Qian Liu Dalian University of TecPerception for Gras
Russell Buchanan, Adrian Röfer, Joao Moura, Abhinav University of Edinburgh,Perception for Gras
Dirk-jan Boonstra, Laurence Willemet, Jelle Luijkx, Mic Delft University of Tech Perception for Gras
Junyu Nan, Jessica Hodgins, Brian Okorn Carnegie Mellon UniversPerception for Gras
Fengjun Mu, Shixiang Sun, Rui Huang, Chaobin Zou, WUniversity of Electroni Perception for Gras
Andrea Monguzzi, Yiannis Karayiannidis, Paolo Rocco,Politecnico di Milano,LuPerception for Gras
Hao-shu Fang, Chenxi Wang, Hongjie Fang, Minghao Go Shanghai Jiao Tong Univ Perception for Gras
Xingxu Li, Nan Ma, Yiheng Han, Shun Yang, Siyi Zhen Beijing University of Te Perception for Gras
Elle Miller, Maximilian Durner, Matthias Humt, Gabrie University of Edinburg Perception for Gras
Junyi Zhu, Young Joong Lee, Yiyue Luo, Tianyu Xu, Cha Massachusetts InstitutePerception for Gras
Michael Murray, Abhishek Gupta, Maya Cakmak University of WashingtoPerception for Gras
Julius Arolovitch, Osher Azulay, Avishai Sintov Carnegie Mellon Universi Perception for Gras
Marwan Taher, Ignacio Alzugaray, Andrew J Davison Imperial College LondonPerception for Gras
Dehao Huang, Chao Tang, Hong Zhang Southern University of Perception for Gras
Arren Glover, Luna Gava, Zhichao Li, Chiara BartolozziIstituto Italiano di Tecn Perception for Gras
Kevin Galassi, Bingbing Wu, Julien Perez, Gianluca PalUniversità di Bologna Perception for Gras
Varun Agrawal, Frank Dellaert Georgia Institute of Te Performance Evalu
Dongsu Lee, Chanin Eom, Minhae Kwon Soongsil University Performance Evalu
Salvatore D'avella, Matteo Bianchi, Ashok Meenakshi SSant'Anna School of AdPerformance Evalu
Ninad Khargonkar, Sai Haneesh Allu, Yangxiao Lu, Ji University of Texas at DPerformance Evalu
Changhe Chen, Mozhgan Pourkeshavarz, Amir RasouliUniversity of Toronto,R Performance Evalu
Wei-chih Hung, Vincent Casser, Henrik Kretzschmar, JWaymo Performance Evalu
Noe Perez-Higueras, Roberto Otero, Fernando CaballerUniversity Pablo de Ola Performance Evalu
Victor Mayoral-Vilches, Jason Jabbour, Yu-Shun HsiaoKlagenfurt University,HaPerformance Evalu
Irene Garcia-Camacho, Alberta Longhini, Michael Welle,Institut de Robòtica i Performance Evalu
Federico Benzi, Cristian Secchi University of Modena anPhysical Human-Robo
Katia Bourahmoune, Karlos Ishac, Marc Garry CarmichUniversity of Technolo Physical Human-Robo
Chenzui Li, Xi Wu, Tao Teng, Sylvain Calinon, Fei CheThe Chinese University P o hysical Human-Robo
Theodoros Stouraitis, Michael Gienger RoboPhren and Honda Re Physical Human-Robo
Mahdi Hejrati, Jouni Mattila Tampere University Physical Human-Robo
Stelios Kotsovolis, Yiannis Demiris Imperial College LondonPhysical Human-Robo
Priyanka Ramasamy, Gunarajulu Renganathan, Yuichi Hiroshima University Physical Human-Robo
S S Sanjeevi Nakka, Vineet Vashista Indian Institute of Tec Physical Human-Robo
Qiao Wang, Ziqi Wang, Marc Garry Carmichael, Dikai LiUniversity of Technolo Physical Human-Robo
Alessia Fusco, Valerio Modugno, Dimitrios Kanoulas, APolitecnico di Torino,U Physical Human-Robo
Sammy Christen, Lan Feng, Wei Yang, Yu-Wei Chao, Ot ETH Zurich,ETH ZURICPhysical Human-Robo
Zhanibek Rysbek, Siyu Li, Afagh Mehri Shervedani, MilUniversity of Illinois at Physical Human-Robo
Cheng Fang, Luka Peternel, Ajay Seth, Massimo SartoriUniversity of Southern DPhysical Human-Robo
Jose Enrique Dominguez-Vidal, Alberto Sanfeliu Institut de Robòtica i I Physical Human-Robo
Ziqi Wang, Marc Garry Carmichael University of Technolo Physical Human-Robo
Saad Yousaf, Gaurav Mukherjee, Raymond King, Ashi The University of TexasPhysical Human-Robo
Andrea Fortuna, Marta Lorenzini, Mattia Leonori, Juan Politecnico di Milano,Ist Physical Human-Robo
Jingtian Yan, Lin Xingqiao, Zhongqiang Ren, Shiqi ZhaCarnegie Mellon Universi Planning under Unce
Lars Lindemann, Matthew Cleaveland, Gihyun Shim, GUniversity of Southern CPlanning under Unce
Shaohang Xu, Haolin Ruan, Wentao Zhang, Yi'an WangHuazhong University of Planning under Unce
Ji Yin, Charles Dawson, Chuchu Fan, Panagiotis TsiotrGeorgia Institute of Te Planning under Unce
Sleiman Safaoui, Tyler Summers The University of Texas Planning under Unce
Moran Barenboim, Moshe Shienman, Vadim Indelman Technion - Israel Institu Planning under Unce
Moran Barenboim, Idan Lev Yehudi, Vadim Indelman Technion - Israel Instit Planning under Unce
Shili Sheng, David Parker, Lu Feng University of Virginia,UnPlanning under Unce
Yasmin Veys, Martina Stadler, Nicholas Roy Massachusetts InstitutePlanning under Unce
Carlos Quintero-Pena, Wil Thomason, Zachary Kingston, Rice University Planning under Unce
Arec Jamgochian, Hugo Buurmeijer, Kyle Wray, AnthonStanford University,n/a Planning under Unce
Jianxiong Shen, Ruijie Ren, Adrià Ruiz, Francesc Mor IRI, CSIC-UPC,IRI-CSI Planning under Unce
Marco Faroni, Dmitry Berenson Politecnico di Milano,UnPlanning under Unce
Emil Wiman, Ludvig Widén, Mattias Tiger, Fredrik HeintLinköping University,AI Planning under Unce
Jing Liang, Peng Gao, Xuesu Xiao, Adarsh Jagan Sat University of Maryland, Planning under Unce
Dongliang Zheng, Panagiotis Tsiotras Georgia Tech Planning under Unce
Kanghyun Ryu, Negar Mehr University of California, Planning under Unce
Nicholas Mohammad, Jacob Higgins, Nicola Bezzo University of Virginia Planning under Unce
Matt Schmittle, Rohan Baijal, Brian Hou, Siddhartha Sr University of WashingtoPlanning under Uncer
David Morilla Cabello, Jonas Westheider, Marija PopovUniversidad de Zaragoza Planning under Uncer
Gokul Puthumanaillam, Xiangyu Liu, Negar Mehr, MelkiUniversity of Illinois U Planning under Uncer
Samer Nashed, Rod Grupen, Shlomo Zilberstein University of MassachusPlanning under Uncer
Kangjie Zhou, Pengying Wu, Yao Su, Han Gao, Ji Ma, Peking University,BeijingPlanning
Instituteunder
for General
Uncer Artificial Intelli
Muhammad Suhail Saleem, Rishi Veerapaneni, Maxim Carnegie Mellon UniversPlanning under Uncer
Ralf Römer, Armin Lederer, Samuel Tesfazgi, Sandra HTechnical University of Planning under Uncer
Anne Theurkauf, Justin Kottinger, Nisar Ahmed, MortezUniversity of Colorado Planning under Uncer
Reiya Takemura, Genya Ishigami Keio University Planning under Uncer
Zhefan Xu, Kenji Shimada Carnegie Mellon UniversPlanning, Schedulin
Manshi Limbu, Zechen Hu, Xuan Wang, Daigo ShishikaGeorge Mason UniversitPlanning, Schedulin
Sang-hyun Cho, Woo-Jin Shin, Jeongsun Ahn, SanghyKorea Advanced Institu Planning, Schedulin
Tong Yang, Li Huang, Yue Wang, Rong Xiong Zhejiang University Planning, Schedulin
Teng Guo, Jingjin Yu Rutgers University Planning, Schedulin
Weiheng Dai, Aditya Bidwai, Guillaume Sartoretti National University of S Planning, Schedulin
Neel Dhanaraj, Jeon Ho Kang, Anirban Mukherjee, HerUniversity of Souther Planning, Schedulin
Xingyao Han, Siyuan Chen, Xinye Xiong, Qiming Liu, Shanghai Jiao Tong Univ Planning, Schedulin
Shengqiang Chen, Yiyu Chen, Ronak Jain, Xiaopan ZhUniversity of Southern CPlanning, Schedulin
Donggyu Jung, Shinsuk Park, Junho Choi Korea Institute of Scie Prosthetics and Exo
Bangyu Lan, Momen Abayazid, Nico Verdonschot, Stefa University of Twente, Prosthetics and Exo
Paniz Sedighi, Xingyu Li, Mahdi Tavakoli University of Alberta Prosthetics and Exo
Haochen Wei, Kai Yu Tong, Michael Yu Wang, Chao CMonash University,The Chinese Prosthetics
University
and Exoof Hong Kong,[em
Albert Lee, Curt A. Laubscher, T. Kevin Best, Robert D University of Michigan Prosthetics and Exo
Majun Song, Zhongyi Li, Weihai Chen, Hao Zheng, Sh Hangzhou Innovation Inst Prosthetics and Exo
Océane Dubois, Agnès Roby-Brami, Ross Parry, NathaSorbonne University,univ Prosthetics and Exo
Zhenzhi Ying, Xianyu Zhang, Shihao Li, Koki NakashimThe University of TokyoProsthetics and Exo
Shuo Ding, Francisco Anaya Reyes, Shounak BhattachNational University of S Prosthetics and Exo
Kyujin Hyeon, Chongyoung Chung, Jihyeong Ma, Ki-U KAIST,Korea AdvancedProstheticsIn and Exo
Miha Dezman, Charlotte Dorothea Marquardt, Tamim AKarlsruhe Institute of T Prosthetics and Exo
Jeongsu Park, Kyeongsu Shi, Hyojun An, Gunhee Lee KAIST,Korea AdvancedProstheticsIn and Exo
Zeynep Özge Orhan, Andrea Dal Prete, Anastasia BolotEPFL,Politecnico di Mil Prosthetics and Exo
Zeynep Özge Orhan, Milad Shafiee, Vincent Juillard, J EPFL Prosthetics and Exo
Gwendolyn Bryan, Patrick W. Franks, Seungmoon Song, IHMC,Skip,Northeastern, Prosthetics and Exo
David Kelly, Ryan Posh, Patrick Wensing University of Notre DamProsthetics and Exo
Francesco Missiroli, Pietro Mazzoni, Nicola Lotti, Enr Heidelberg University, Prosthetics and Exo
Conner Pulling, Je Hon Tan, Yaoyu Hu, Sebastian ScheCarnegie Mellon Univer Range Sensing
Eva Goichon, Guillaume Caron, Pascal Vasseur, FumioUPJV,CNRS,Université Range de Sensing
David Hunt, Shaocheng Luo, Amir Khazraei, Xiao Zhang, Duke University,Duke unRange Sensing
Georgi Pramatarov, Matthew Gadd, Paul Newman, Dani University of Oxford,OxfRange Sensing
Sehua Ji, Weinan Chen, Zerong Su, Yisheng Guan, Jieh Guangdong University ofRange Sensing
Jun Zhu, Hongyi Li, Zhepeng Wang, Shengjie Wang, T Tsinghua University,Le Range Sensing
Ruslan Agishev, Tomas Petricek, Karel Zimmermann Czech Technical UniversRange Sensing
Nicolai Steinke, Daniel Goehring, Raúl Rojas Freie Universität Berlin Range Sensing
Guoyu Lu University of Georgia Range Sensing
Victor Reijgwart, Michael Pantic, Roland Siegwart, Lione
ETH Zurich,ETH Zürich Reactive and Senso
Tanel Kossas, Walid Remmas, Roza Gkliva, Asko Rist Tallinn University of Te Reactive and Senso
Takato Miura, Naoki Akai, Kohei Honda, Susumu Hara Nagoya University Reactive and Senso
Felix Eberle, Riddhiman Laha, Haowen Yao, AbdeldjalliTechnical University of Reactive and Senso
Shaohang Xu, Wentao Zhang, Chin Pang Ho, Lijun ZhuHuazhong University of Reactive and Senso
Letian Ai, Yihao Liu, Mehran Armand, Amir KheradmanJohns Hopkins UniversitReactive and Senso
Jianfeng Gao, Zhi Tao, Noémie Jaquier, Tamim Asfour Karlsruhe Institute of T Reactive and Senso
Fabrice Zeug, Marvin Becker, Matthias A. Muller Gottfried Wilhelm Leibn Reactive and Senso
Hassan Jardali, Mahmoud Ali, Lantao Liu Indiana University Reactive and Senso
Baris Akbas, Aleyna Soylemez, Huseyin Taner Yuksel, Kadir Has University Rehabilitation Robo
Myles Cooper, Santiago Canete, Asa Eckert-Erdheim, Ai Harvard University,HarvRehabilitation Robo
Takuma Shimoyama, Tomoyuki Noda, Tatsuya Teramae The University of Elec Rehabilitation Robo
Rafael J. Escarabajal, José Luis Pulloquinga, Pau Zamo Universidad Politécnica Rehabilitation Robo
Raffaele Giannattasio, Stefano Maludrottu, Gaia Zinni, Italian Institute of Tech Rehabilitation Robo
Nat Wannawas, Clara Diaz-pintado, Jyotindra Narayan,Imperial College London, Rehabilitation Robo
Andreas Christou, Antonio J. Del-ama, Juan C. MorenoThe University of EdinbuRehabilitation Robo
Matthew Lamsey, Meredith Wells, You Liang Tan, MadeGeorgia Institute of Te Rehabilitation Robo
Chenyang Cao, Zichen Yan, Renhao Lu, Junbo Tan, X Tsinghua University,Cent Reinforcement Lear
Fan Yang, Wenxuan Zhou, Zuxin Liu, Ding Zhao, DavidUniversity of Michigan,CReinforcement Lear
Weijian Zhang, Jianshu Wang, Yu Yang Nanjing University,Nati Reinforcement Lear
Luigi Berducci, Shuo Yang, Rahul Mangharam, Radu GTU Wien,University of PReinforcement Lear
Nico Messikommer, Yunlong Song, Davide ScaramuzzUniversity of Zurich Reinforcement Lear
Homayoun Honari, Mehran Ghafarian Tamizi, HomayouUniversity of Victoria Reinforcement Lear
Lev Grossman, Brian Plancher Berkshire Grey,Barnard Reinforcement Lear
Josselin Somerville Roberts, Julia Di Stanford University Reinforcement Lear
Chengxin Li, Yihui Hu, Zewen Zheng, Xiang Gao, YongSouth China Normal UnReinforcement Lear
Ju-hyeok Ryu, Chan Kim, Seong-woo Kim Seoul National UniversitReinforcement Learn
Ya-Chien Chang, Sicun Gao University of Californi Reinforcement Learn
Miao Li, Wenhao Ding, Ding Zhao Carnegie Mellon UniversReinforcement Learn
Shaohao Zhu, Jiacheng Zhou, Anjun Chen, Mingming Ba Zhejiang University,CollReinforcement Learn
Yiwen Hou, Haoyuan Sun, Jinming Ma, Feng Wu University of Science a Reinforcement Learn
Hao-Lun Hsu, Haocheng Meng, Shaocheng Luo, Junchen Duke University,HarvardReinforcement Learn
Hyukjun Kwon, Kangwon Kim, Junyoung Lee, Hyunsei D L GIST,HYU(Hanyang UniverReinforcement Learn
Mingen Li, Changhyun Choi University of California Reinforcement Learn
Wataru Hatanaka, Ryota Yamashina, Takamitsu MatsuRicoh Company, Ltd.,R Reinforcement Learn
Gyan Tatiya, Jonathan Francis, Ho-hsiang Wu, YonatanTufts University,Bosch CRepresentation Lear
Hanwen Ren, Ahmed H. Qureshi Purdue University Representation Lear
Yuze He, Peng Wang, Yubin Hu, Wang Zhao, Ran Yi, Tsinghua University,TheRepresentation Lear
Dominik Joho, Jonas Schwinn, Kirill Safronov KUKA Deutschland GmbRepresentation Lear
Stephen Hausler, David Hall, Sutharsan Mahendren, CSIRO,Commonwealth Representation
Sci Lear
Junsheng Zhou, Xin Wen, Baorui Ma, Yu-shen Liu, YueTsinghua University,NVID Representation Lear
Junyao Shi, Jianing Qian, Yecheng Jason Ma, Dinesh University of Pennsylva Representation Lear
Chia-liang Kuo, Yu-Wei Chao, Yi-ting Chen National Yang Ming ChiRepresentation Lear
Peter So, Rafael Ignacio Cabral Muchacho, Robin KirscTechnical University of Representation Lear
Alireza Rezazadeh, Athreyi Badithela, Karthik Desing University of Minnesota,Representation Lear
Sneha Silwal, Karmesh Yadav, Tingfan Wu, Jay Vakil, Meta,Georgia Tech,Meta Representation Lear
Kartikeya Singh, Charuvahan Adhivarahan, Karthik Danuniversity at buffalo,Uni Representation Lear
Skand Peri, Iain Lee, Chanho Kim, Fuxin Li, Tucker HeOregon State University,Representation Lear
Vasileios Vasilopoulos, Suveer Garg, Jinwook Huh, BhoSamsung Research Ameri Representation Lear
Jianing Qian, Anastasios Panagopoulos, Dinesh Jayar University of Pennsylva Representation Lear
Jiankai Sun, Yan Xu, Mingyu Ding, Hongwei Yi, Chen Stanford University,TheRepresentation Learning II
Yinghan Sun, Linfang Zheng, Hua Chen, Wei Zhang Southern University of RGB-D Sensing and
Isaac Kasahara, Shubham Agrawal, Kazim Selim Engin, Samsung Research Ameri RGB-D Sensing and
Yuhao Yang, Jun Wu, Yue Wang, Mr Zhang Guangjian,Chongqing University ofRGB-D Sensing and
Ang Li, Anning Hu, Wei Xi, Danping Zou, Wenxian Yu Shanghai Jiao Tong Univ RGB-D Sensing and
Siddharth Tourani, Jayaram Gurram, Sarvesh Thakur IIIT Hyderabad,Internat RGB-D Sensing and
Shenghai Yuan, Yizhuo Yang, Thien Hoang Nguyen, Thie NANYANG TECHNOLOGICAL RGB-D Sensing and
Evin Pınar Örnek, Aravindhan Krishnan, Shreekant GTU Munich,Amazon Lab,RGB-D Sensing and
Yangxiao Lu, Yuqiao Chen, Nicholas Ruozzi, Yu Xiang the University of Texas RGB-D Sensing and
Helin Cao, Sven Behnke University of Bonn RGB-D Sensing and
Jiyuan Wang, Chunyu Lin, Lang Nie, Shujuan Huang, XBeijing JiaoTong UniverRGB-D Sensing and
Peng Gao, Yu Shen, Ming C. Lin University of MassachusRGB-D Sensing and
David Jin, Sushrut Karmalkar, Harry Zhang, Luca Carl MIT,University of WiscoRGB-D Sensing and
Zhaole Sun, Hang Zhou, Nanbo Li, Longfei Chen, JihonTsinghua University, theRGB-D Sensing and
Carter Sifferman, Yeping Wang, Mohit Gupta, Michael University of Wisconsin RGB-D Sensing and
Zhichao Deng, Xiangtai Li, Xia Li, Yunhai Tong, Shen Sun Yat-sen university, RGB-D Sensing and
Qiao Gu, Alihusein Kuwajerwala, Sacha Morin, KrishnaUniversity of Toronto,U RGB-D Sensing and
Jingyi Xiang, Holly Dinkel, Harry Zhao, Naixiang Gao, BUniversity of Illinois RGB-D Sensing and
Jean Lahoud, Fahad Khan, Hisham Cholakkal, Rao AnMBZUAI,Linkoping UnivRGB-D Sensing and
Dong-wook Yang, Hyun-su Park, Keon-ik Jang, Jae-h Korea Advanced Institu Robot Design
Crystal Winston, Leo Casey Stanford University,Goog Robot Design
Muhan Li, David Matthews, Sam Kriegman Northwestern UniversityRobot Design
Xuanchun Yin, Junliang Xie, Pengyu Zhou, Sheng Wen,South China AgriculturalRobot Design
Milad Asgari, Ilian Bonev, Clement Gosselin École de technologie suRobot Design
Robin Hall, Gabriel Espinosa, Shou-shan Chiang, Cag Worcester Polytechnic IRobot Design
Robin Hall, Cagdas Onal Worcester Polytechnic IRobot Design
Jehyeok Kim, Clement Gosselin Université Laval Robot Design
Azamat Yeshmukhametov, Koichi Koganezawa Nazarbayev University,TRobot Design
Sergey Levine, Chelsea Finn, Ken Goldberg, LawrenceUC Berkeley,Stanford Un Robot Manipulation
Isabel Leal, Krzysztof Choromanski, Deepali Jain, Av Google Deepmind,Google Robot Manipulation
Won Kyung Do, Ankush Dhawan, Mathilda Kitzmann, Stanford University Robot Manipulation
Jared Mejia, Victoria Dean, Tess Hellebrekers, Abhina Carnegie Mellon UniversRobot Manipulation
Homanga Bharadhwaj, Abhinav Gupta, Vikash Kumar, Carnegie Mellon UniversRobot Manipulation
Thomas Cohn, Seiji Shaw, Max Simchowitz, Russ TedrMassachusetts InstituteRobot Manipulation
Hongchao Zhang, Luyao Niu, Andrew Clark, Radha PoWashington University inRobot Safety I
Matti Vahs, Christian Pek, Jana Tumova KTH Royal Institute of Robot Safety I
Prashant Solanki, Jasper Van Beers, A. Jamshidnejad,Delft University of Tech Robot Safety I
Bolun Dai, Rooholla Khorrambakht, Prashanth Krishnam New York University,New Robot Safety I
Yutong Zhang, Fei Liu, Xiao Liang, Michael Yip University of California Robot Safety I
Nick Rober, Karan Mahesh, Tyler Paine, Max L. GreeneMassachusetts InstituteRobot Safety I
Bowen Weng, Guillermo Castillo, Yun-seok Kang, AyonIowa State University,ThRobot Safety I
Aryaman Gupta, Kaustav Chakraborty, Somil Bansal Indian Institute of Tech Robot Safety I
Ryan Cosner, Igor Sadalski, Jana Woo, Preston Culbe California Institute of T Robot Safety I
Dawei Sun, Benjamin Yang, Sayan Mitra UIUC,University of Illi Robot Safety II
Manas Sashank Juvvi, David Smith Sundarsingh, Ratn Indian Institute of Scien Robot Safety II
Gianni Lunardi, Asia La Rocca, Matteo Saveriano, AndrUniversity of Trento Robot Safety II
Asia La Rocca, Matteo Saveriano, Andrea Del Prete University of Trento Robot Safety II
Jingxing Qian, Siqi Zhou, Nicholas Ren, Veronica ChatrUniversity of Toronto,TeRobot Safety II
Alexandre St-Jean, Francis Dorval, Jean-Sebastien PlaUniversité de SherbrookRobot Safety II
Matti Vahs, Jana Tumova KTH Royal Institute of Robot Safety II
Jordan Lekeufack Sopze, Anastasios Angelopoulos, Andr University of California Robot Safety II
Neil Janwani, Ersin Das, Thomas Touma, Skylar Wei, TCalifornia Institute of Robot Safety II
Adam Misik, Driton Salihu, Xin Su, Heike Brock, Eckeh Siemens Technology, Tec Robot Vision
Siddharth Ancha, Philip Osteen, Nicholas Roy Massachusetts InstituteRobot Vision
Yuxiang Peng, Chuchu Chen, Guoquan Huang University of Delaware Robot Vision
Yu Lin, Zhiheng Li, Yubo Cui, Zheng Fang Northeastern UniversityRobot Vision
Yunxuan Mao, Xuan Yu, Zhuqing Zhang, Kai Wang, Yue Zhejiang University,Hu Robot Vision
Zichen Zhang, Yunshuang Li, Osbert Bastani, Abhish Allen Institute for AI,U Robot Vision
Alejandro Velasquez, Cindy Grimm, Joseph Davidson Oregon State UniversityRobotics and Automa
Yanjiang Huang, Jiepeng Liu, Xianmin Zhang Guangdong Provincial ke Robotics and Automa
Lei Fan, Dongdong Fan, Yiwen Ding, Yong Wu, Hongx The University of New Robotics and Automa
Chen Gao, Franz Daxinger, Lukas Roth, Fabiola Maffra,ETH Zürich,ETH Zurich,U Robotics and Automa
Jonathan Jaramillo, Aaron Wilhelm, Nils Napp, Justine Cornell University Robotics and Automa
Elias Ariel Marks, Matteo Sodano, Federico Magistri, University of Bonn,PhotRobotics and Automa
Chung Hee Kim, Moonyoung Lee, Oliver Kroemer, GeoCarnegie Mellon UniversRobotics and Automa
Yuseung Jo, Hyoung Il Son Chonnam National Unive Robotics and Automa
Gianmarco Roggiolani, Federico Magistri, Tiziano Guada University of Bonn Robotics and Automa
Amel Dechemi, Dimitrios Chatziparaschis, Joshua ChenUniversity of California, Robotics and Automa
Feng-li Lian, Lu-ching Wang, Yen-cheng Chu National Taiwan UniversRobotics and Automa
Jordan James, Heather K. Manching, Amanda M. HulseUniversity of Texas at ARobotics and Automa
Sharmistha B Pandey, David Colliaux, Ayan ChaudhuryISRO,Sony CSL Paris,Ind Robotics and Automa
Felix Esser, Radu Alexandru Rosu, Andre Cornelissen,University of Bonn Robotics and Automa
Jeonghyun Byun, Inkyu Jang, Dongjae Lee, H. Jin Kim Seoul National UniversitRobotics and Automa
Houman Masnavi, Jatan Shrestha, Karl Kruusamäe, ArInstitute of Technology, Robotics and Automa
Viswa Narayanan Sankaranarayanan, Akshit Saradagi,Lulea University of Tec Robotics and Automa
Adam Mukuddem, Paul Amayo University of Cape Tow Robotics and Automa
Mark Niemeyer, Julian Arkenau, Sebastian Pütz, JoachDFKI,German ResearchRobotics Ce and Automa
Qinghui Pan, Dong Wang, Jie Lian, Yongxiang Dong, Dalian University of Te Robotics and Automa
Masaki Asano, Takanori Fukao University of Tokyo, Gr Robotics and Automa
Roni Kasemi, Lara Lammer, Stefan Thalhammer, MarkUBT,ACIN TU WIEN,TURobotics W and Automa
Alexander You, Nidhi Parayil, Josyula Gopala Krishn Oregon State UniversityRobotics and Automa
Dimitrios Chatziparaschis, Hanzhe Teng, Yipeng Wang,UC Riverside,UniversityRobotics and Automa
Harry Freeman, George Kantor Carnegie Mellon UniversRobotics and Automa
Akshay Kumar Burusa, Eldert J. Van Henten, Gert KootWageningen University Robotics and Automa
Dasharadhan Mahalingam, Aditya Patankar, Khiem Phi,Stony Brook University Robotics and Automa
Julia Hindel, Nikhil Gosala, Kevin Bregler, Abhinav Val University of Freiburg,F Robotics and Automa
Yue Linn Chong, Jan Weyler, Philipp Lottes, Jens Behley University of Bonn Robotics and Automa
Elie Ayoub, Heshan Fernando, William Larrivée-hardy, FPInnovations,FP Innovat Robotics and Automa
Yu-hsiang Lai, Chia-yun Chuang, Yu-qiang Chen, Feng-National Taiwan universiRobotics and Automa
Meher Venkata Ramakrishna Malladi, Tiziano Guadagnino University of Bonn,Univ Robotics and Automa
Nicholas Harrison, Nathan Wallace, Salah Sukkarieh The University of Sydney Robotics and Automa
Dac Dang Khoa Nguyen, Gavin Paul, Alen Alempijevic University of Technolog Robotics and Automa
Sang-Ho Hyon, Ryo Ando, Eiji Sono, Shunichi Sugimoto Ritsumeikan University Robotics and Automa
Mads Alber Kuhlmann-jørgensen, Johannes Pankert, Lu ETH Zurich,ETH Zueric Robotics and Automa
Miana Smith, Amira Abdel-rahman, Neil Gershenfeld MIT,Massachusetts InstiRobotics and Automa
Joseph Rowell, Lintong Zhang, Maurice Fallon University of Oxford, OxRobotics and Automa
Samuel Lensgraf, Ankita Sarkar, Adithya Pediredla, DevDartmouth College Robotics and Automa
Daniel Eriksson, Reza Ghabcheloo, Geimer Marcus Tampere University,KarlRobotics and Automa
Takayuki Osa, Naoto Osajima, Masanori Aizawa, Tats University of Tokyo,KyuRobotics and Automa
Ejup Hoxha, Jinglun Feng, Diar Sanakov, Jizhong XiaoThe City College of NewRobotics and Automa
Chao Zhao, Shuai Yuan, Chunli Jiang, Junhao Cai, Ho Hong Kong University ofRobotics
Science with
and Larg
Technology,The Hong
An Dinh Vuong, Minh Nhat Vu, Le Trung Hieu, Baoru HFPT Software,TU Wien,Robotics
A with Larg
Raghav Arora, Shivam Singh, Karthik Swaminathan, AhIIIT Hyderabad,InternatiRobotics with Larg
K.r. Zentner, Ryan Julian, Brian Ichter, Gaurav Sukhat University of Southern Robotics with Larg
Lingfeng Sun, Devesh Jha, Chiori Hori, Siddarth Jain University of California Robotics with Larg
Zhirui Dai, Arash Asgharivaskasi, Thai Duong, Shusen UC San Diego,UniversityRobotics with Larg
Shreyas Sundara Raman, Vanya Cohen, Ifrah Idrees, EBrown University,UniverRobotics with Larg
Chao Tang, Dehao Huang, Wenqi Ge, Weiyu Liu, HongSouthern University of Robotics with Larg
Guo Ye, Qinjie Lin, Zening Luo, Han Liu Northwestern UniversityRobotics with Larg
Gustavo Olivas-martínez, Armando Miranda-moya, CarInstituto Tecnológico d Robust/Adaptive Co
Gan Yu, Joel Reis, Carlos Silvestre Shanghai Jiao Tong Univ Robust/Adaptive Co
Radosław Patelski, Dariusz Pazderski Poznan University of T Robust/Adaptive Co
Qie Sima, Yu Luo, Tianying Ji, Fuchun Sun, Huaping LiTsinghua University,UniRobust/Adaptive Co
Annie Mao, Joseph Moore, Louis Whitcomb Johns Hopkins UniversitRobust/Adaptive Co
Dvij Kalaria, Qin Lin, John Dolan Carnegie Mellon UniversRobust/Adaptive Co
Armand Jordana, Avadesh Meduri, Etienne Arlaud, Justi New York University,IN Robust/Adaptive Co
Bowen Su, Fan Zhang, Panfeng Huang Northwestern PolytechniRobust/Adaptive Co
Mohamed Djeha, Pierre Gergondet, Abderrahmane Kh Université de Montpel Robust/Adaptive Co
Jon Škerlj, Mazin Hamad, Jean Elsner, Abdeldjallil NacTechnical University of Safety in HRI
Zheng Shen, Matteo Saveriano, Fares Abu-Dakka, SamTU Munich,University ofSafety in HRI
Shohei Fujii, Quang-Cuong Pham Nanyang Technological Safety in HRI
Catherine Lollett, Advaith Sriram, Mitsuhiro Kamezaki, Waseda University,The S Uafety in HRI
Jasper Geldenbott, Karen Yan Ming Leung University of WashingtoSafety in HRI
Youngim Nam, Cheolhyeon Kwon Ulsan National Institut Safety in HRI
Xiaotong Zhang, Jinger Chong, Kamal Youcef-Toumi Massachusetts InstituteSafety in HRI
Zhiquan Zhang, Tianyu Li, Nadia Figueroa University of Pennsylva Safety in HRI
Yifan Jia, Christopher M. Poskitt, Peixin Zhang, Jingy Zhejiang University,Sin Safety in HRI
Demetrious T. Kutzke, Ashwin Wariar, Junaed Sattar University of Minnesota Search & Rescue Ro
Pierre-Luc Richard, Bellemare Jonathan, Philippe Hamel Hydro-Quebec Research Search & Rescue Ro
Yukihiro Maezawa, Yuichi Ambe, Yu Yamauchi, Masash Tohoku University,Osaka Search & Rescue Ro
Arnau Romero, Carmen Delgado, Lanfranco Zanzi, Raul i,CAT Foundation,NEC La Search & Rescue Ro
Akash Patel, Mario Alberto Valdes Saucedo, ChristoforLuleå University of Tec Search & Rescue Ro
Yanshuo Li, Kuo-shih Tseng National Central UniversSearch & Rescue Ro
Yanshuo Li, Kuo-shih Tseng National Central UniversSearch & Rescue Ro
Martin Zoula, Jan Faigl Czech Technical UniversSearch & Rescue Ro
Adam Seewald, Marvin Chancán, Connor Mccann, Seongh Yale University,Harvard Search & Rescue Ro
Franklin Kenghagho Kenfack, Michael Neumann, PatricUniversity of Bremen,U Semantic Scene Und
Roberto Bigazzi, Lorenzo Baraldi, Shreyas Kousik, Rit University of Modena and Semantic Scene Und
Jianing Yang, Xuweiyi Chen, Shengyi Qian, Nikhil Ma University of Michigan, Semantic Scene Und
Junwon Seo, Sungdae Sim, Inwook Shim Agency for Defense DevSemantic Scene Und
Jiale Li, Hang Dai, Yu Wang, Guangzhi Cao, Chun LuoZhejiang University,Uni Semantic Scene Und
Markus Käppeler, Kürsat Petek, Niclas Vödisch, Wolfr University of Freiburg, Semantic Scene Und
Hongmin Mu, Gang Zhang, Mengchu Zhou, Zhengcai Beijing University of Ch Semantic Scene Und
Rongtao Xu, Changwei Wang, Duzhen Zhang, Man Zhan Institute of Automation, Semantic Scene Und
Adam Rashid, Chung Min Kim, Justin Kerr, Letian Fu, UC Berkeley,University Semantic Scene Und
Giang Nguyen, Daniel Beßler, Simon Stelter, Mihai PomUniversity of Bremen,UnSemantic Scene Und
Jingwen Wang, Juan Jose Tarrio, Lourdes Agapito, Pabl University College Lo Semantic Scene Und
Scott Fredriksson, Akshit Saradagi, George NikolakopoLuleå University of Tec Semantic Scene Und
Kashu Yamazaki, Taisei Hanyu, Khoa Vo, Trong ThangUniversity of Arkansas,W Semantic Scene Und
Kadir Yilmaz, Jonas Schult, Alexey Nekrasov, Bastian RWTH Aachen UniversitSemantic Scene Und
Rodrigo Marcuzzi, Lucas Nunes, Louis Wiesmann, EliasUniversity of Bonn Semantic Scene Und
Jiaxu Kang, Bolei Chen, Ping Zhong, Haonan Yang, ShCentral South UniversitySemantic Scene Und
Mario Alberto Valdes Saucedo, Akash Patel, Akshit SarLulea University of Tec Semantic Scene Und
Chad Samuelson, Joshua Mangelson University,Brigham YounSemantic Scene Und
Zhiyuan Wu, Jiaqi Li, Yi Feng, Chengju Liu, Wei Ye, Qi Tongji University Semantic Scene Und
Hao Huang, Shuaihang Yuan, Congcong Wen, Yu Hao,New York University,NeSemantic Scene Und
Matthias Zeller, Vardeep Singh Sandhu, Benedikt Mersch CARIAD SE,University oSemantic Scene Und
Joao Marcos Correia Marques, Albert Zhai, Shenlong University of Illinois Semantic Scene Und
Wenjie Gao, Jiawei Fu, Yanqing Shen, Haodong Jing, Xi'an Jiaotong University, Semantic Scene Und
Ukcheol Shin, Kyunghyun Lee, In So Kweon, Jean Oh CMU(Carnegie Mellon Un Semantic Scene Und
Elias Greve, Martin Büchner, Niclas Vödisch, Wolfram University of Freiburg, Semantic Scene Und
Minsu Kim, Giseop Kim, Kyong Hwan Jin, Sunwook ChKorea Institute of Sci Semantic Scene Und
Junming Wang, Zekai Sun, Xiuxian Guan, Tianxiang S The University of Hong Semantic Scene Und
Cong Fan, Shengkai Zhang, Kezhong Liu, Shuai WangWuhan University of TecSensor Fusion I
Ye Li, Hanjiang Hu, Zuxin Liu, Xiaohao Xu, Xiaonan H University of Michigan,CSensor Fusion I
Jiahang Cao, Xu Zheng, Yuanhuiyi Lyu, Jiaxu Wang, RThe Hong Kong UniversiSensor Fusion I
Jun Cen, Shiwei Zhang, Yixuan Pei, Kun Li, Hang Zhe The Hong Kong UniversiSensor Fusion I
Aniruddh Sikdar, Jayant Teotia, Suresh Sundaram Indian Institute of Scien Sensor Fusion I
Zeping Zhang, Tianran Liu, Robert Laganière University of Ottawa Sensor Fusion I
Hao Dong, Weihao Gu, Xianjing Zhang, Jintao Xu, Rui ETH Zürich,HAOMO.AI Sensor Tec Fusion I
Kailai Li, Ziyu Cao, Uwe D. Hanebeck Linköping University,KarSensor Fusion I
Cheng Chi, Xin Zhang, Jiahui Liu, Yulong Sun, Zihao Shanghai Jiao Tong Univ Sensor Fusion I
Hanyu Zhou, Zhiwei Shi, Hao Dong, Shihan Peng, Yi CHuazhong University of Sensor Fusion II
Tanmoy Dam, Sanjay Bhargav Dharavath, Sameer Alam, Saab NTU Joint Lab, Nan Sensor Fusion II
Han Li, Yukai Ma, Yaqing Gu, Kewei Hu, Yong Liu, Xin Zhejiang University,zhejSensor Fusion II
Zhengcheng Shen, Linh Kästner, Yi Gao, Jens LambreTU Berlin,T-Mobile, TU B Sensor Fusion II
Youjie Zhou, Guofeng Mei, Yiming Wang, Fabio Poiesi,Shandong university,UnSensor Fusion II
Ryoichi Ishikawa, Shuyi Zhou, Yoshihiro Sato, Takeshi The University of TokyoSensor Fusion II
Tianbo Pan, Zidong Cao, Lin Wang Hong Kong University Sensor Fusion II
Arturo Del Castillo Bernal, Philippe Decoste, James Ri McGill University Sensor Fusion II
Simon-pierre Deschênes, Dominic Baril, Matej Boxan, Université Laval,Norlab,Sensor Fusion II
Shuo Xin, Zhen Zhang, Mengmeng Wang, Xiaojun Hou,Zhejiang University,ChinSensors and Auditi
Lea Steffen, Thomas Trapp, Arne Roennau, Rüdiger DiFZI Research Center forSensors and Auditi
Hikaru Arita, Hayato Nakamura, Takuto Fujiki, Kenji TaKyushu University Sensors and Auditi
Massimiliano De Sa, Venkata Naga Prasanth Kotaru, KUniversity of California, Sensors and Auditi
John Thomas, Francois Chaumette Inria Rennes,Inria centeSensors and Auditi
Shuyang Zhang, Jinhao He, Bohuan Xue, Wu Jin, Pengy The Hong Kong UniversiSensors and Auditi
Wen Fan, Haoran Li, Dandan Zhang University of Bristol,Im Sensors and Auditi
Inkyu An, Guoyuan An, Taeyoung Kim, Sung-Eui YoonETRI,KAIST Sensors and Auditi
Swapnil Sontakke, Pradyoth Hegde, Prashant BannulmIndian Institute of Sensors and Auditi
Yuxuan Luo, Zekun Wu, Zhouhui Lian Peking University Service Robotics
Federico Magistri, Rodrigo Marcuzzi, Elias Ariel Marks,University of Bonn,PhotService Robotics
Jeremy Nash, Quintin Dwight, Lucas Saldyt, Haoda WaJet Propulsion LaboratorService Robotics
Takahiro Miki, Joonho Lee, Lorenz Wellhausen, Marco ETH Zurich,ETH Zürich Service Robotics
Mimo Shirasaka, Tatsuya Matsushima, Soshi Tsunashim The University of Tokyo,Service Robotics
Yuchun Guo, Zhiqing Lu, Yanling Zhou, Xin Jiang Harbin Institute of Tec Service Robotics
Anzhe Chen, Hongxiang Yu, Yue Wang, Rong Xiong Zhejiang University Service Robotics
Hongxiang Yu, Anzhe Chen, Kechun Xu, Dashun Guo, Zhejiang University,Nank Service Robotics
Anurisha Piyathma Dunuwila, Lahiru Gunawardhana, HUniversity of Moratuwa,Service Robotics
Zhenhua Gong, Chuanxin Ning, Jiejunyi Liang, Ting ZhSoochow University,Huaz Service Robotics
Hideyuki Tanaka, Kunihiro Ogata National Institute of AI Service Robotics
Patrick Mania, Simon Stelter, Gayane Kazhoyan, MichaUniversity of Bremen,UnService Robotics
M. A. Viraj J. Muthugala, Bhagya Samarakoon, Raihan Singapore University of Service Robotics
Frank Joublin, Antonello Ceravola, Pavel Smirnov, Fel Honda Research InstituService Robotics
Xuchen Han, Joseph Masterjohn, Alejandro Castro Toyota Research InstituSimulation and Ani
Bo Ling, Yan Lyu, Dongxiao Li, Guanyu Gao, Yi Shi, Southeast University,NaSimulation and Ani
Christian Jestel, Karol Rösner, Niklas Dietz, Nicolas B Fraunhofer IML Simulation and Ani
José Antonio Fernández-fernández, Ralph Lange, Stefan RWTH Aachen University Simulation and Ani
Li-ming Chao, Liang Li Max Planck Institute of Simulation and Ani
Antoine Richard, Junnosuke Kamohara, Kentaro Uno, University of Luxembourg Simulation and Ani
Jack Lu, Kelvin Wong, Chris Zhang, Shun Da Suo, RaqUniversity of Waterloo,USimulation and Ani
Gang Yang, Siyuan Luo, Yunhai Feng, Zhixin Sun, Chen National University of S Simulation and Ani
Marvin Wiedemann, Ossama Ahmed, Anna Dieckhoefer, Fraunhofer Institute for Simulation and Ani
Renlang Huang, Minglei Zhao, Jiming Chen, Liang Li Zhejiang University,ZhejSLAM I
James Ross, Nimet Kaygusuz, Oscar Alejandro MendeUniversity of Surrey SLAM I
Shida Xu, Kaicheng Zhang, Ziyang Hong, Yuanchang LImperial College LondonSLAM I
Kaicheng Zhang, Yining Ding, Shida Xu, Ziyang Hong Heriot-Watt University,I SLAM I
Morten Nissov, Nikhil Khedekar, Kostas Alexis NTNU,NTNU - Norwegian SLAM I
Pushyami Kaveti, Shankara Narayanan Vaidyanathan, Northeastern University,SLAM I
Jie Yin, Ang Li, Wei Xi, Wenxian Yu, Danping Zou Shanghai Jiao Tong Univ SLAM I
Zhe Xin, Yufeng Yue, Liangjun Zhang, Chenming Wu Meituan,Beijing Institu SLAM I
Yue Zhuge, Haiyong Luo, Runze Chen, Yushi Chen, Ji Institute of Computing SLAM II
Binqian Jiang, Shaojie Shen Hong Kong University oSLAM II
Saurabh Gupta, Tiziano Guadagnino, Benedikt Mersch,University
I of Bonn,DexoSLAM II
Kai Huang, Junjiao Zhao, Zhongyang Zhu, Chen Ye, TiTongji University,Tongji SLAM II
Daniel Casado Herraez, Matthias Zeller, Le Chang, IgnacUniversity of Bonn & CASLAM II
Emilio Olivastri, Alberto Pretto University of Padua,UniSLAM II
Yu Han, Ziwei Long, Yanting Zhang, Wu Jin, Zhijun FanDonghua University,Tong SLAM II
Yutong Wang, Chaoyang Jiang, Xieyuanli Chen beijing institute of tec SLAM II
Wenhui Wei, Jiantao Li, Kaizhu Huang, Jiadong Li, Xin University of Science SLAM II
George Terzakis, Manolis Lourakis Varjo Technologies,Foun SLAM III
Junkai Niu, Sheng Zhong, Yi Zhou Hunan University SLAM III
Hriday Bavle, Jose Luis Sanchez-Lopez, Muhammad Sha University of Luxembourg SLAM III
Stefan Schubert, Peer Neubert, Sourav Garg, Michael M Chemnitz University of SLAM III
Jui-te Huang, Ruoyang Xu, Akshay Hinduja, Michael K Carnegie Mellon UniversSLAM III
Oguzhan Ilter, Iro Armeni, Marc Pollefeys, Daniel BaratETH Zürich,Stanford UnSLAM III
Xiongfeng Peng, Zhihua Liu, Weiming Li, Ping Tan, S Samsung R&D Institute SLAM III
David Skuddis, Norbert Haala University of Stuttgart SLAM III
Yuezhang Lv, Yunzhou Zhang, Xiaoyu Zhao, Wu Li, JiaNortheastern UniversitySLAM III
Seth Isaacson, Pou-Chun Kung, Mani Ramanagopal, Ra University of Michigan,USLAM IV
Yibin Wu, Tiziano Guadagnino, Louis Wiesmann, LasseUniversity
K of Bonn SLAM IV
Qi Chen, Guanghao Li, Xiangyang Xue, Jian Pu Fudan University SLAM IV
Jesse Morris, Yiduo Wang, Viorela Ila University of sydney,UnSLAM IV
Chang Wu, Yuan You, Yifei Yuan, Xiaotong Kong, YingUniversity of Electroni SLAM IV
Zheng Liu, Xiyuan Liu, Fu Zhang University of Hong KongSLAM IV
Zhiming Chen, Kun Zhang, Hua Chen, Michael Yu WanHong Kong University ofSLAM Science
IV and Technology,Southern
Jiarong Lin, Yuan Chongjian, Yixi Cai, Haotian Li, Yu The University of Hon SLAM IV
Pushyami Kaveti, Matthew Giamou, Hanumant Singh, Northeastern UniversitySLAM IV
Xin Su, Sebastian Eger, Adam Misik, Dong Yang, RastiTechnical University o SLAM V
Hanyeol Lee, Jaehyung Jung, Chan Gook Park Seoul National UniversitSLAM V
Jiyeon Nam, Soojeong Hyeon, Youngjun Joo, Dongki ASRI, Seoul National UnSLAM V
Mohamad Qadri, Zachary Manchester, Michael Kaess Carnegie Mellon UniversSLAM V
Gihan Charith Premachandra Hanchapola Appuha, RanSingapore University of SLAM V
Qiucan Huang, Yuchen Liang, Zhijian Qiao, Shaojie ShHong Kong University oSLAM V
Eunju Jeong, Jina Lee, Suyoung Kang, Pyojin Kim Sookmyung Women's Univ SLAM V
Yuxiaotong Lin, Jiming Chen, Liang Li ZJU,Zhejiang University,SLAM V
Christina Kassab, Matias Mattamala, Lintong Zhang, Ma University of Oxford SLAM V
Gabriel Moreira, Manuel Marques, Joao Paulo Costeir Carnegie Mellon Universi SLAM VI
Simon Boche, Sebastián Barbas Laina, Stefan LeuteneTechnical University of SLAM VI
Matthew Hanlon, Boyang Sun, Marc Pollefeys, HermanETH Zurich SLAM VI
Jing Zeng, Yanxu Li, Jiahao Sun, Qi Ye, Yunlong Ran, Zhejiang University SLAM VI
He Yin, Jong Jin Park, Marcelino Almeida, Martin Labr Amazon.com, Inc.,Amazo SLAM VI
Yang Xu, Ronghao Zheng, Senlin Zhang, Meiqin Liu, Zhejiang University,Uni SLAM VI
Jianeng Wang, Jonathan Gammell University of Oxford SLAM VI
Alessandro Fornasier, Pieter Van Goor, Eren Allak, R University of Klagenfurt SLAM VI
Bin Zhao, Yunzhou Zhang, Junjie Huang, Xichen ZhangNortheastern UniversitySLAM VI
Quecheng Qiu, Xuyang Zhang, Shunyi Yao, Yu'an Chen School of Data Science,Social HRI
Minae Kwon, Hengyuan Hu, Vivek Myers, Siddharth KaStanford University,Fac Social HRI
Matthias Rehm, Antonia Krummheuer, Carlos Gomez Aalborg University Social HRI
Amar Halilovic, Senka Krivic Ulm University,UniversitSocial HRI
Yanying Zhou, Jochen Garcke University of Bonn,UniveSocial HRI
Lucrezia Grassi, Zhouyang Hong, Carmine Tommaso Re University of Genova,UnSocial HRI
Iman Bakhoda, Pourya Shahverdi, Katelyn Rousso, Just Intelligent Robotics Lab Social HRI
Randy Gomez, Deborah Szapiro, Sara Cooper, Nabil Bou Honda Research InstitutSocial HRI
Zhidong Su, Weihua Sheng Oklahoma State UniversSocial HRI
Feiyang Deng, Shilong Yao, Kwai Man Luk City University of Hong Soft Robot Applicati
Peiyu Luo, Shilong Yao, Yiyao Yue, Hong Yan, Jiank Southern University of Soft Robot Applicati
Yongkang Jiang, Xin Tong, Zhongqing Sun, Yanmin Zhou Tongji University,Shen Soft Robot Applicati
Fenglin Han, Lei Fei, Run Zou, Weijian Li, Jinghao Zh Central South UniversitySoft Robot Applicati
Cora Maria Sourkounis, Ditzia Susana Garcia Morales Leibniz University Han Soft Robot Applicati
Obumneme Godson Osele, Kentaro Barhydt, Nicholas C Stanford University,MasSoft Robot Applicati
Nguyen Hai, Tadashi Kozuno, Cristian Beltran-Herna Northeastern Universi Soft Robot Applicati
Samuel Pilch, Daniel Müller, Oliver Sawodny University of Stuttgart Soft Robot Applicati
Philip. H Johnson, Kai Junge, Charles Whitfield, Josie University of Lincoln,Éc Soft Robot Applicati
Rui Chen, Xinrui Tao, Changyong (chase) Cao, Pei Jia Chongqing University,Ca Soft Robot Applicati
Peter Wharton, Tsam Lung You, George Jenkinson, RicUniversity of Bristol Soft Robot Applicati
Kailuan Tang, Chenghua Lu, Yishan Chen, Yin Xiao, S Harbin Institute of Te Soft Robot Applicati
Mo Deng, Fengya Fan, Xi Wei University of Science a Soft Robot Applicati
Pengyuan Wang, Zheng Zheng, Jiazhen Sun, Yuqiang H L arbin Institute of Tech Soft Robot Applicati
Zhenxing Wang, Xiao He, Yuhang Zhang, Cheng Zhang, Chinese Academy of Scie Soft Robot Applicati
Xinpei Ai, Hengmao Yue, Wei Wang Hanyang University Soft Robot Applicati
Arvyn De, Divyam Kumar, Ian Kwuan, Alex Qiu, Ai-Pin Georgia Institute of Te Soft Robot Applicati
Xiangxing Wang, Xuan Pei, Xinyang Wang, Taogang HBeijing Jiaotong UniversSoft Robot Material
Guirec Maloisel, Espen Knoop, Christian Schumacher, Disney Research,The Wa Soft Robot Material
Sophie Paul, Matthew Devlin, Elliot Hawkes University of California Soft Robot Material
Sandra Q. Liu, Edward Adelson Massachusetts InstituteSoft Robot Material
Tianqi Yue, Hermes Gadelha, Jonathan Rossiter University of Bristol,UnivSoft Robot Materials and Design I
Alexander Keller, Tianqi Yue, Qiukai Qi, Andrew Conn, University of Bristol Soft Robot Material
Peiwen Yang, Shuguang Li Tsinghua University,TsinSoft Robot Material
Fanxing Kong, Jie Zhao, Hegao Cai, Yanhe Zhu Harbin Institute of Tech Soft Robot Material
Yangfan Li, Jun LIU, Wenyu Liang, Zhuangjian Liu Institute of High Per Soft Robot Material
Dong Wang, Chengru Jiang, Guo-Ying Gu Shanghai Jiao Tong Univ Soft Robot Material
Xin Fu, Daohui Zhang, Liyan Mo, Kai Li, Xingang Zhao Shenyang Institute of Soft Robot Material
Hanna Matusik, Chao Liu, Daniela Rus MIT,Massachusetts InstiSoft Robot Material
Disheng Xie, Yujie Su, Xiangqian Shi, Zheng Li, Kai YuThe Chinese University Soft Robot Material
Hao Liu, Changchun Wu, Senyuan Lin, Yonghua Chen The University of Hong Soft Robot Material
Nam Gyun Kim, Dongoh Seo, Shinwoo Park, Jee-HwanKorea Advanced Institu Soft Robot Material
Mijaíl Jaén Mendoza Flores, Nicholas Naclerio, Elliot University of California Soft Robot Material
Linda Paterno, Canberk Sozer, Sujit Sahu, Arianna MenScuola Superiore Sant Soft Robot Material
Kyle L. Walker, Alix Partridge, Hsing-yu Chen, Rahul The National Robotarium Soft Robot Material
Chaoxiang Ye, Guido De Croon, Salua Hamaza Delft University of Tech Soft Sensors and Ac
Annan Zhang, Lillian Chin, Daniel L Tong, Daniela Rus Massachusetts InstituteSoft Sensors and Ac
Loïc Mosser, Laurent Barbé, Lennart Rubbert, Pierre Icube - Université de s Soft Sensors and Ac
Hexin Li, Jiang Li, Ruichen Zhen, Ming Cheng, Kehan Harbin Institute of Tech Soft Sensors and Ac
Ilias Zournatzis, Sotiris Kalaitzakis, Panagiotis Polygeri Hellenic Mediterranean Soft Sensors and Ac
Seoyeon Ham, Soe Lin Paing, Brian Byunghyun Kang, Hanyang University,ArizSoft Sensors and Ac
Daniel Van Lewen, Catherine Wang, Hun Chan Lee, AnBoston University Soft Sensors and Ac
Viola Del Bono, Max Mccandless, Frank Juliá Wise, ShBoston University,Boston Soft Sensors and Ac
Linhang Ju, Hanze Jia, Yanjun Shi, Xilun Ding, Yang Beihang University,BeijiSoft Sensors and Ac
Barnabas Gavin Cangan, Sizhe Tian, Stefan Escaida Na ETH Zurich,Inria, Unive Soft Sensors and Ac
Cheng Pan, Kieran Gilday, Emily Sologuren, Kai JungeSwiss Federal Institut Soft Sensors and Ac
Jiang Zou, Shakiru Olajide Kassim, Jieji Ren, Vahid VaShanghai Jiao Tong U Soft Sensors and Ac
Xingyu Chen, Jialei Shi, Helge Wurdemann, Thomas GUniversity College LondSoft Sensors and Ac
Joonwon Kang, Sudong Lee, Yong-Lae Park Seoul National Universi Soft Sensors and Ac
Shivani Deglurkar, Charles Xiao, Luke Gockowski, Mega University of California Soft Sensors and Ac
Feifei Chen, Zenan Song, Shitong Chen, Guo-Ying Gu,Shanghai Jiao Tong Univ Soft Sensors and Ac
Jialun Liu, Mennaallah Soliman, Dana Damian The University of SheffieSoft Sensors and Ac
Joanna Jones, Marco Pontin, Dana Damian University of Sheffield Soft Sensors and Ac
Carsten Oldemeyer, Matthias Hellerer, Matthias ReinerGerman Aerospace Cent Software for Robot
Jean-pierre Busch, Lennart Reiher, Lutz Eckstein RWTH Aachen University Software for Robot
Or Wertheim, Dan Rouven Suissa, Ronen Brafman ben gurion university o Software for Robot
Matthias Mayer, Jonathan Külz, Matthias Althoff Technical University of Software for Robot
Hajo Erwich, Bart Duisterhof, Guido De Croon Delft University of Tech Software for Robot
Angelos Mavrogiannis, Christoforos Mavrogiannis, Yia University of Maryland, Software for Robot
Nicholas Cote, Annabella Macaluso, Sachin Chitta Autodesk, Inc,UniversitySoftware for Robot
Pascal Stoop, Tharaka Ratnayake, Giovanni Toffetti OST,Zurich Applied ScieSoftware for Robot
Ettore Sani, Antonio Sgorbissa, Stefano Carpin University of Genova,Uni Software for Robot
José Ignacio Delgado-centeno, Paula Harder, Valentin Universite du Luxembourg Space Robotics I
Mahya Ramezani, Mohammadamin Alandihallaj, AndreUniversity of LuxembouSpace Robotics I
Achira Boonrath, Feng Liu, Eleonora Botta, Souma Ch University at Buffalo, S Space Robotics I
Tae Ha Park, Simone D’amico Stanford University Space Robotics I
Bahador Beigomi, Zhenghong (george) Zhu York University Space Robotics I
Arunkumar Rathinam, Haytam Qadadri, Djamila AouadUniversity of LuxembourSpace Robotics I
Lennart Werner, Pedro Proença, Andreas Nuechter, RoETH Zürich,California InSpace Robotics I
Kyle McCleary, Swaminathan Gurumurthy, Paulo Fisch,Carnegie Mellon UniversSpace Robotics I
Georgios Rekleitis, Evangelos Papadopoulos National Technical UniveSpace Robotics II
Jingdong Zhao, Zirui Wang, Ziyi Liu, Liangliang Zhao, Harbin Institute of Tech Space Robotics II
David Rodríguez-Martínez, Kentaro Uno, Kenta Sawa, École polytechnique fé Space Robotics II
Micah Oevermann, Derek Pravecek, Joseph Garrett JibTexas A&M University Space Robotics II
Mehmed Yüksel, Wiebke Brinkmann, Marko Jankovic, H DFKI GmbH,DFKI Robotic Space Robotics II
Daichi Hirano, Nobutaka Tanishima, Tony G. Chen Japan Aerospace ExplorSpace Robotics II
Ziao Qin, Jingzhou Song, Xinglong Gong, Changrui LiuBeijing University of P Space Robotics II
Kostas Nanos, Efstathios Chachamis, Evangelos PapaNational Technical UniveSpace Robotics II
Mohammad Mahdi Kalantari, Erol Ozgur, Mohammad Alkh Clermont Auvergne INP,Surgical Robotics I
Junlei Hu, Dominic Jones, Mehmet Dogar, Pietro Valdas University of Leeds Surgical Robotics I
Nural Yilmaz, Brendan Burkhart, Anton Deguet, Peter Marmara University,John Surgical Robotics I
Keshara Weerasinghe, Seyed Hamidreza Roodabeh, KUniversity of Virginia Surgical Robotics I
Rokas Bendikas, Valerio Modugno, Dimitrios Kanoulas,UCL,University College Surgical Robotics I
Radian Gondokaryono, Mustafa Haiderbhai, Sai AneesUniversity of Toronto,In Surgical Robotics I
Juan Antonio Barragan, Jintan Zhang, Haoying Zhou, Johns Hopkins UniversitSurgical Robotics I
Samuel Schmidgall, Axel Krieger, Jason Eshraghian Johns Hopkins University Surgical Robotics I
Jiawei Fu, Yonghao Long, Kai Chen, Wang Wei, Qi Do Institute of Artificial Surgical Robotics I
Jie Wang, Chengquan Hu, Jingyi Kang, jiayuan Liu, Lo Tsinghua University Surgical Robotics II
Christopher D'ambrosia, Florian Richter, Zih-Yun Chiu, University of California Surgical Robotics II
Jiaxin Guo, Jiangliu Wang, Zhaoshuo Li, Tongyu Jia, QThe Chinese University Surgical Robotics II
Xiao Liang, Fei Liu, Yutong Zhang, Yuelei Li, Shan Lin,University of California Surgical Robotics II
Brian Y Cho, Alan Kuntz University of Utah Surgical Robotics II
Jibiao Chen, Junyan Yan, Yufu Qiu, Haiyang Fang, Ji The Chinese University Surgical Robotics II
Kay Hutchinson, Ian Reyes, Zongyu Li, Homa AlemzadUniversity of Virginia,IB Surgical Robotics II
Rebekka Charlotte Peter, Steffen Peikert, Ludwig Haid Carl Zeiss AG,Friedrich Surgical Robotics II
Qinxi Yu, Masoud Moghani, Karthik Dharmarajan, VinceUniversity of Toronto,UCSurgical Robotics II
Jingwei Song, Keke Yang, Zheng Zhang, Meng Li, TuoyUniversity of Michigan, Surgical Robotics II
Keshav Iyengar, S.m.hadi Sadati, Christos Bergeles, University College LondSurgical Robotics II
Jake Childs, Caleb Rucker EndoTheia, Inc.,UniversSurgical Robotics II
Jacynthe Francoeur, Dimitri A. Lezcano, Yernar Zhetpi Polytechnique Montréal,Surgical Robotics II
Jingjie Wu, Kevin Yu, Ithza Lopez, Alexa Aguilar Izqui The University of Texas Surgical Robotics II
Izadyar Tamadon, S.m.hadi Sadati, Virginia Mamone, Vi University of Twente,Ki Surgical Robotics II
Long Lei, Baoliang Zhao, Xiaozhi Qi, Rui Mi, Hai Ye, Shenzhen Institute of Surgical Robotics II
Andreas Brandstätter, Scott Smolka, Scott Stoller, Ash Technische Universität Swarm Robotics
Tugay Alperen Karagüzel, Victor Retamal Guiberteau, N Vrije Universiteit Amst Swarm Robotics
Luigi Feola, Andreagiovanni Reina, Mohamed S. Talamal University of Rome "La Swarm Robotics
Xiaoyang Qin, Yongliang Yang, Mengyun Pan, Long Cui Shenyang Institute of A Swarm Robotics
Petras Swissler, Michael Rubenstein New Jersey Institute of Swarm Robotics
Steven Ceron, Wei Xiao, Daniela Rus Massachusetts InstituteSwarm Robotics
David Garzon Ramos, Mauro Birattari Université Libre de BruxSwarm Robotics
Christian Llanes, Zahi Kakish, Kyle Williams, Samuel The Georgia Institute o Swarm Robotics
Sriram N N, Dinesh Jayaraman, Manmohan ChandrakeNEC Labs America,Univer Task and Motion Pla
Brandon Vu, Toki Migimatsu, Jeannette Bohg Stanford University Task and Motion Pla
Tobias Löw, Sylvain Calinon Idiap Research InstituteTask and Motion Pla
Teng Xue, Amirreza Razmjoo, Sylvain Calinon idiap/EPFL,Idiap Researc Task and Motion Pla
Junjie Gao, Peijia Xie, Xuheng Gao, Zhirui Sun, Jian Harbin Institute of Tec Task and Motion Pla
Marc Toussaint, Joaquim Ortiz De Haro, Valentin Hart TU Berlin,ETH Zürich,T Task and Motion Pla
Wanting Jin, Alcherio Martinoli EPFL Task and Motion Pla
Kim Tien Ly, Valeriy Semenov, Mattia Risiglione, WolfgUniversity of Oxford,Oxfo Task and Motion Pla
Svetlana Levit, Joaquim Ortiz De Haro, Marc ToussaintTU Berlin Task and Motion Pla
Enrico Saccon, Ahmet Tikna, Davide De Martini, Edoard University of Trento,UnivTask and Motion Pla
Sujin Jeon, Suyeon Shin, Byoung-Tak Zhang Seoul National UniversitTask and Motion Pla
Patrick Hammer, Peter Isaev, Lei Feng, Robert Johan KTH Royal Institute of Task and Motion Pla
Yoonchang Sung, Rahul Shome, Peter Stone The University of Texas Task and Motion Pla
Kandai Watanabe, Georgios Fainekos, Bardh Hoxha, Mo University of Colorado Task and Motion Pla
Mario Fiore, Ciro Natale Università degli Studi deTask and Motion Planning II
Alessio Saccuti, Riccardo Monica, Jacopo Aleotti University of Parma Task and Motion Pla
Azin Shamshirgaran, Sandeep Manjanna, Stefano CarpUniversity of California Task Planning
Shuangyu Xie, Chengsong Hu, Di Wang, Joe Johnson,Texas A&M University,Mo Task Planning
Jack Collins, Mark Robson, Jun Yamada, Mohan SridhaUniversity of Oxford,Un Task Planning
Robin Kirschner, Carina M. Micheler, Yangcan Zhou, STU Munich, Institute fo Task Planning
Khen Elimelech, Zachary Kingston, Wil Thomason, Vard Rice University Task Planning
David Paulius, Alejandro Agostini, Dongheui Lee Brown University,UniverTask Planning
Md Sadman Sakib, Yu Sun University of South Flor Task Planning
Fan Zeng, Shouhei Shirafuji, Changxiang Fan, Masahiro the University of Toky Task Planning
Shuge Wu, Chunzheng Wang, Jiayi Pan, Dongming Han Beihang University Task Planning
Davide Torielli, Leonardo Franco, Maria Pozzi, Luca M Humanoids and Human Telerobotics
Cent and Te
Basak Gulecyuz, Vincent Von Büren, Xiao Xu, EckeharTechnical University of Telerobotics and Te
Alejandro Suarez, Antonio Gonzalez-morgado, Aníbal OUniversity of Seville,UniTelerobotics and Te
Burak Kizilkaya, Changyang She, Guodong Zhao, MuhUniversity of Glasgow,UTelerobotics and Te
Wen Fan, Xiaoqing Guo, Enyang Feng, Jialin Lin, YuanUniversity of Bristol,Ci Telerobotics and Te
Mads Mørch Antonsen, Francesco Chinello, Qi Zhang Aarhus University Telerobotics and Te
Lingyun Chen, Abdeldjallil Naceri, Abdalla Swikir, San Technical University of Telerobotics and Te
Jaehyun Park, Il Seop Choi, Sang-woo Choi, Keehoon Pohang University of Telerobotics and Te
Gabriel Quere, Freek Stulp, David Filliat, João Silvério DLR,DLR - Deutsches Ze Telerobotics and Te
Florent Audonnet, Jonathan Grizou, Andrew Hamilton University of Glasgow,STelerobotics and Te
Antonin Dallard, Mehdi Benallegue, Fumio Kanehiro, LIRMM,AIST Japan,Natio Telerobotics and Te
Joonhyung Lee, Sangbeom Park, Jeongeun Park, KyunKorea University,Chung-Telerobotics and Te
Giovanni Cortigiani, Bernardo Brogi, Alberto Villani, To University of Siena Telerobotics and Te
Fabian Clemens Weigend, Xiao Liu, Shubham Dilip So Arizona State Universit Telerobotics and Te
Patrick Naughton, James Seungbum Nam, Andrew Strat University of Illinois a Telerobotics and Te
Mingyu Cai, Karankumar Patel, Soshi Iba, Songpo Li University of Californi Telerobotics and Te
Amartya Purushottam, Christopher Xu, Yeongtae Jung University of Illinois, Telerobotics and Te
Batool Ibrahim, Imad Elhajj, Daniel Asmar American University of Telerobotics and Te
Jiajun Fei, Zhidong Deng Tsinghua University Transfer Learning
Cao Haozhi, Yuecong Xu, Jianfei Yang, Pengyu Yin, ShNanyang Technological Transfer Learning
Ruiqi Zhu, Tianhong Dai, Oya Celiktutan King's College London,ITransfer Learning
Hongyu Sun, Yongcai Wang, Wang Chen, Haoran Deng, Renmin University of Transfer Learning
Jiaming Liu, Rongyu Zhang, Xiaoqi Li, Xiaowei Chi, Peking University,Nanji Transfer Learning
Gergely Sóti, Xi Huang, Christian Wurll, Björn Hein Karlsruhe University of Transfer Learning
Kefei Wu, Xuming He, Yang Wang, Xiaopei Liu ShanghaiTech Universi Transfer Learning
Dongqingwei Dongqingwei, Peng Zeng, Guangxi Wan, Shenyang Institute of A Transfer Learning
Zhenyu Liao, Zonghao Dong, Ricardo Manríquez, AnkitTohoku
R University,LISV,Video Session
Jin-deok Lee, Hyun-min Joe Kyungpook National UniVideo Session
Ning Zhao, Jian Qi, Zhiyuan Yang, Sikai Zhao, Xin Sui Harbin Institute of Tech Video Session
Ying Shi, Xiaoshun Liu, Ke Tan, Zhenfeng Gu, Junhui Huazhong University of Video Session
Katsuhiko Fujiki, Takumi Hachimine, Satoshi Makita Fukuoka Institute of Te Video Session
Donghun Noh, Hyunwoo Nam, Kyle Gillespie, Yeting LiUniversity of California Video Session
Sravan Chittupalli, Liu Jun, Dong Huang Carnegie Mellon UniversVideo Session
Anna Adamczyk, Alexander Moortgat-pick, Daniel And Technical University of Video Session
Carlos Del Reguero Valle, Diego Emilio Alvarez GuerrerTecnologico de MonterrVideo Session
Haoye Dong, Liu Jun, Dong Huang Carnegie Mellon UniversVideo Session
Oscar Ochoa, Enrico Mendez, David Olivera Guzman, Vic Tecnologico de Monterre Video Session
Johannes Tenhumberg, Leon Sievers, Dominik Winkelbau German Aerospace Cent Video Session
Hasra Dodampegama, Mohan Sridharan University of Edinburgh Video Session
Han Wang Nanyang Technological Video
U Session
Michikuni Eguchi, Koki Kato, Tatsuya Oshima, Shunya University of Tsukuba,OVideo Session
Sehwa Chun, Chihaya Kawamura, Kenji Ohkuma, ToshiThe University of TokyoVideo Session
Kenjiro Tadakuma, Masahiro Watanabe, Satoshi TadokTohoku University Video Session
Tyler Morris, Eric Vaughan, Sydney Walker, Laython H University of TennesseeVideo Session
Xiu Zhang, Angela Peloso, Emiliano Votta, Elena De MPolitecnico di Milano Video Session
Maciej Stępień AGH University of Krak Video Session
Jayden Hong, Zengjie Zhang, Amir Mehdi Soufi EnayatUniversity of Victoria,E Video Session
Sepehr Ghassemi, Dennis Hong UCLA Video Session
Renan Lima Baima, Eduard Hartwich, Loïck Chovet, Mig University of Luxembourg,Video Session
Joseph Berman, John Patterson, He (Helen) Huang North Carolina State UniVideo Session
Petri Tikka VTT Techincal researchVideo Session
Justin Quan, Mingzhang Zhu, Dennis Hong UCLA,University of Cali Video Session
Seongjin Bien, Jon Škerlj, Felix Eberle, Ainoor Teimo Technical University of Video Session
Zonghao Zuo, Haoxuan Wang, Xia He, Zhuyin Shao, Jiaq Beihang University Video Session
Xiaonan Huang, Will Johnson, Shiyang Lu, Kun Wang, U Kniversity of Michigan, Video Session
Hideyuki Tsukagoshi, Nobuyuki Arai, Ichiro Kiryu, To Tokyo Institute of TechnVideo Session
Xia He, Zonghao Zuo, Zhuyin Shao, Haoxuan Wang, Fei Beihang University Video Session
Hao Liu, Chongyang Wang, Zhenxing Wang, Dingjia Li Chinese Academy of SciVideo Session
Qinghua Zhang, Hongwei Jing, Kerui Sun, Yuancheng Li, Harbin Institute of Tech Video Session
Yuan Yang, Wen Liang Yeoh, Teerapapa Luecha, Jeewo Kyushu University,SagaVideo Session
Hongwei Jing, Tianjiao Zheng, Sikai Zhao, Gangfeng L Harbin Institute of Tech Video Session
Martina Lippi, Michael Welle, Andrea Gasparri, Danica University of Roma Tre,Vision Systems
Kang Yufan, Guillaume Caron, Ryoichi Ishikawa, Adrie The University of TokyoVision Systems
Agniva Sengupta, Makki Karim, Adrien Bartoli Institut Pascal,UCA Vision Systems
Jingpei Lu, Florian Richter, Shan Lin, Michael Yip University of California Vision Systems
Jinkun Cao, Jiangmiao Pang, Kris Kitani Carnegie Mellon UniversVision Systems
Leonardo Saraceni, Ionut Marian Motoi, Daniele Nardi, Sapienza University of Vision Systems
Chao Hu, Ping Huang, Wei Wang Harbin Engineering UnivVision Systems
Weihan Wang, Chieh Chou, Ganesh Sevagamoorthy, Kevi Stevens Institute of Te Vision Systems
Alexander Millane, Helen Oleynikova, Emilie Wirbel, NVIDIA,ETH Zurich,ValeVision Systems
Justin Wasserman, Girish Chowdhary, Abhinav Gupta, University of Illinois a Vision-Based Navig
Payam Nourizadeh, Michael Milford, Tobias Fischer QUT Centre for RoboticVision-Based Navig
Tianrui Guan, Yurou Yang, Harry Cheng, Muyuan Lin, University of Marylan Vision-Based Navig
Lixuan Tang, Chuanfang Ning, George Adaimi, Auke Ijsp EPFL,École PolytechniqVision-Based Navig
Faraz Lotfi, Khalil Virji, Farnoosh Faraji, Lucas Berry McGill University Vision-Based Navig
Chengguang Xu, Christopher Amato, Lawson L.S. WonNortheastern UniversityVision-Based Navig
Takuya Kitade, Wataru Yamada, Keiichi Ochiai, MichitaNTT DOCOMO, INC.,Keio Vision-Based Navig
Henrique Morimitsu, Zhu Xiaobin, Roberto Marcondes C University of Science a Vision-Based Navig
Yanjun Sun, Yue Qiu, Yoshimitsu Aoki, Hirokatsu Kata Keio University,NationalVision-Based Navig
Lauri Aleksanteri Suomela, Jussi Oskari Kalliola, Harr Tampere University,TamVision-Based Navig
Nuo Xu, Wen Wang, Rong Yang, Mengjie Qin, ZheyuanZhejiang Lab,Dalhousie Vision-Based Navig
Jiaming Wang, Harold Soh National University of S Vision-Based Navig
Wenzhe Cai, Siyuan Huang, Guangran Cheng, Yuxing Southeast University,ShVision-Based Navig
Joshua Barton Knights, Stephen Hausler, Sridha Srid Queensland University Vision-Based Navig
Pascal Roth, Julian Nubert, Fan Yang, Mayank Mittal, ETH Zurich,ETH Zürich Vision-Based Navig
Xiaomin Lin, Nare Karapetyan, Kaustubh Joshi, TiancheUniversity of Maryland, Vision-Based Navig
Haresh Karnan, Elvin Yang, Garrett Warnell, Peter St The University of TexasVision-Based Navig
Jing Ye, Yang Liu, Congjing Yu, Changzhen Qiu, Zhiy Sun Yat-Sen University,Visual Learning I
Fangbo Qin, Taogang Hou, Shan Lin, Kaiyuan Wang, MInstitute of Automation, Visual Learning I
Mingjie Pan, Jiaming Liu, Renrui Zhang, Peixiang HuanPeking University,CUHKVisual Learning I
Xiaodong Gu, Weihao Yuan, Zuozhuo Dai, Siyu Zhu, CAlibaba Group,Hong Kon Visual Learning I
Zhenyu Jiang, Hanwen Jiang, Yuke Zhu The Unversity of Texas V a isual Learning I
Zijun Long, George Killick, Richard Mccreadie, Gerar University of Glasgow,SVisual Learning I
Yunze Liu, Changxi Chen, Zifan Wang, Li Yi Tsinghua University Visual Learning I
Yuxuan Liu, Zhenhua Xu, Huaiyang Huang, Lujia Wang,Hong Kong University oVisual Learning I
Zhongyang Zhang, Shuyang Cui, Kaidong Chai, HaoweUniversity of Californi Visual Learning I
Arul Selvam Periyasamy, Sven Behnke University of Bonn Visual Learning II
Ruoyu Fan, Wang Zhao, Matthieu Lin, Qi Wang, Yong-Tsinghua University,GuiVisual Learning II
Long Bai, Guankun Wang, Jie Wang, Xiaoxiao Yang, Hu The Chinese University Visual Learning II
Mayank Lunayach, Sergey Zakharov, Dian Chen, RaresGeorgia
A Institute of Tec Visual Learning II
Xiangyuan Wang, Kuangyi Chen, Wen Yang, Lei Yu, YaWuhan University,SynSe Visual Learning II
Yucheng Xu, Nanbo Li, Arushi Goel, Zonghai Yao, Zij University of Edinburgh Visual Learning II
Guoyu Lu University of Georgia Visual Learning II
Jieqi Shi, Peiliang Li, Xiaozhi Chen, Shaojie Shen the Hong Kong UniversitVisual Learning II
Boris Chidlovskii, Leonid Antsfeld Naver Labs Europe Visual Learning II
Xiaojie Xu, Tianshuo Xu, Fulong Ma, Yingcong Chen The Hong Kong UniversiVisual Learning III
Junyi Cao, Zhichao Li, Naiyan Wang, Chao Ma Shanghai Jiao Tong Univ Visual Learning III
Jiaxu Wang, Junhao He, Ziyi Zhang, Renjing Xu Hong Kong University oVisual Learning III
Guoyu Lu University of Georgia Visual Learning III
Rui Yu, Jiachen Liu, Zihan Zhou, Sharon X. Huang University of Louisville Visual Learning III
Wenjie Li, Jia Liu, Yanyan Wang, Dayong Ren, Wei Hao Nanjing University,HohaiVisual Learning III
Xiang Li, Varun Belagali, Jinghuan Shang, Michael S RStony Brook University, Visual Learning III
Jianfeng Gao, Xiaoshu Jin, Franziska Krebs, Noémie J Karlsruhe Institute of T Visual Learning III
Xu Zhang, Linyan Cui, Jihao Yin Beihang University,Beiha Visual Learning III
Sara Hatami Gazani, Matthew Tucsok, Iraj Mantegh, H University of Victoria,U Visual Perception a
Xu Bian, Wenzhao Chen, Xiaoyu Tian, Donglai Ran Xi’an Jiaotong Universit Visual Perception a
Zixing Lei, Zhenyang Ni, Ruize Han, Shuo Tang, Che Shanghai Jiao Tong UniVisual Perception a
Yunhao Li, Jing Wu, Lingzhe Zhao, Peidong Liu Westlake University Visual Perception a
Jianjia Xin, Lichun Wang, Kai Xu, Chao Yang, Baocai YBeijing University of te Visual Perception and Learning I
Ronghan Chen, Yang Cong, Yu Ren Sheyang Institute of A Visual Perception a
Min Hyuk Kim, Min Ju Kim, Seok Bong Yoo Chonnam National Unive Visual Perception a
Yuejiang Dong, Fang-lue Zhang, Song-hai Zhang Tsinghua University,VictVisual Perception a
Jiayuan Fan, Shijie Chen, Tao Chen, Mohammad MahdFudan University,FudanVisual Perception a
Junyu Zhu, Lina Liu, Yu Tang, Feng Wen, Wanlong Li, zhejiang University,Zhe Visual Perception a
Yijie Zhou, Likun Cai, Xianhui Cheng, Zhongxue Gan, Fudan University,Univers Visual Perception a
Bowen Yao, Yongjian Deng, Yuhan Liu, Hao Chen, YouBeijing University of Te Visual Perception a
Ahalya Ravendran, Mitch Bryson, Donald G DansereauThe Commonwealth Scient Visual Perception a
Jiaming Liu, Qizhe Zhang, Xiaoqi Li, Jianing Li, Gua Peking University,NanjinVisual Perception a
Jisong Kim, Geonho Bang, Kwangjin Choi, Minjae Seo Hanyang university,Hany Visual Perception a
Likun Wang, Zi Wang, Peter Kendall, Kevin Gumma, AlHunan University, UniveVisual Perception a
Zhengxuan Qiu, Jie Yang, Jiankun Wang Southern University of Visual Perception a
Anne Sielemann, Stefan Wolf, Masoud Roschani, Jens Fraunhofer IOSB,FraunhVisual Perception a
Zaixing He, Quanzhi Li, Xinyue Zhao, Jin Wang, Huar Zhejiang University,ZheVisual Perception A
Zhiyuan Zhang, Michael Koch, Daniel Ahmed Acoustic Robotics SysteVisual Perception A
Tat Hieu Bui, Yeong Gwang Son, Seung Jae Moon, QuSungkyunkwan UniversitVisual Perception A
Boyoung Kim, Junhong Min Samsung Electronics Visual Perception A
Alexander Prutsch, Horst Possegger, Horst Bischof Graz University of TechVisual Perception A
Mert Asim Karaoglu, Viktoria Markova, Nassir Navab, Technical University o Visual Perception A
Kexin Chen, Du Yuyang, Tao You, Mobarakol Islam, Z Department of ComputerVisual Perception A
Zhen Li, Ya-wen Deng, Qiang Ye, Weihong Yu, Haoxiang Institute of Automation Visual Perception A
Albert Enyedy, Ashay Aswale, Berk Calli, Michael GennWorcester Polytechnic In Visual Servoing
Ryogo Kai, Yuzuka Isobe, Sarthak Mahesh Pathak, Ka Chuo University Visual Servoing
Maximilian Argus, Abhijeet Nayak, Martin Büchner, Sil University of Freiburg Visual Servoing
Eduardo Godinho Ribeiro, Raul De Queiroz Mendes, Mar University of São Paulo Visual Servoing
Sotiris Aspragkathos, George Karras, Kostas KyriakopoNTUA,University of ThesVisual Servoing
Mohammadreza Shetab-bushehri, Miguel Aranda, YoucUniversité Clermont AuvVisual Servoing
Jiawen Zhu, Huayi Tang, Zhi-qi Cheng, Jun-yan He, Bi Dalian University of Te Visual Servoing
Alessandro Saviolo, Pratyaksh Prabhav Rao, Vivek Ra New York University,TecVisual Servoing
Peng Wang, Yongcai Wang, Deying Li Renmin University of ChVisual Servoing
Shaoting Peng, Margaret Wang, Julie A. Shah, Nadia FUniversity of Pennsylva Visual Tracking
Wen-hsuan Chu, Adam Harley, Pavel Tokmakov, AchalCarnegie
D Mellon UniversVisual Tracking
Shihao Feng, Pengpeng Liang, Jin Gao, Erkang ChengZhengzhou University,InVisual Tracking
Xinglong Sun, Adam Harley, Leonidas Guibas Stanford & UIUC,StanforVisual Tracking
Sandro Papais, Robert Ren, Steven Lake Waslander University of Toronto Visual Tracking
Chang Won Lee, Steven Lake Waslander University of Toronto Visual Tracking
Kevin Chappellet, Masaki Murooka, Guillaume Caron, CNRS,AIST,National In Visual Tracking
Qingmao Wei, Bi Zeng, Jianqi Liu, Li He, Guotian ZengGuandong University ofVisual Tracking
Jianjing Sun, Shuang Wu, Jun Dong, Junming He Anhui University,Hefei Visual-Inertial SLA
Haolong Li, Joerg Stueckler Max Planck Institute for Visual-Inertial SLA
Wenzhe Yang, Yan Zhuang, Dongting Luo, Xuetao Zh Dalian University of Te Visual-Inertial SLA
Jun Hu, Xiaoming Lang, Feng Zhang, Yinian Mao, Gu Meituan Inc,Meituan,MeiVisual-Inertial SLA
Weijian Xie, Guanyi Chu, Quanhao Qian, Yihao Yu, S Zhejiang University, S Visual-Inertial SLA
Matthew Lisondra, Junseo Kim, Riku Murai, Kourosh ZaToronto Metropolitan UnVisual-Inertial SLA
Weipeng Guan, Peiyu Chen, Yuhan Xie, Peng Lu The University of Hong Visual-Inertial SLA
Dhruv Kumar, Shishir Gopinath, Karthik Dantu, Steve KSimon Fraser University,Visual-Inertial SLA
Yifu Wang, Yonhon Ng, Inkyu Sa, Alvaro Joaquin ParraTencent,The University V o isual-Inertial SLA
liu xiaotong, Binglu Wang, Zhijun Li University of Science a Wearable Robotics
Sojiro Sugiura, Jayant Unde, Yaonan Zhu, Yasuhisa Nagoya University Wearable Robotics
Sajeeva Abeywardena, Ildar Farkhatdinov University of Surrey,Qu Wearable Robotics
Luka Miskovic, Tilen Brecelj, Miha Dezman, Tadej PetriJožef Stefan Institute,Wearable Robotics
Zhixin Tu, Yijun Fang, Yuquan Leng, Chenglong Fu Southern University of Wearable Robotics I
Yu Chen, Gong Chen, Jing Ye, Xiangjun Qiu, Xiang Li Tsinghua University,Sh Wearable Robotics
Yu Chen, Minda Wang, Yifan Wang Nanyang Technological Wearable Robotics
Taeyeon Kim, Eunseok Song, Seongbin An, Hyunjin ChKorea Advanced Institu Wearable Robotics I
Hongwu Li, Junchen Liu, Ziqi Wang, Haotian Ju, Tianj Harbin Institute of Tech Wearable Robotics I
Shubhranil Sengupta, Jee-Hwan Ryu Korea Advanced InstitutWearable Robotics I
Rami Jradi, Hala Rifai, Samer Mohammed UPEC,University of Paris Wearable Robotics I
Oussama Bey, Rami Jradi, Huiseok Moon, Hala Rifai, University Paris-Est CréWearable Robotics I
Junzhe Su, Masahiro Hirano, Yuji Yamakawa The University of TokyoWheeled Robots
Zhi-ren Chen, Wei-shun Yu, Pei-Chun Lin NTU,National Taiwan UnWheeled Robots
Pei-chun Huang, I-Chia Chang, Wei-shun Yu, Pei-ChunNational Taiwan UniversWheeled Robots
Adwait Mane, Christian Hubicki FAMU-FSU College of Engi Wheeled Robots
Victor Klemm, Yvain De Viragh, David Rohr, Roland S ETH Zurich,Inria RenneWheeled Robots
Tyler Bishop, Keran Ye, Konstantinos Karydis University of California, Wheeled Robots
Sang-Yun Jeon, Rakjoon Chung, Dongjun Lee Seoul National Universi Wheeled Robots
Botian Sun, Qinglin Lang, Minghe Li, Xuefeng Wang Peking University Wheeled Robots
Kai Yan Beihang University Wheeled Robots
Francesco Roscia, Michele Focchi, Andrea Del Prete, DIstituto Italiano di Tecn Whole-Body Motion
Alberto Rigo, Muqun Hu, Satyandra K. Gupta, Quan N University of Southern CWhole-Body Motion
Rosario Scalise, Ege ÇAäÿlar, Byron Boots, Chad C. University of WashingtoWhole-Body Motion
Tasuku Makabe, Kei Okada, Masayuki Inaba The University of TokyoWhole-Body Motion
Grazia Zambella, Robert Schuller, George Mesesan, Anto TU Wien,German Aerospac Whole-Body Motion
Yiming Li, Yan Zhang, Amirreza Razmjoo, Sylvain Cali Idiap Research Institut Whole-Body Motion
Xuwei Wu, Alin Albu-Schäffer, Alexander Dietrich German Aerospace Cent Whole-Body Motion
Tifanny Portela, Gabriel Margolis, Yandong Ji, Pulkit A EPFL,Massachusetts Ins Whole-Body Motion
Donghoon Baek, Yu-chen (johnny) Chang, Joao Ramo University of Illinois U Whole-Body Motion
Abstract
When used in a real-world noisy environment, the capacity to generalize to multiple domains is essential for any au
Point cloud is a popular and widely used geometric representation, which has attracted significant attention in 3D vi
In scenarios of Human-Robot Interaction (HRI), it is often assumed that the robot should cooperate with the closest
In large-scale metric localization, an incorrect result during retrieval will lead to an incorrect pose estimate or loop c
This paper presents Action-SGFA, a novel action feature alignment approach to learn unified joint embeddings acro
With innovation in fields such as autonomous driving and augmented reality, point cloud-based place recognition ha
Correlation filter (CF)-based approaches have gained widespread attention in the field of unmanned aerial vehicle (
As robotic systems increasingly encounter complex and unconstrained real-world scenarios, there is a demand to r
Vision transformers have demonstrated impressive performance in various robotics and automation applications, su
Series Elastic Actuator-based exoskeleton can offer precise torque control and transparency when interacting with
Actuation concepts such as Series Elastic Actuation (SEA), Parallel Elastic Actuation (PEA), and Biarticular Actuatio
The safer interaction between humans and robots is one of the challenges in robotics. To protect humans and robo
Variable stiffness actuator (VSA) can imitate natural muscles in their compliance capbility, which can provide flexibl
Elastic actuation taps into elastic elements' energy storage for dynamic motions beyond rigid actuation. While Serie
Recent trends in robotic actuation have highlighted the need for low-cost, high performance, and efficient gearing. W
A novel origami-inspired foldable pneumatic actuator is proposed in this letter to satisfy the comprehensive requirem
A pneumatic linear actuator is presented and evaluated. Designed to operate in demanding environments such as M
Sitting down and standing up from a chair and, similarly, moving heavy objects up and down between factory lines a
This letter addresses the problem of trajectory planning in a marsupial robotic system consisting of an unmanned a
Unmanned aerial vehicles (UAVs) must fly at low altitudes to execute certain missions when operating in complex m
Global navigation satellite systems (GNSS) denied environments/conditions require unmanned aerial vehicles (UAV
Ground testing is of paramount importance to verify and validate space operations and the associated control algor
Sub-SI{50}{gram} nano-drones are gaining momentum in both academia and industry. Their most compelling applic
This article proposes a novel algorithm for aerobatic trajectory generation for a vertical take-off and landing (VTOL)
In nature, when encountering unexpected uncertainty, animals tend to react quickly to ensure safety as the top prio

Nano-size drones hold enormous potential to explore unknown and complex environments. Their small size makes
Robotic sketching in real-world scenarios poses a challenging problem with diverse applications in art, robotics, and
Playing table tennis with a human player is a challenging robotic task due to its dynamic nature. Despite a number o
Existing robotic musicians utilize detailed handcrafted instrument models to generate or learn policies for playing be
The Michigan Robotics MBot is a low-cost mobile robot platform that has been used to train over 1,400 students in
Multirotors flying in close proximity induce aerodynamic wake effects on each other through propeller downwash. C
We study how to design learning-based adaptive controllers that enable fast and accurate online adaptation in chan
In recent years, wide-area visual detection systems have been widely applied in various industrial and security sect
Autonomous robot navigation within the dynamic unknown environment is of crucial significance for mobile robotic a
Large language models (LLMs) are accelerating the development of language-guided robot planners. Meanwhile, s
Accurate disturbance estimation is essential for safe robot operations. The recently proposed neural moving horizon
Spotting blemished areas automatically on a textured surface is a particular challenge, as both nominal and defectiv
Generalizable articulated object manipulation is essential for home-assistant robots. Recent efforts focus on imitatio
Motivated by the substantial achievements of Large Language Models (LLMs) in the field of natural language proce
Abstract—Olfaction sensing in autonomous robotics faces challenges in dynamic operations, energy efficiency, an
Effective interaction modeling and behavior prediction of dynamic agents play a significant role in interactive motion
Scene flow estimation determines a scene's 3D motion field, by predicting the motion of points in the scene, especi
Effective coordination is crucial for motion control with reinforcement learning, especially as the complexity of agent
Ethically compliant autonomous systems (ECAS) are the prevailing approach to building robotic systems that perfor
In recent years, reinforcement learning and imitation learning have shown great potential for controlling humanoid r
Enabling bipedal walking robots to learn how to maneuver over highly uneven, dynamically changing terrains is cha
When deploying modern machine learning-enabled robotic systems in high-stakes applications, detecting distributio
Online decision making under uncertainty in partially observable domains, also known as Belief Space Planning, is
This paper focuses on the problem of detecting and reacting to changes in the distribution of a sensorimotor contro
State estimation is crucial for the performance and safety of numerous robotic applications. Among the suite of estim
Deep learning models have revolutionized various domains but have also raised concerns regarding their security a
As life expectancy increases, social and health assistance requires sustainable and affordable solutions possibly us
Cooking robots can enhance the home experience by reducing the burden of daily chores. However, these robots m
Object rearrangement is pivotal in robotic-environment interactions, representing a significant capability in embodie
A flurry of recent work has demonstrated that pre-trained large language models (LLMs) can be effective task plann
Humans interpret scenes by recognizing both the identities and positions of objects in their observations. For a robo
This work stages Foosball as a versatile platform for advancing scientific research, particularly in the realm of robot
The language-conditioned robotic manipulation aims to transfer natural language instructions into executable action
Large Language Models (LLMs) have demonstrated the ability to perform semantic reasoning, planning and write c
We propose and demonstrate a compositional framework for training and verifying reinforcement learning (RL) syst
The hierarchy of global and local planners is one of the most commonly utilized system designs in autonomous rob
In visual SLAM (VSLAM) systems, loop closure plays a crucial role in reducing accumulated errors. However, VSLA
In this letter, a paradigm for the classification and manipulation of novel objects is established and demonstrated th
Generalist robot manipulators need to learn a wide variety of manipulation skills across diverse environments. Curre
Recent planning methods based on Large Language Models typically employ the In-Context Learning paradigm. Co
As autonomous driving technology matures, end-to-end methodologies have emerged as a leading strategy, promis
For effective human-robot interaction, robots need to understand, plan, and execute complex, long-horizon tasks de
We present Adaptive Skill Coordination (ASC) – an approach for accomplishing long-horizon tasks like mobile pic
Artificial cognitive architectures traditionally rely on complex memory models to encode, store, and retrieve informat
Aerial robots have the potential to play a crucial role in assisting humans with complex and dangerous tasks. Never
In the rapidly evolving landscape of Unmanned Aerial Vehicles (UAVs), regulation enforcement is critical. Unfortuna
Existing motion planning approaches for over-actuated unmanned aerial vehicle (UAV) platforms can achieve online
In this paper, an Active Cooperative Localization system for Quadrotor Unmanned Aerial Vehicles is developed. Th
A recent advance in vector field path following is the introduction of the Parametric Guiding Vector Field method. It
Automatic landing of tethered unmanned aerial vehicles (UAVs) is an important issue. Typically, UAVs rely on locat
Drone applications require perception all around the vehicle for obstacle avoidance during drone navigation. Due to
We propose a new approach to visual perception for relative localization of agents within large-scale swarms of UA
Autonomous exploration in dynamic environments necessitates a planner that can proactively respond to changes a
We propose an interpretable framework for reading analog gauges that is deployable on real world robotic systems
Estimating the region of attraction RoA for a robot controller is essential for safe application and controller composit
Model-predictive control (MPC) is a powerful tool for controlling highly dynamic robotic systems subject to complex
This study proposes a novel movable microfluidic chip in which a microfluidic chip is integrated into a robotic manip
Parameter convergence in adaptive control is crucial for improving the stability and robustness of robotic systems. N
We present Sim-on-Wheels, a safe, realistic, and vehicle-in-loop framework to test autonomous vehicles’ perfo
Generating safety-critical scenarios is essential for testing and verifying the safety of autonomous vehicles. Traditio
In this work, we present a learning method for lateral and longitudinal motion control of an ego-vehicle for vehicle pu
To operate in open-ended environments where humans interact in complex, diverse ways, autonomous robots mus
In light of the challenges and costs of real-world testing, autonomous vehicle developers often rely on testing in sim
Recent advancements in large language models (LLMs) have enabled a new research domain, LLM agents, for sol
In autonomous driving, leveraging cooperative perception through vehicle-to-vehicle (V2V) communication is consid
Accurate trajectory prediction is crucial for safe and efficient autonomous driving, but handling partial observations p
Multi-agent motion prediction is a crucial concern in autonomous driving, yet it remains a challenge owing to the am
This work proposes a new representation of pedestrian crossing scenarios and a hybrid modeling approach, RePed
Multi-Robot and Multi-Agent Systems demonstrate collective (swarm) intelligence through systematic and distribute
In this article, we present a novel approach, namely distributed matching-by-clone hungarian-based algorithm (DMC
A novel deep neural network (DNN) based controller for omnidirectional robots is proposed. The controller decompo
We present LGX (Language-guided Exploration), a novel algorithm for Language-Driven Zero-Shot Object Goal Na
Deciding where to go next is a challenging task for humans. However, for robots in unknown environments, this bec
Large Language Models (LLMs) have been widely utilized to perform complex robotic tasks. However, handling ext
Modeling the interactions among agents based on their historical trajectories is key to precise multi-agent trajectory
Large Language Models (LLMs) have shown promise in the autonomous driving sector, particularly in generalizatio
In many intelligent transportation systems, predicting the future motion of heterogeneous traffic participants is a fun
This paper presents a framework for odometry estimation in automotive application using six asynchronously opera
Achieving reliable autonomous navigation during nighttime remains a substantial obstacle in the field of robotics. Al
In recent years, imitation-based driving planners have reported considerable success. However, due to the absence
An accurate road topological structure is of great importance for autonomous driving in complex urban environment
Navigating complex and dynamic environments requires autonomous vehicles (AVs) to reason about both visible an
Accurate and robust localization remains a significant challenge for autonomous vehicles. The cost of sensors and
Accurate prediction is important for operating an autonomous vehicle in interactive scenarios. Prediction must be fa
Generating safe and non-conservative behaviors in dense, dynamic environments remains challenging for automat
Trajectory planning is one of the key challenges to the rapid and large-scale deployment of autonomous driving. Th
Most conventional wheeled robots can only move in flat environments and simply divide their planar workspaces int
Empowering robots to navigate in a socially compliant manner is essential for the acceptance of robots moving in h
Target tracking is a fundamental task in the domain of robotics. The effectiveness of target tracking hinges upon va
This work introduces Talk2BEV, a large vision-language model (LVLM) interface for bird’s-eye view (BEV) map
The effective reasoning of integrated multimodal perception information is crucial for achieving enhanced end-to-en
Although visual navigation has been extensively studied using deep reinforcement learning, online learning for real-
This paper introduces UniGen, a novel approach to generating new traffic scenarios for evaluating and improving a
LiDAR super-resolution is a relatively new problem in which we seek to fill in the blanks between measured points w
In this paper, we present the Circular Accessible Depth (CAD), a robust traversability representation for an unmann
Anticipating the motion of all humans in dynamic environments such as homes and offices is critical to enable safe
While existing Learning from intervention (LfI) methods within the human-in-the-loop reinforcement learning (HiL-RL
Vehicle planning is receiving increasing attention with the emergence of diverse driving simulators and large-scale d
In complex unstructured environments, accurate terrain traversability analysis is a fundamental requirement for the
Infrastructure-based vehicle localization, in comparison to single-agent approaches, offers several advantages inclu
Searching Gas Source actively and efficiently in unknown hazard environments is an important but challenging issu
Image-goal navigation in unknown environments with sensor error is of considerable difficulty for autonomous robot
We introduce OpenBot-Fleet, a comprehensive open-source cloud robotics system for navigation. OpenBot-Fleet u
Widely adopted motion forecasting datasets substitute the observed sensory inputs with higher-level abstractions s
The use of industrial robots represents a key technology for increasing productivity and efficiency in manufacturing.
The existing internet-scale image and video datasets cover a wide range of everyday objects and tasks, bringing th
The grand aim of having a single robot that can manipulate arbitrary objects in diverse settings is at odds with the p
We introduce Dream2Real, a robotics framework which integrates vision-language models (VLMs) trained on 2D da
The pre-train and fine-tune paradigm in machine learning has had dramatic success in a wide range of domains be
Motion forecasting has become an increasingly critical component of autonomous robotic systems. Onboard compu
Robot software is abundant with variables that represent real-world physical units (e.g., meters, seconds). Operatio
In this work, we achieved self-takeoff of an eagle-scale flapping-wing robot for the first time. Inspired by the takeoff
Flippers are important components for improving the locomotion efficiency and stability of bionic underwater robots.
A model-based approach for lateral maneuvering of flapping wing UAVs in closed spaces is presented. Bird-size or
Fish locomotion which adopts body and/or caudal fin swimming mode consists of different motions, such as Cruisin
Moving on natural muddy terrains, where soil composition and water content vary significantly, is complex and chal
By living and moving in groups, fish can gain many benefits, such as heightened predator detection, greater hu
An increasing number of underwater robots inspired by Carangidae are developed, which is characterized by high e
Tuning body stiffness like fish to improve swimming efficiency and speed has been adopted by many fish-inspired r
Fish schools present high-efficiency group behaviors to collective migration and dynamic escape from the predator
We present a framework for learning visually-guided quadruped locomotion by integrating exteroceptive sensing an
In this paper we adapt the theory of form closure to define the form closed region, i.e., the subset of a snake robot's
Decimeter scale robots in human environments are small relative to obstacles they encounter, making them prone t
Legged robots are becoming increasingly agile in exhibiting dynamic behaviors such as running and jumping. Usua
Scientists have long theorized that the cheetah’s tail contributes to its impressive maneuvrability at high speeds
Animals possessing spinal columns often exhibit exceptional agility for highly dynamic locomotion. The spine grants
This paper presents the design and analysis of Pegasus, a quadrupedal wheeled robot grounded in biomimicry prin
In the natural world, insects exhibit remarkable locomotion capabilities through a combination of running and jumpin
Subterranean burrowing is inherently difficult for robots because of the high forces experienced as well as the high
Insect-scale micro-aerial vehicles (MAV) are becoming increasingly important for sensing and mapping spatially con
Jumping is a key locomotion for miniature robots, but it is difficult for a robot to jump a long distance without flipping
Springtails are tiny arthropods that crawl and jump. They jump by temporarily storing elastic energy in resilin elastic
Several insect species are able to locomote across the air-water interface by leveraging surface tension to remain a
This paper presents the VLEIBot* (Very Little Eel-Inspired roBot), a 45-mg/23-mm3 microrobotic swimmer that is pr
Insects have long been recognized for their ability to navigate and return home using visual cues from their nest's e
Dragonflies show impressive flying skills by achieving both high efficiency and agility. They can perform distinctive f
Flying microrobots have garnered growing research interest owing to their technological intricacies and suitability fo
Embodying in vitro biological neural networks (BNNs) with robots to explore the rise of intelligence in these simpler
Continuum robot arms that can access confined spaces are useful in many applications, such as invasive surgery,
Current multi-modal hybrid robots with flight and wheeled modes have fallen into the dilemma that they can only av
Sidewinding, a locomotion strategy characterized by the coordination of lateral and vertical body undulations, is freq
Inertial appendages (e.g., tails and reaction wheels) have shown their reorientation capability to enhance robots' mo
While many inkjet printers employ only four types of ink (i.e. CKMY) to produce a wide range of colors, numerous te
The cable-driven redundant manipulator (CDRM) has significant potential for applications in narrow and hazardous
The prevalence of ineffective corrective surgeries for ocular motor disorders calls for a robotic eye platform in aiding
Hand-eye calibration is a critical task in robotics, as it directly affects the efficacy of critical operations such as mani
Extrinsic calibration for LiDAR and camera is an essential prerequisite for sensor fusion. Recently, automatic and ta
Accurate estimation of stereo camera extrinsic parameters is crucial to guarantee the performance of stereo match
The relative poses of visual perception sensors distributed over a vehicle’s body may vary due to dynamic force
RGB-D cameras are crucial in robotic perception, given their ability to produce images augmented with depth data.
As autonomous systems evolve, the precise calibration of lidar and camera sensors remains a pivotal concern. Am
Accurate and reliable sensor calibration is critical for fusing LiDAR and inertial measurements in autonomous drivin
Extrinsic calibration is an essential prerequisite for the applications of camera-LiDAR fusion. Existing methods eithe
Accurate calibration is crucial for using multiple cameras to triangulate the position of objects precisely. However, it
LiDAR-camera fusion plays a pivotal role in 3D reconstruction for self-driving applications. A fundamental prerequis
Collaborative robots have promising potential for widespread use in small-and-medium-sized enterprise (SME) man
Robots need to estimate the material and dynamic properties of objects from observations in order to simulate them
In autonomous driving, the LiDAR-Camera system plays a crucial role in a vehicle's perception of 3D environments
Hand-eye calibration is the problem of solving the transformation from the end-effector of a robot to the sensor atta
Many robots utilize commercial force/torque sensors to identify inertial properties of unknown objects. However, suc
The fusion of LiDARs and cameras has been increasingly adopted in autonomous driving for perception tasks. The
Calibrating the extrinsic parameters of sensory devices is crucial for fusing multi-modal data. Recently, event came
As a new sensor, 4D radar (x, y, z, velocity) has great potential for V2X, due to its 3D point cloud, direct doppler ve
Reinforcement learning (RL) for bipedal locomotion has recently demonstrated robust gaits over moderate terrains
Autonomous robots must navigate reliably in unknown environments even under compromised exteroceptive perce
Understanding how humans leverage semantic knowledge to navigate unfamiliar environments and decide where to
Robotic learning for navigation in unfamiliar environments needs to provide policies for both task-oriented navigatio
Geometric regularity, which leverages data symmetry, has been successfully incorporated into deep learning archit
Collision checking is the primitive operation of motion planning that consumes most time. Machine learning algorithm
Autonomous mobile robots require predictions of human motion to plan a safe trajectory that avoids them. Because
To explore safe interactions between a mobile robot and dynamic obstacles, this paper presents a comprehensive a
This paper proposes a GPU-accelerated optimization framework for collision avoidance problems where the contro
The latest robot navigation methods for dynamic environments assume that the states of obstacles, including their g
Collision Avoidance/Mitigation System (CAMS) for autonomous vehicles is a crucial technology that ensures the sa
We interleave sampling based motion planning methods with pruning ideas from minimum spanning tree algorithms
It is challenging for the mobile robot to achieve autonomous and mapless navigation in the unknown environment w
In this work, we present DiPPeR, a novel and fast 2D path planning framework for quadrupedal locomotion, leverag
Collision-avoiding motion planning for articulated robotic arms is one of the major challenges in robotics. The difficu
 Robot navigation using deep reinforcement learning (DRL) has shown great potential in improving the performance
The past few years have seen immense progress on two fronts that are critical to safe, widespread mobile robot de
We present an efficient method for certifying non-collision for piecewise-polynomial motion plans in algebraic repara
Image compression is increasingly important in applications like intelligent driving and smart surveillance systems. T
Path planning in a changing environment is a challenging task in robotics, as moving objects impose time-dependen
Model predictive control (MPC) may provide local motion planning for mobile robotic platforms. The challenging asp
In this paper, we present a probabilistic and unconstrained model predictive control formulation for robot navigation
We present MIM (Multi-Layer Intensity Map), a novel 3D object representation for robot perception and autonomous
Efficient navigation through uneven terrain remains a challenging endeavor for autonomous robots. We propose a n
Collision is typically avoided in robot navigation for safety guarantee. However, when a robot's exteroceptive senso
We aim to solve the problem of data-driven collision-distance estimation given 3-dimensional (3D) geometries. Con
Line-Of-Sight (LOS) check plays a crucial role in collision avoidance and time comsuming, particularly in scenarios
The interest in using reinforcement learning (RL) controllers in safety-critical applications such as robot navigation a
We present a novel trajectory traversability estimation and planning algorithm for robot navigation in complex outdo
A hyper-redundant manipulator(HRM) can flexibly accomplish tasks in narrow spaces. However, its excessive degr
This paper presents a novel extension of the Control Barrier Function (CBF) as the low-level safety controller for au
Recent advances in vision-language models (VLMs) have led to improved performance on tasks such as visual que
Neural Radiance Fields (NeRFs) are gaining significant interest for online active object reconstruction due to their e
Active object reconstruction using autonomous robots is gaining great interest. A primary goal in this task is to maxi
3D object detection models that exploit both LiDAR and camera sensor features are top performers in large-scale a
Reconstructing real-world objects and estimating their movable joint structures are pivotal technologies within the fi
Moving object segmentation (MOS) provides a reliable solution for detecting traffic participants and thus is of great
3D scene reconstruction from 2D images has been a long-standing task. Instead of estimating per-frame depth map
We present Normal Field Learning (NFL), a robust yet practical solution to perceive 3D layouts of transparent objec
Precise reconstruction and manipulation of the crumpled cloths is challenging due to the high dimensionality of cloth
Optical flow estimation is very challenging in situations with transparent or occluded objects. In this work, we addres
er Vision for Transportation
The scene flow estimation aims at accurately achieving the motion of 3D points, imposing challenges like mis-regist
With the increasing demands for perception accuracy in autonomous driving, there is a growing focus on fine-graine
Inspired by frame-based methods, state-of-the-art event-based optical flow networks rely on the explicit constructio
Collaborative perception leverages rich visual observations from multiple robots to extend a single robot's perceptio
Generative modeling of 3D LiDAR data is an emerging task with promising applications for autonomous mobile robo
Real-time processing is crucial in autonomous driving systems due to the imperative of instantaneous decision-mak
There is a recent need in the LiDAR perception field for unifying multiple tasks in a single strong network with impro
Owing to their tremendous adaptability to free-form objects, soft grippers with fin-ray structures have a wide range o
In typical in-hand manipulation tasks represented by object pivoting, the real-time perception of rotational slippage h
Non-flat surfaces pose difficulties for robots operating in unstructured environments. Reconstructions of uneven sur
Control techniques like MPC can realize contact-rich manipulation which exploits dynamic information, maintaining
In this paper, we present the TacShade: a newly designed 3D-printed soft optical tactile sensor. The sensor is deve
In robotics, it's crucial to understand object deformation during tactile interactions. A precise understanding of defor
This paper addresses the problem of pushing manipulation with nonholonomic mobile robots. Pushing is a fundame
In this article, we define "robotic contact juggling" to be the purposeful control of the motion of a 3-D smooth object
Reliable grasping and manipulation in daily tasks and unstructured environments require accurate contact modeling
To assist with everyday human activities, robots must solve complex long-horizon tasks and generalize to new setti
Prior to the deployment of robotic systems, pre-training the deep-recognition models on all potential visual cases is
Lifelong learning or continual learning is the problem of training an AI agent continuously while also preventing it fro
Continual reinforcement learning, which aims to help robots acquire skills without catastrophic forgetting, obviating
Open-world robotic tasks such as autonomous driving pose significant challenges to robot control due to unknown a
Large Language Models (LLMs) have emerged as a new paradigm for embodied reasoning and control, most recen
Large Language Models (LLMs) have been shown to act like planners that can decompose high-level instructions in
The sense of touch is essential for robots to perform various daily tasks. Artificial Neural Networks have shown sign
We introduce LOTUS, a continual imitation learning algorithm that empowers a physical robot to continuously and e
Micro tandem bi-copters are capable of passing through narrow gaps owing to their particular slender shape. Howe
We present ArrayBot, a distributed manipulation system consisting of a 16x16 array of vertically sliding pillars integr
Industrial robots are designed as general-purpose hardware with limited ability to adapt to changing task requireme
Cleaning glass surfaces is a prevailing maintenance problem in high-rise buildings. In the traditional methods of cle
Manipulating deformable objects arises in daily life and numerous applications. Despite phenomenal advances in in
For a multi-robot team that collaboratively explores an unknown environment, it is of vital importance that the collec
Perception is necessary for autonomous navigation in an unknown area crowded with obstacles. It’s challengin
It is promising but challenging to design flocking control for a robot swarm to autonomously follow changing pattern
Recent advancements in age-related macular degeneration treatments necessitate precision delivery into the subre
Panoramic cameras offer a 4Ï€ steradian field of view, which is desirable for tasks like people detection and tracking
An object handover between a robot and a human is a coordinated action which is prone to failure for reasons such
We introduce the CEA Industrial Multi-Scale Object Localization Dataset (CEA-IMSOLD), a new BOP format datase
Semantic Scene Completion (SSC) aims to jointly generate space occupancies and semantic labels for complex 3D
Despite the substantial progress in deep learning, its adoption in industrial robotics projects remains limited, primari
Dynamic environments in factories, surgical robotics, and warehouses increasingly involve humans, ma- chines, rob
Numerous datasets and benchmarks exist to assess and compare Simultaneous Localization and Mapping (SLAM)
In robotic vision, a de-facto paradigm is to learn in simulated environments and then transfer to real-world applicatio
Most existing robotic datasets capture static scene data and thus are limited in evaluating robots' dynamic performa
In collaborative human-robot manipulation, a robot must predict human intents and adapt its actions accordingly to
Machine Learning (ML) has replaced traditional handcrafted methods for perception and prediction in autonomous v
Social navigation and pedestrian behavior research has shifted towards machine learning-based methods and conv
We present a scalable, bottom-up and intrinsically diverse data collection scheme that can be used for high-level re
Corner case scenarios are an essential tool for testing and validating the safety of autonomous vehicles (AVs). As t
A key challenge for robotic manipulation in open domains is how to acquire diverse and generalizable skills for robo
Machine learning provides a powerful tool for building socially compliant robotic systems that go beyond simple pre
Intelligent driving systems aim to achieve a zero-collision mobility experience, requiring interdisciplinary efforts to en
Autonomous navigation on inland waters requires an accurate understanding of the environment in order to react to
Intelligent vehicle systems require a deep understanding of the interplay between road conditions, surrounding entit
Over the past few years, there has been remarkable progress in research on 3D point clouds and their use in auton
Accurately localization between multi-robots is crucial for many underwater applications, such as tracking, convoyin
The potential for deploying autonomous systems can be significantly increased by improving the perception and int
Robotic grasping is one of the essential skills in robotics. From industrial to housework, robots are required to hand
Visual Place Recognition (VPR) in indoor environments is beneficial to humans and robots for better localization an
Data collection for forestry, timber, and agriculture currently relies on manual techniques which are labor-intensive a
3D point clouds are widely used for robot perception and navigation. LiDAR sensors can provide large scale 3D poi
Point cloud primitive segmentation aims to segment the surface point cloud into various geometric types of primitive
Cross-modality point cloud registration is confronted with significant challenges due to inherent differences in moda
The ability to estimate joint parameters is essential for various applications in robotics and computer vision. In this p
Autonomous vehicles rely on LiDAR sensors to perceive the environment. Adverse weather conditions like rain, sno
Efficient interest point detection and description in images play a crucial role in many tasks such as multi-robot SLA
Point cloud registration is essential in computer vision and robotics. Recently, transformer-based methods have ach
A robot's ability to anticipate the 3D action target location of a hand's movement from egocentric videos can greatly
Since autonomous driving systems usually face dynamic and ever-changing environments, continual test-time adap
Slip detection plays a pivotal role in the dexterity of robotics, improving the reliability and precision of manipulations
We propose a novel 3-D human motion and object interaction prediction model that is aware of commonsense know
A robot self-model is a task-agnostic representation of the robot's physical morphology that can be used for motion
Affordance detection and pose estimation are of great importance in many robotic applications. Their combination h
In this paper, we propose a novel method for plane clustering specialized in cluttered scenes using an RGB-D came
This paper presents a novel strategy to train keypoint detection models for robotics applications. Our goal is to deve
Tremendous variations coupled with large degrees of freedom in UAV-based imaging conditions lead to a significan
Scene transfer for vision-based mobile robotics applications is a highly relevant and challenging problem. The utility
Detecting air flows caused by phenomena such as heat convection is valuable in multiple scenarios, including leak
Relative drone-to-drone localization is a fundamental building block for any swarm operations. We address this task
3D instance segmentation is a fundamental and critical task for enabling robots to operate effectively in unstructure
Precise scene understanding is key for most robot monitoring and intervention tasks in agriculture. In this work we p
Self-supervised depth estimation has evolved into an image reconstruction task that minimizes a photometric loss.
Estimating robot pose and joint angles is pivotal in advanced robotics, underpinning applications like robot collabora
Advanced Driver-Assistance Systems (ADAS) have successfully integrated learning-based techniques into vehicle
The ability to predict future structure features of environments based on past perception information is extremely ne
Despite its importance in both industrial and service robotics, mobile manipulation remains a significant challenge a
We present a method for generating, predicting, and using Spatiotemporal Occupancy Grid Maps (SOGM), which e
In policy learning for robotic manipulation, sample efficiency is of paramount importance. Thus, learning and extract
Trajectory prediction plays a crucial role in the autonomous driving stack by enabling autonomous vehicles to antici
Physical human-robot interaction has been an area of interest for decades. Collaborative tasks, such as joint compl
Recent work has shown the promise of creating generalist, transformer-based, models for language, vision, and se
Pedestrian trajectory prediction plays an important role in autonomous driving systems and robotics. Recent work u
Multi-Agent Path Finding (MAPF) is an important issue in multi-agent cooperation. Many studies apply Multi-Agent R
There has been a significant research interest in employing large language models to empower intelligent robots wi
Motion prediction has been an essential component of autonomous driving systems since it handles highly uncertai
This paper explores active sensing strategies that employ vision-based tactile sensors for robotic perception and cla
Bio-inspired neural networks have the potential to advance our understanding of neural computation and improve th
In recent years, numerous real-time stereo matching methods have been introduced, but they often lack accuracy.
3D reconstruction is an important task in computer vision and is widely used in robotics and autonomous driving. W
In this paper, we address the Sim2Real gap in the field of vision-based tactile sensors for classifying object surface
Integrating robots into populated environments is a complex challenge that requires an understanding of human soc
With the proliferation of sensor devices, the extensive utilization of three-dimensional data in multimedia continues
In this work we present CppFlow - a novel and performant planner for the Cartesian Path Planning problem, which
A critical goal of autonomy and artificial intelligence is enabling autonomous robots to rapidly adapt in dynamic and
Sophisticated models can accurately describe deformations of continuum robots while being computationally dema
Motion planning is a mature area of research in robotics with many well-established methods based on optimization
Object rearrangement is a fundamental problem in robotics with various practical applications ranging from managin
This study presents a benchmark for evaluating action-constrained reinforcement learning (RL) algorithms. In action
Existing grasp prediction approaches are mostly based on offline learning, while, ignoring the exploratory grasp lear
The prevailing grasp prediction methods predominantly rely on offline learning, overlooking the dynamic grasp learn
Visual-based robot pose estimation is a fundamen- tal challenge, involving the determination of the camera’s p
This paper introduces a method, Generative Adversarial Networks for Detecting Erroneous Results (GANDER), lev
Task-oriented grasping models aim to predict a suitable grasp pose on an object to fulfill a task. These systems hav
Automating the assembly of objects from their parts is a complex problem with innumerable applications in manufac
Advancements in robot learning for object manipulation have shown promising results, yet certain everyday objects
This paper focuses on benchmarking the robustness of visual representations toward policy learning in the context
Recent research efforts have yielded significant advancements in manipulating objects under homogeneous setting
The use of multi-camera views simultaneously has been shown to improve the generalization capabilities and perfo
While it is generally acknowledged that force feedback is beneficial to robotic control, applications of policy learning
Robots need to have a memory of previously observed, but currently occluded objects to work reliably in realistic en
Industrial bin picking for tangled-prone objects requires the robot to either pick up untangled objects or perform sep
Fabric manipulation is a long-standing challenge in robotics due to the enormous state space and complex dynamic
Robotic autonomous grasp requires the system to perform multiple functions such as gripper and robot control, mak
Robotic push-grasping in densely cluttered environments presents significant challenges due to unbalanced synerg
Shape servoing, a robotic task dedicated to controlling objects to desired goal shapes, is a promising approach to d
What makes generalization hard for imitation learning in visual robotic manipulation? This question is difficult to app
In this work, we present STOPNet, a framework for 6-DoF object suction detection on production lines, with a focus
For robots to be useful outside labs and specialized factories we need a way to teach them new useful behaviors qu
Many practically relevant robot grasping problems feature a target object for which all grasps are occluded, e.g., by
We investigate the active manipulation of objects using model-free and long-horizon DRL (Deep Reinforcement Lea
Articulated objects like cabinets and doors are widespread in daily life. However, directly manipulating 3D articulate
Bin picking is an important building block for many robotic systems, in logistics, production or in household use-cas
Multi-goal robot manipulation tasks with sparse rewards are difficult for reinforcement learning (RL) algorithms due
In this paper, we tackle the problem of grasping transparent and specular objects. This issue holds importance, yet
Offline learning usually requires a large dataset for training. In this paper, we focus on vision-based robotic manipul
Robotic manipulation requires accurate perception of the environment, which poses a significant challenge due to it
Manipulating unseen articulated objects through visual feedback is a critical but challenging task for real robots. Exi
We introduce MORPH, a method for co-optimization of hardware design parameters and control policies in simulati
Reinforcement learning solely from an agent's self-generated data is often believed to be infeasible for learning on r
Robotic affordances, providing information about what actions can be taken in a given situation, can aid robotic man
Despite the prevalence of robotic manipulation tasks in various real-world applications of different requirements and
Fast and robust object grasping in clutter is a crucial component of robotics. Most current works resort to the whole
Recently, several works achieve end-to-end visual servoing (VS) for robotic manipulation by replacing traditional co
We present a novel vision-based, 6-DoF grasping framework based on Deep Reinforcement Learning (DRL) that is
Point cloud prediction is an important yet challenging task in the field of autonomous driving. The goal is to predict f
The popular methods for semi-supervised semantic segmentation mostly adopt a unitary network model using conv
Routine visual inspections of concrete structures are imperative for upholding the safety and integrity of critical infra
This paper presents a novel Fully Binary Point Cloud Transformer (FBPT) network model which has the potential to
The millimeter-wave radar sensor maintains stable performance under adverse environmental conditions, making it
In many robotic applications, especially those involving humans and the environment, linguistic and visual informati
Robot manipulation relying on learned object-centric descriptors became popular in recent years. Visual descriptors
An accurate and generalizable dense 3D reconstruction system has attracted much attention. However, existing 3D
Deep learning methods for perception are the cornerstone of many robotic systems. Despite their potential for impre
Equipping multi-fingered robots with tactile sensing is crucial for achieving the precise, contact-rich, and dexterous
Controlling the shape of deformable linear objects using robots and constraints provided by environmental fixtures h
Robotic systems often face challenges when attempting to grasp a target object due to interference from surroundin
Tactile servoing is an effective approach to enabling robots to safely interact with unknown environments. One of th
In this work, we delve into the intricate synergy among non-prehensile actions like pushing, and prehensile actions
The ability to plan and control robotic in-hand manipulation is challenged by several issues, including the required a
Recent works on the pretraining for robot manipulation have demonstrated that representations learning from large
Precise perception of contact interactions is essential for fine-grained manipulation skills for robots. In this paper, w
Despite decades of steady research progress, the robotic hand is still far behind the human hand in terms of dexter
In this paper, we introduce a double-stage transfer learning framework based on expert data. It employs probabilist
Insertion is an essential skill for robots in both modern manufacturing and services robotics. In our previous study, w
Deformable linear object (DLO) manipulation is needed in many fields. Previous research on deformable linear obje
We provide the mechanical and dynamical requirements for a robotic finger capable of performing a large number o
We discuss the precise cooperative motion of a dual manipulator. In the inverse kinematics of cooperative redunda
Shape control of deformable objects is challenging and important. This paper proposes a model-free controller usin
The empirical success of Reinforcement Learning (RL) in contact-rich manipulation leaves much to be understood f
Planning robot dexterity is challenging due to the non-smoothness introduced by contacts, intricate fine motions, an
The ability to grasp and manipulate small objects in cluttered environments remains a significant challenge. This let
The reconfigurable unmanned aerial vehicle (RUAV) has the ability to create and break physical links to self-assem
Limiting latency is essential for critical robot applications such as collision avoidance or target tracking and is challe
Tethers have great potential in multi-robot systems from enabling retrieval of deployed robots and facilitating power
This paper proposes two decentralized multi-agent optimal control methods that combine the computational efficien
Coalition is an important mean of multi-robot systems to collaborate on common tasks. An adaptive coalition strateg
Multi-robot systems have increasingly become instrumental in tackling coverage problems. However, the challenge
We present new algorithms for Distributed Factor Graph Optimization (DFGO) problems that arise in the probabilist
This letter comprehensively investigates the performance of six state-of-art distributed task allocation algorithms (i.e
Due to the low motion efficiency and maneuverability of underwater robots with six degrees of freedom, it is challen
Morphing quadrotors that can be potentially applied to confined spaces such as warehouses, tanks, and pipelines h
In this work, we investigate task planning for mobile robots under linear temporal logic (LTL) specifications. This pro
This paper investigates the problem of designing control policies that satisfy high-level specifications described by s
The determination of the dynamic properties of a robot is especially important for designing highly accurate and effi
In this paper, we propose a new algorithm for solving the forward dynamics of multibody systems consisting of rigid
We derive a family of efficient constrained dynamics algorithms by formulating an equivalent linear quadratic regula
Co-simulation is a useful approach in the modelling of robotic systems composed of multiple parts. In co-simulation
Autonomous navigation in unstructured environments is a challenging task due to the complex and dynamic nature
Contact-rich manipulation tasks often exhibit a large sim-to-real gap. For instance, industrial assembly tasks freque
This paper proposes to use a high-compliance soft wrist to improve the performance of robotic peg-in-hole in uncer
Accurate 6D pose estimation of object is important for robot assembly. This letter presents a novel method for achie
The automated assembly of complex products requires a system that can automatically plan a physically feasible s
This paper addresses the problem of establishing efficient intralogistic systems, focusing on the generation of roadm
Multi-session mapping serves as the pre-requisite for autonomous robots to fulfill various long-term tasks (e.g., map
Graph-patrolling problems in the adversarial domain typically embed models and assumptions about how hostile ev
Graph patrolling algorithms provide effective strategies to coordinate mobile robots in the context of autonomous su
Centralized control of a multi-agent system improves upon distributed control especially when multiple agents share
As the usage of bi-copters increases in military and civilian fields, the demand for reliable bi-copters is on the rise. T
To speed up and reduce power consumption per cycle in robotic manipulation, one option is to exploit intentional co
During the operation of industrial robots, unusual events may endanger the safety of humans and the quality of prod
An autonomous service robot should be able to interact with its environment safely and robustly without requiring hu
Quadruped robots, in recent years, have been increasingly used in extremely harsh and dangerous conditions. Con
To create a self-repairing 3D printer, it must continue operating even after experiencing corruption. This work focus
Unmanned aerial vehicles (UAVs) are becoming an integral part of both industry and society. In particular, the quad
With the proliferation of autonomous vehicles (AVs) comes the need to ensure they abide by safe driving properties
In manufacturing, minimizing operational delays is crucial for efficiency and resilience. Therefore efficiently handling
Existing exploration algorithms mainly generate frontiers using random sampling or motion primitive methods within
Inspection of powerlines is a hard problem that requires humans to operate in remote locations and dangerous con
This paper introduces the open-source framework, GIRA, which implements fundamental robotics algorithms for rec
Autonomous terrain traversal of articulated tracked robots can reduce operator cognitive load to enhance task effici
LiDAR-based localization methods, such as the Iterative Closest Point(ICP) algorithm, can suffer in geometrically u
Efficiently performing intervention tasks underwater is crucial in various commercial and scientific sectors; however
Time-effective and accurate source localization with mobile robots is crucial in safety-critical scenarios, e.g. leakage
This paper presents a strategy for predicting the turning radius of a sailing robot with consideration of aerodynamic
Tunnel construction using the drill-and-blast method requires the 3D measurement of the excavation front to evalua
Solar power is becoming an increasingly popular option for energy production in commercial and private application
Automatic inspection of parabolic-trough solar plants is key to preventing failures that can harm the environment an
Limited resources and resulting energy crises occurring all over the world highlight the importance of energy efficien
Environmental monitoring via UAVs offers unprecedented aerial observation capabilities. However, the limited flight
Robotic gas distribution mapping improves the understanding of a hazardous gas dispersion while putting the huma
This paper proposes a novel decision-making framework for planning “when” and “where” to deploy r
The creation of precise and high-resolution crop point clouds in agricultural fields has become a key challenge for h
Emissions of particulate matter into the atmosphere are essential to characterize, in terms of properties such as pa
Accurate and robust navigation in unstructured environments requires fusing data from multiple sensors. Such fusio
Tactile sensing is significant for robotics since it can obtain physical contact information during manipulation. To cap
Tactile sensing plays a critical role in enabling robots to interact safely with target objects in dynamic and unstructur
Tactile exploration of surfaces is a key component of everyday life, allowing us to make complex inferences about o
Tactile sensing has become a popular sensing modality for robot manipulators, due to the promise of providing robo
Object pushing presents a key non-prehensile manipulation problem that is illustrative of more complex robotic man
In this paper, we propose X-lateral teleoperation: a novel hybrid unilateral-bilateral teleoperation framework. Bilatera
GelSight sensors that estimate contact geometry and force by reconstructing the deformation of their soft elastome
Compared to fully-actuated robotic end-effectors, underactuated ones are generally more adaptive, robust, and cos
Camera-based tactile sensors can provide high resolution positional and local geometry information for robotic man
Tactile sensing plays a pivotal role in human perception and manipulation tasks, allowing us to intuitively understan
Tactile sensing is a necessary capability for a robotic hand to perform fine manipulations and interact with the envir
The advancements in vision-based tactile sensors have boosted the aptitude of robots to perform contact-rich mani
Visual-tactile sensors that use a camera to capture the deformation of a soft gel layer have become popular in rece
Combining 3D vision with tactile sensing could unlock a greater level of dexterity for robots and improve several ma
In this work, we use MEMS microphones as vibration sensors to simultaneously classify texture and estimate conta
Soft tactile optical sensors have opened up new possibilities for endowing artificial robotic hands with advanced tou
Bimanual manipulation with tactile feedback will be key to human-level robot dexterity. However, this topic is less ex
Tactile exploration plays a crucial role in under- standing object structures for fundamental robotics tasks such as g
The outdoor environment has many uneven surfaces that put the robot at risk of sinking or tipping over. Recognizin
When manipulating an object to accomplish complex tasks, humans rely on both vision and touch to keep track of t
Robotic unfolding of cloth is challenging due to the wide range of textile materials and their ability to deform in unpre
Soft-bubble tactile sensors have the potential to capture dense contact and force information across a large contact
Vertical farming, a sustainable key for urban agriculture, has garnered attention for its land use optimization and en
Most vision-based tactile sensors use elastomer deformation to infer tactile information, which can not sense some
Tactile and textile skin technologies have become increasingly important for enhancing human-robot interaction and
Whole-arm tactile feedback is crucial for robots to ensure safe physical interaction with their surroundings. In this pa
Simulating tactile perception could potentially leverage the learning capabilities of robotic systems in manipulation ta
Accurate perception of the surrounding environment stands as a primary objective for robots. Through tactile intera
Data-driven methods have been successfully applied to images from vision-based tactile sensors to fulfill various m
We introduce a novel approach that combines tactile estimation and control for in-hand object manipulation. By inte
This work presents a novel shape evaluation and optimization approach for shape sensing, specifically targeting the
It is difficult for robots to retrieve objects in densely cluttered lateral access scenes with movable objects as jammin
Iontronic tactile sensors are promising to measure spatial-temporal contact information with high performance. How
Assistive robots have the potential to support disabled and elderly people in daily dressing activities. An intermediat
The rapidly evolving field of robotics necessitates methods that can facilitate the fusion of multiple modalities. Spec
Using tactile signal is a natural way to perceive potential dangers and safeguard robots. One possible method is to
Manipulation in tight environment is challenging but increasingly common in vision-guided robotic applications. The
The system observability analysis is of practical importance, for example, due to its ability to identify the unobservab
The sense of touch enables humans to perform many delicate tasks on deformable objects and/or in a vision-denie
Although the potential demand for force sensors in both robotics and automation is high, the complexity of their stru
Intraoperative contact sensing has the potential to reduce the risk of surgical errors and enhance manipulation capa
This paper presents a control framework for legged robots that enables self-perception and resistance to external d
This paper is concerned with the contact force estimation problem of robot manipulators based on imperfect dynam
Flexible manufacturing lines are required to meet the demand for customized and small batch-size products. Even t
In this paper we present a grammar and control synthesis framework for online modification of Event-based Signal
This paper introduces a sampling-based strategy synthesis algorithm for nondeterministic hybrid systems with comp
In this paper, we consider the problem of safety analysis of a closed-loop control system with anytime perception se
The robustness of signal temporal logic not only assesses whether a signal adheres to a specification but also prov
In the realm of robotics, numerous downstream robotics tasks leverage machine learning methods for processing, m
In this work, we address the problem of control synthesis for a homogeneous team of robots given a global tempora
This paper introduces an iterative approach to multi-agent route planning under chance constraints. A heterogeneo
Autonomous robots must utilize rich sensory data to make safe control decisions. To process this data, compute-co
In environments like offices, the duration of a robot's navigation between two locations may vary over time. For inst
Automaton based approaches have enabled robots to perform various complex tasks. However, most existing auto
As robots become more prevalent, the complexity of robot-robot, robot-human, and robot-environment interactions
Reactive planning enables the robots to deal with dynamic events in uncertain environments. However, existing me
This work develops a fast task allocation framework for heterogeneous multi-robot systems subject to both tempora
Deploying robots in real-world environments, such as households and manufacturing lines, requires generalization
This study developed a high-precision paint deposition model that considers the position and direction of a spray-pa
A robotic behavior model that can reliably generate behaviors from natural language inputs in real time would subst
This paper contemplates the possibility of asking robots questions and having them use their ability to go out into th

The grasping of transparent objects is challenging but of significance to robots. In this article, a visual-tactile fusion
Versatile and secure grasping in robotic systems remains a difficult challenge to address when objects possess a w
Objects we interact with and manipulate often share similar parts, e.g. handles, that allow us to transfer our actions
In robotic applications such as bin-picking or block-stacking, learned predictive models have been developed for ma
Grasping is one of the most fundamental skills for humans to interact with objects. However, it remains a challengin
We present FuncGrasp, a framework that can infer dense yet reliable grasp configurations for unseen objects using
Grasp Quality Networks are important components of grasping-capable autonomous robots, as they allow them to e
Robotic manipulation is a key enabler for automation in the fulfillment logistics sector. Such robotic systems require
In multi-object stacking scenarios, exploring the relationships among objects and determining the correct sequence
This paper presents a novel acoustic soft tactile (AST) skin technology operating with sound waves. In this innovati
Robotic grasping refers to making a robotic system pick an object by applying forces and torques on its surface. Ma
Grasping with five-fingered humanoid hands is a complex control problem. Throughout the entire grasping motion, a
We consider the problem of closed-loop robotic grasping and present a novel planner which uses Visual Feedback
This paper and its companion propose a new fractal robotic gripper, drawing inspiration from the century-old Fracta
Accurate grasping is the key to several robotic tasks including assembly and household robotics. Executing a succe
Advancing robotic grasping and manipulation requires the ability to test algorithms and/or train learning models on l
Robot learning methods have recently made great strides but generalization and robustness challenges still hinder
Recently, pre-trained vision models have gained significant attention in motor control, showcasing impressive perfo
Robotic systems that aspire to operate in uninstrumented real-world environments must perceive the world directly
This paper introduces a novel approach to improve robotic grasping in dynamic environments by integrating Gauss
The problem of grasping objects using a multi-finger hand has received significant attention in recent years. Howev
One goal of dexterous robotic grasping is to allow robots to handle objects with the same level of flexibility and adap
The disposal of waste electrical and electronic equipment (WEEE) presents a sustainability challenge, particularly fo
The progressive prevalence of robots in human suited environments has given rise to a myriad of object manipulati
Robotic grasping is fundamental to many real-world applications, and new approaches must be systematically evalu

This study proposes a novel robotic gripper driven by a single motor. The main task is to pick up objects in confined
Various sensorized grippers have been developed to handle delicate objects safely. These grippers have sensors m
The development of a robotic gripper for handling objects of various sizes, shapes, weights, and degrees of hardne
Industrial robots are capable of moving at high speed. Each time they come into contact with their environment, e.g
In this paper, we use impact-induced acceleration in conjunction with periodic stick-slip to successfully and quickly t
This paper presents a novel Bionic Soft Finger (BSF) that aims to overcome the limitations of conventional rigid ma
Small-scale robots hold significant promise in the field of minimally invasive surgery (MIS). In this paper, we presen
This paper presents a novel approach to designing, a low-cost gripper that is highly repeatable and functionally inte
While robotic assembly is a well researched topic, recycling and disassembly of products are also becoming ever m
This paper introduces a novel Fractal Hand robotic gripper. The hand has only 1 actuator, but (2^{n+1}-1) joints, wh
Food handling is a challenging task for robotic grippers, which are required to manipulate highly deformable and fra
A robotic system's hardware and control policy must be co-optimized to ensure they complement each other to inte
Squirrels exhibit agile leaping between tree branches, often using non-prehensile gripping with compliant and passi
The seafood processing industry provides fertile ground for robotics to impact the future-of-work from multiple persp
High cost and lack of reliability has precluded the widespread adoption of dexterous hands in robotics. Furthermore
Object grasping is an important ability required for various robot tasks. In particular, tasks that require precise force
We present a solution to the separate challenges faced by suction cups and gecko adhesives for one-sided graspin
Imitation learning methods have proven effective in learning robotic tasks by leveraging multiple human-controlled d
Recently, tactile sensing has attracted great interests in robotics, especially for exploring unstructured objects. Sens
Hand motion tracking is essential in many fields, e.g., immersive virtual reality, teleoperation of robotic hand, and ha
This study presents the Prosthetic Upper-Limb Sensory Enhancement (PULSE), a novel dual feedback device com
Surface vibration tactile feedback is capable of conveying various semantic information to humans via handheld ele
The integration of high-level assistance algorithms in surgical robotics training curricula may be beneficial in establis
Continuum robots are well-suited for applications in delicate and constrained environments, such as minimally invas
A miniature haptic stimulation device utilizes focused ultrasound to deliver a tactile haptic sensation to the finger. Th
Tactile perception is important for robotic systems that interact with the world through touch. Touch is an active sen
Shape-changing displays enable real-time visual- ization and haptic exploration of 3D surfaces. However, many sha
3D reconstruction based on monocular videos has attracted wide attention, and existing reconstruction methods us
Rehabilitative robotics, particularly lower-limb exoskeletons (LLEs), have gained increasing importance in aiding pa
This research contributes to the field of Human-Robot Collaboration (HRC) within dynamic and unstructured environ
Advanced Human-In-The-Loop (HITL) control strategies for robot hands based on surface electromyography (sEMG
Handing objects to humans is an essential capability for collaborative robots. Previous research works on human-ro
In the context of rapid advancements in industrial automation, vision-based robotic grasping plays an increasingly c
In a Human-in-the-Loop paradigm, a robotic agent is able to act mostly autonomously in solving a task, but can req
Assessing proficiency in small unmanned aerial vehicles (sUAVs) pilots is complex and not well understood, but inc
Work related musculoskeletal disorders (WMSDs) are very common. Repetitive motion, which is often present in in
Preference-based reinforcement learning (RL) poses as a recent research direction in robot learning, by allowing hu
Robots operating alongside humans often encounter unfamiliar environments that make autonomous task completi
Developing autonomous vehicles that can safely interact with pedestrians requires large amounts of pedestrian and
Shared electric micromobility has surged to a popular model of urban transportation due to its efficiency in short-dis
Drone racing has become a popular international competition and has attained wide attention in recent years. Howe
This work presents a reactive control strategy for loading and sudden unloading of an elbow effort-compensation de
It is difficult to guarantee an empty living environment to prevent unexpected contact between the object being man
An emerging trend in the field of human-robot collaboration is the disassembly of end-of-life products. Safety is a cr
Robots are increasingly finding their way into home environments, where they can assist with household tasks like
Predicting pedestrian movements remains a complex and persistent challenge in robot navigation research. We mu
Reinforcement learning (RL) methods for social robot navigation show great success navigating robots through larg
The effective deployment of robots in risky and crowded environments (RCE) requires the specification of robot pla
As robots become more versatile, fostering effective interactions with humans within shared spaces becomes param
With the increase of robots navigating through crowded environments in our daily lives, the demand for designing a
In highly interactive driving scenarios, the actions of one agent greatly influence those of its neighbors. Planning sa
Exploring topologically distinctive trajectories provides more options for robot motion planning. Since computing tim
Motion planning algorithms are important components of autonomous robots, which are difficult to understand and d
Most existing 6-DoF robot grasping solutions depend on strong supervision on grasp pose to ensure satisfactory pe
Collaboration between humans and robots is becoming increasingly crucial in our daily life. In order to accomplish e
Machine learning and deep learning have been used extensively to classify physical surfaces through images and t
In complex tasks (beyond a single targeted controller) requiring robots to collaborate with multiple human users, two
In Human-Robot Interaction scenarios, human factors like trust can greatly impact task performance and interaction
This paper proposes an interactive navigation framework by using large language and vision-language models, allo
In the past decade, there has been significant advancement in designing wearable neural interfaces for controlling n
In shared autonomy, human-robot handover for object delivery is crucial. Accurate robot predictions of human hand
Theory of Mind (ToM) is a fundamental cognitive architecture that endows humans with the ability to attribute menta
In this paper, building on recent advances in the fields of gaming AI and social robotics, we present a new approach
Participating in sports is of great significance to people’s physical and mental well-being. While physical activity
We previously proposed a motorcycle with a balance assist function that can generate a self-balancing moment by
Meal preparation is an important instrumental activity of daily living (IADL). While existing research has explored rob
This paper describes the use of a simulation-to-real training pipeline using autoregressive neural networks (MIntNe
In this paper, a method for generating a map from path information described using natural language (textual path)
Current methods of deploying robots that operate in dynamic, uncertain environments, such as Uncrewed Aerial Sy
Routine inspections for critical infrastructures such as bridges are required in most jurisdictions worldwide. Such rou
Integration of robots into scientific user facilities, such as the National Synchrotron Light Source II, improves their ef
Our goal is to design a framework for holistic evaluation of human-robot collaboration systems. To this end we utiliz
The human arm has an intrinsic capacity to absorb energy during physical human-robot interaction (pHRI), which ca
Language-guided active sensing is a robotics subtask where a robot with an onboard sensor interacts efficiently wit
Despite the increased use of lower limb exoskeletons as gait training and mobility assistive devices, their controllers
We propose, examine prototypes of, and collect user input on morphing robotic surface, “robot-room” eleme
Trust in automation is usually assessed with post-interaction questionnaires. For human robot collaboration it would
Human following for mobile robots has emerged as a promising technique with widespread applications. To ensure
This paper presents a Bayesian optimization framework for the automatic tuning of shared controllers which are def
Recent work has considered trust-aware decision making for human-robot collaboration (HRC) with a focus on mod
We focus on the problem of how we can enable a robot to collaborate seamlessly with a human partner, specifically
Despite advancements in human-robot teamwork, limited progress was made in developing AI assistants capable o
This paper addresses human-robot collaborative (HRC) precision assembly that complements natural human ability
Simultaneous Localization and Mapping (SLAM) has emerged as a prime autonomous mobile agent localization alg
The accumulation of a sufficient amount of data for training deep neural networks is a major hindrance in the applic
Mass customization and shorter manufacturing cycles are becoming more important among small and medium-size
The new face of modern industrial scenarios involves shared workspaces where humans and robots work closely to
While the traditional control of surgical robots relies on fully manual teleoperations, human-robot collaborative syste
In this paper we present SIREN: a novel audio-based communication system for underwater human-robot interactio
This article introduces a framework for complex human-robot collaboration tasks, such as the co-manufacturing of f
Autonomous assistance of people with motor impairments is one of the most promising applications of autonomous
While surgical robots, such as the da Vinci Surgical System, have become prevalent in minimally invasive surgeries
Today's robot policies exhibit subpar performance when faced with the challenge of generalizing to novel environme
During certain human-robot collaboration tasks, the operator interacts with the robot by hand guidance to adjust the
This work presents an ergonomic and interactive human-robot collaboration (HRC) framework, through which new c
The utilization of shared control in the realm of surgical robotics augments precision and safety by amalgamating hu
In recent years, the need for robots to transition from isolated industrial tasks to shared environments, including hum
"Hey, robot. Let's tidy up the kitchen. By the way, I have back pain today". How can a robotic system devise a share
As robots are deployed in human spaces, it is important that they are able to coordinate their actions with the peopl
In the actual industrial scenarios, human operators and robots work together sharing the workspace. Such proximity
Safety and efficiency are two important aspects of human-robot collaboration (HRC). Most existing control methods
The human-robot collaboration scenarios are characterized by the presence of human operators and robots that wo
Visual language navigation (VLN) is an embodied task demanding a wide range of skills encompassing understand
The synergic orchestration of the cognitive and psychological dimensions characterizes human intelligence. Accord
Many solutions to address the challenge of robot learning have been devised, namely through exploring novel ways
Prior robot painting and drawing work, such as FRIDA, has focused on decreasing the sim-to-real gap and expandi
People with Visual Impairments (PVI) typically recognize objects through haptic perception. Knowing objects and m
This paper presents a framework to navigate visually impaired people through unfamiliar environments by means o
Movement primitives (MPs) are compact representations of robot skills that can be learned from demonstrations an
This paper proposes the utilization of Supernumerary Robotic Limbs (SuperLimbs) for augmenting astronauts durin
Robotic caregivers could potentially improve the quality of life of many who require physical assistance. However, in
Human memory of a robot's competence, and resulting subjective perceptions of that robot, are influenced by nume
Long-term human motion prediction (LHMP) is essential for safely operating autonomous robots and vehicles in pop
For robots to perform assistive tasks in unstructured home environments, they must learn and reason on the seman
In this paper, we present a novel system for Human-Robot Interactive Creation of Artworks (HRICA). Different from
Smart robotic wheelchairs, as well as other assistive robotic devices, can provide an effective form of independent m
In this paper, we tackle the task of image captioning with the complementarity of visible light images and infrared im
Creating visually pleasing stylized ink paintings from 3D models is a challenge in robotic manipulation. We propose
This paper presents a mixed-reality human-robot teaming system. It allows human operators to see in real-time whe
This paper presents a probabilistic Model-based Reinforcement Learning (MBRL) approach for learning the Energy
Legged robots have demonstrated high efficiency and effectiveness in unstructured and dynamic environments. Ho
In this paper, we propose an adaptive controller for virtual passive biped dynamic walking on unknown uneven terra
Transferring human motion skills to humanoid robots remains a significant challenge. In this study, we introduce a W
For humanoid robots, online motion generation on non-flat terrain remains an ongoing research challenge. Comput
In this letter, we present a method for realizing the heat-evoked nociceptive withdrawal reflex (NWR) in the upper lim
This paper presents a novel approach to fall prediction for bipedal robots, specifically targeting the detection of pote
Shape-changing robotic mannequin is a humanoid robot for imitating shapes of human bodies. The diversity of hum
This paper proposes a novel sensor fusion based on Unscented Kalman Filtering for the online estimation of joint-to
Due to the lack of enough real multi-agent data and time-consuming of labeling, existing multi-agent cooperative pe
Recent 3D object detectors typically utilize multi-sensor data and unify multi-modal features in the shared bird’s
In autonomous driving, cooperative perception makes use of multi-view cameras from both vehicles and infrastructu
Classifying intentions of other traffic agents is an essential task for intelligent transportation systems. To simplify thi
Stereo matching under foggy scenes remains a challenging task since the scattering effect degrades the visibility a
Road pavement detection and segmentation are critical for developing autonomous road repair systems. However,
To develop the next generation of intelligent LiDARs, we propose a novel framework of parallel LiDARs and constru
In autonomous driving, 3D occupancy prediction outputs voxel-wise status and semantic labels for more comprehen
Due to the difficulty of acquiring large-scale 3D human keypoint annotation, previous methods for 3D human pose e
Monocular Semantic Occupancy Prediction aims to infer the complete 3D geometry and semantic information of sc
Multi-object tracking (MOT) is a cornerstone capability of any robotic system. Tracking quality is largely dependent
The diverse agents in multi-agent perception systems may be from different companies. Each company might use t
The multi-agent perception system collects visual data from sensors located on various agents and leverages their
Motion forecasting is crucial in enabling autonomous vehicles to anticipate the future trajectories of surrounding age
Cooperative perception can effectively enhance individual perception performance by providing additional viewpoin
Point clouds play a crucial role in robots and intelligent vehicles. Noise caused by fog droplets seriously degrades th
Vehicle-Infrastructure Cooperative (VIC) 3D object detection is a challenging task for balancing communication ban
Current state-of-the-art autonomous driving vehicles mainly rely on each individual sensor system to perform perce
There is an ever-growing zoo of modern neural network models that can efficiently learn end-to-end control from vis
For a general-purpose robot, it is desirable to imitate human demonstration videos that can effectively solve long-ho
Offline reinforcement learning (RL) aims to optimize the policy, based on pre-collected data, to maximize the cumul
We propose DINOBot, a novel imitation learning framework for robot manipulation, which leverages the image-leve
Teaching robots novel skills with demonstrations via human-in-the-loop data collection techniques like kinesthetic te
Preference-based reward learning is a popular technique for teaching robots and autonomous systems how a huma
Deformable object manipulation has been considered a challenging task for robots for its complex dynamics and th
Deep Learning (DL) has been successfully used in robotics for perception tasks and end-to-end robot control. In the
We present a sampling-based approach to reasoning about the caging-based manipulation of rigid and a simplified
This paper presents a novel in-hand rolling manipulation method in which a ball on a cloth attached to fingertips is c
The development of robotic hand that can imitate human movements has always been an important research topic.
Simulation to Real-World Transfer allows affordable and fast training of learning-based robots for manipulation task
Robotics policies are always subjected to complex, second order dynamics that entangle their actions with resulting
We present in-hand manipulation tasks where a robot moves an object in grasp, maintains its external contact mod
Recently, reinforcement learning has led to dexterous manipulation skills of increasing complexity. Nonetheless, lea
Executing contact-rich manipulation tasks necessitates the fusion of tactile and visual feedback. However, the distin
We present a control framework for achieving a robust object grasp and manipulation in hand. In-hand manipulation
This paper presents a large-scale climbing robot that employs a parallel mechanism with three translational degree
We introduce a Cable Grasping-Convolutional Neural Network (CG-CNN) designed to facilitate robust cable graspin
In this paper, a numerically efficient flexible control scheme for the absolute accuracy of industrial robots is presente
Surface treatment tasks such as grinding, sanding or polishing are a vital step of the value chain in many industries
In the context of smart cities, autonomous vehicles, such as unmanned delivery vehicles and taxis are gradually ga
Images captured by robotics under low-light conditions are often plagued by several challenges, including diminishe
Robotic systems for manipulation in millimeter scale often use a camera with high magnification for visual feedback
To enable the computation of effective randomized patrol routes for single- or multi-robot teams, we present RoSSO
Despite the potential of mobile manipulators and applications where robots require a force-controlled physical intera
Approaches for stochastic nonlinear model predictive control (SNMPC) typically make restrictive assumptions abou
A self-driving vehicle must understand its environment to determine the appropriate action. Traditional autonomy sy
Safety verification in motion planning is known to be computationally burdensome, despite its importance in robotics
Fuel efficiency is a crucial aspect of long-distance cargo transportation by oil-powered trucks that economize on cos
This paper presents a robustified time-optimal motion planning approach for navigating an Autonomous Mobile Rob
Model Predictive Control lacks the ability to escape local minima in nonconvex problems. Furthermore, in fastchang
Synthesizing planning and control policies in robotics is a fundamental task, further complicated by factors such as
The success of the model predictive path integral control (MPPI) approach depends on the appropriate selection of
In this work, we introduce LazyBoE, a multi-query method for kinodynamic motion planning with forward propagatio
In this work, a tube-based nearly optimal solution to motion planning in unknown workspaces is presented. The adv
When planning with an inaccurate dynamics model, a practical strategy is to restrict planning to regions of state-act
Efficiently navigating off-road environments presents a number of challenges arising from their unstructured nature.
Robotic systems for manipulation tasks are increasingly expected to be easy to configure for new tasks. While in th
When facing a new motion-planning problem, most motion planners solve it from scratch, e.g., via sampling and ex
The utilization of vector-based deep learning techniques has great prospects in the realm of autonomous driving, pa
The trade-off between computation time and path optimality is a key consideration in motion planning algorithms. W
Manipulation of articulated and deformable objects can be difficult due to their compliant and under-actuated nature
Kinodynamic motion planners allow robots to perform complex manipulation tasks under dynamics constraints or w
Calibration of fixtures in robotic work cells is essential but also time consuming and error-prone, and poor calibratio
There is an emerging industrial demand for predictive maintenance algorithms that exhibit high levels of predictive a
Robotics aims to develop manipulation skills approaching human performance. However, skill complexity is often ov
This study proposes a new semi-analytical design and implementation method for nonlinear partial differential equa
This paper presents an automated sewing system de- signed for smart garment manufacturing, incorporating mach
Surface finishing such as grinding or polishing is a time-consuming task, involves health risks for humans and is stil
Modern manufacturing requires effective integration of production control and maintenance scheduling to improve p
Modern algorithms allow robots to reach a greater level of autonomy and fulfill more challenging tasks. However, on
Lane changing is a fundamental but challenging operation for moving vehicles. Connected and Automated Vehicles
Federated learning has been widely applied in autonomous driving since it enables training a learning model among
Large-scale scenario databases may contain hundreds of thousands of scenarios for the verification and validation
Advances in autonomy offer the potential for dramatic positive outcomes in a number of domains, yet enabling their
This paper presents a novel approach to modeling human driving behavior, designed for use in evaluating autonom
Predicting temporally consistent road users' trajectories in a multi-agent setting is a challenging task due to the unk
The development of connected autonomous vehicles (CAVs) facilitates the enhancement of traffic efficiency in com
Traffic simulation plays a crucial role in evaluating and improving autonomous driving planning systems. After being
Advanced driver assistance systems (ADAS) have been developed to enhance driving safety by issuing timely warn
In safety-critical domains like autonomous driving (AD), errors by the object detector may endanger pedestrians and
Fast and efficient semantic segmentation of large-scale LiDAR point clouds is a fundamental problem in autonomou
Tracking objects in three-dimensional space is critical for autonomous driving. To ensure safety while driving, the tr
In recent years, autonomous driving has garnered significant attention due to its potential for improving road safety
Planning a public transit network is a challenging optimization problem, but essential in order to realize the benefits
This article addresses the robust trajectory tracking problem for a Quadrotor-Slung-Load System (QSLS), which co
The tremendous hype around autonomous driving is eagerly calling for emerging and novel technologies to support
Mobile autonomy relies on the precise perception of dynamic environments. Robustly tracking moving objects in 3D
Traffic congestion is a persistent problem in our society. Existing methods for traffic control have proven futile in alle
The utilization of Large Language Models (LLMs) within the realm of reinforcement learning, particularly as planner
Motion prediction and cost evaluation are vital components in the decision-making system of autonomous ve-hicles
Autonomous vehicles require motion forecasting of their surrounding multi-agents (pedestrians and vehicles) to ma
As machine learning models become increasingly prevalent in motion forecasting for autonomous vehicles (AVs), it
Limited visibility and sensor occlusions pose pressing safety challenges for advanced driver-assistance systems (A
Conventional control, such as model-based control, is commonly utilized in autonomous driving due to its efficiency
In this paper, we focus on the autonomous multiagent taxi routing problem for a large urban environment where the
The safety of autonomous vehicles (AV) has been a long-standing top concern, stemming from the absence of rare
Environments for autonomous driving can vary from place to place, leading to challenges in designing a learning m
Predicting the future behavior of agents is a fundamental task in autonomous vehicle domains. Accurate prediction
Despite large advances in recent years, real-time capable motion planning for autonomous road vehicles re- mains
In autonomous driving, predicting the behavior (turning left, stopping, etc.) of target vehicles is crucial for the self-dr
The driver warning system that alerts the human driver about potential risks during driving is a key feature of an adv
Merging into dense highway traffic for an autonomous vehicle is a complex decision-making task, wherein the vehic
Reinforcement Learning is a highly active re- search field with promising advancements. In the field of autonomous
Lane change in dense traffic typically requires the recognition of an appropriate opportunity for maneuvers, which re
In the burgeoning field of autonomous vehicles (AVs), trajectory prediction remains a formidable challenge, especia
Real-time, accurate prediction of human steering behaviors has wide applications, from developing intelligent traffic
Exoskeleton-assisted home-based rehabilitation plays a vital role in upper limb rehabilitation of stroke patients in the
We address the problem of (a) predicting the trajectory of an arm reaching motion, based on a few seconds of the m
 This paper explores the feasibility of employing EEG-based intention detection for real-time robot assistive control.
Microexpressions are expressions that people inadvertently express, and therefore often represent a person's true
Anticipating driver intention is an important task when vehicles of mixed and varying levels of human/machine auton
This work explores the capacity of large language models (LLMs) to address problems at the intersection of spatial
n Recognition
In this paper, we present a new paradigm for evaluating egocentric pedestrian trajectory prediction algorithms. Base
In human-robot collaboration (HRC) settings, hand motion intention prediction (HMIP) plays a pivotal role in ensurin
The ability to grasp objects broadens the application range of unmanned aerial vehicles (UAVs) by allowing interact
The human shoulder, with its glenohumeral joint, tendons, ligaments, and muscles, allows for the execution of comp
This study presents the design and validation of a variable stiffness actuator incorporating multiple cam mechanism
There is a lack of cleaning robots dedicated to the scrubbing of contaminated surfaces. Contaminated surfaces in d
Cable-driven continuum manipulators have gained considerable attention due to their high dexterity and inherent st
The application requirements of the tendon-sheath mechanism in the field of precision machinery are becoming inc
Tendon driven continuum robots are often considered to navigate through- and operate in cluttered environments. W
Some cable-driven parallel robots (CDPRs) can be rapidly deployed on-site. To achieve such deployability, the fixed
Bowden cables serve as essential components in various mechanical systems, facilitating power transmission from
In robotics and biomechanics, accurately determining joint parameters and computing the corresponding forward an
Water jumping motion extends the robot's movement space and flexibility. However, the jumping performance is inf
In this paper, we present an iterative method to quickly traverse multi-dimensional paths considering jerk constraint
This letter presents an investigation into the mathematical relationships between the positions and orientations at th
We present an analytic solution to the 3D Dubins path problem for paths composed of an initial circular arc, a straig
Performance measures are essential to characterize a robot's ability to carry out manipulation tasks. Generally, the
Robotic arms are highly common in various automation processes such as manufacturing lines. However, these hig
Trajectory segmentation refers to dividing a trajectory into meaningful consecutive sub-trajectories. This paper focu
A new, and extremely efficient, computational modeling paradigm is introduced here for specific finite elasticity prob
In this paper, we propose a metric called hitting flux which is used in the motion generation and controls for a robot
Automated storage and retrieval systems (ASRS) are a key component of the modern storage industry, and are use
The 4-DOF (3T1R) 4PPa-2PaR parallel manipulator is developed for high-speed pick-and-place operations. Howev
Sagging cable-driven parallel robots (CDPRs) are often modelled by using the Irvine's model. We will show that the
Cable elasticity can significantly impact the accuracy of Cable-Driven Parallel Robots (CDPRs). However, it’s fr
The wrench and twist feasibility are the workspace conditions that indicate whether the mobile-platform (MP) of the
This paper presents the design, analysis, and performance evaluation of RicMonk, a novel three-link brachiation rob
External and internal convertible (EIC) form-based motion control (i.e., EIC-based control) is one of the effective ap
Safety certificates based on energy functions can provide demonstrable safety for complex robotic systems. Howev
Safety has been a critical issue for the deployment of learning-based approaches in real-world applications. To add
Robots are more capable of achieving manipulation tasks for everyday activities than before. But the safety of mani
A robot in a human-centric environment needs to account for the human’s intent and future motion in its task an
We developed a modality attention motion generation model on the basis of multi-modality prediction. This model p
Various robots have been developed so far; however, we face challenges in modeling the low-rigidity bodies of som
Robots acting in human environments will often face new situations and can benefit from transferring prior experien
Pre-training a diverse set of neural network controllers in simulation has enabled robots to adapt online to damage
In this study, we present an implementation strategy for a robot that performs peg transfer tasks in Fundamentals o
Learning from humans allows non-experts to program robots with ease, lowering the resources required to build co
Movement Primitive (MP) is a promising Learning from Demonstration (LfD) framework, which is commonly used to
The proposed work presents a framework based on Graph Neural Networks (GNN) that abstracts the task to be ex
Humans perform complex tasks involving force interactions daily. Learning from demonstration, a method for transf
Skill propagation among robots without human involvement can be crucial in quickly spreading new physical skills to
In many contact-rich tasks, force sensing plays an essential role in adapting the motion to the physical properties of
Building a general and efficient path planning framework in uncertain nonconvex environments is challenging due to
Large-scale self-supervised models have recently revolutionized our ability to perform a variety of tasks within the v
As robots tackle complex object arrangement tasks, it becomes imperative for them to be able to generalize to com
Direct teaching is a straightforward way of teaching new motion to robots. Active methods with torque sensors, for e
We propose WayEx, a new method for learning complex goal-conditioned robotics tasks from a single demonstratio
In surgery, the application of appropriate force levels is critical for the success and safety of a given procedure. Wh
Due to their flexibility and ease of use, Dynamical Movement Primitives (DMPs) are widely used in robotics applicat
While humans can use parts of their arms other than the hands for manipulations like gathering and supporting, wh
It is crucial for robots to be aware of the presence of constraints in order to acquire safe policies. However, explicitly
Learning from Demonstration (LfD) enables robots to acquire new skills by imitating expert demonstrations, allowing
Relative placement tasks are an important category of tasks in which one object needs to be placed in a desired po
Imitation learning mitigates the resource-intensive nature of learning policies from scratch by mimicking expert beha
Established techniques that enable robots to learn from demonstrations are based on learning a stable dynamical s
Deploying machine learning algorithms for robot tasks in real-world applications presents a core challenge: overcom
We propose a Dynamical System (DS) approach to learn complex, possibly periodic motion plans from kinesthetic d
Autonomous Dynamic System (DS)-based algorithms hold a pivotal and foundational role in the field of Learning fro
Learning a stable autonomous dynamic system (ADS) encoding human motion rules has been shown as an effectiv
We present the problem of inverse constraint learning (ICL), which recovers constraints from demonstrations to aut
Current learning from demonstration (LfD) frameworks have difficulty dealing with an unreliable, limited number of d
Model-based reinforcement learning (RL) has shown great promise due to its sample efficiency, but still struggles w
Self-supervised skill learning aims to acquire useful behaviors that leverage the underlying dynamics of the environ
This paper addresses the problem of robotic cutting during disassembly of products for materials separation and re
Pre-training robots with a rich set of skills can substantially accelerate the learning of downstream tasks. Prior work
In reinforcement learning (RL), learning directly from pixels, is commonly known as vision-based RL. Effective state
Learning highly dynamic behaviors for robots has been a longstanding challenge. Traditional approaches have dem
Model-based RL is a promising approach for real-world robotics due to its improved sample efficiency and generaliz
Learning strategic robot behavior---like that required in pursuit-evasion interactions---under real-world constraints is
Offline reinforcement learning (RL) has been shown to be successfully modeled as a sequence modeling problem,
Snake robots have showcased remarkable compliance and adaptability in their interaction with environments as the
Reinforcement Learning (RL) has progressed from simple control tasks to complex real-world challenges with large
Multi-agent reinforcement learning has emerged as a promising approach for the control of multi-robot systems. Ne
Models with fewer parameters are necessary for the neural control of memory-limited, performant robots. Finding th
Deep reinforcement learning can enable robots to autonomously acquire complex behaviors such as legged locomo
In this paper, we review the question of which action space is best suited for controlling a real biped robot in combin
Reinforcement learning (RL) has achieved remarkable success in complex robotic systems (eg. quadruped locomo
We present TartanDrive 2.0, a large-scale off-road driving dataset for self-supervised learning tasks. In 2021 we rel
In recent years, significant progress has been made in the field of underwater image enhancement (UIE). However,
To avoid unmanned ground vehicles being obstructed by deformed terrain in off-road, effective vehicle mobility ana
Traversability estimation in rugged, unstructured environments remains a challenging problem in field robotics. Ofte
This letter presents a novel method for geographical localization by registering satellite maps with LiDAR point clou
Efficient and reliable navigation in off-road environments poses a significant challenge for robotics, especially when
In wild environments, motion planning for mobile robots faces the challenge of local optimal path traps due to limite
Data-driven methods have great advantages in modeling complicated human behavioral dynamics and dealing with
In this work, we investigate the potential of improving multi-task training and leveraging it for transferring in the reinf
Goal-Conditioned Reinforcement Learning (GCRL) can enable agents to spontaneously set diverse goals to learn a
Object-centric representation is an essential abstraction for forward prediction. Most existing forward models learn t
Current reinforcement learning algorithms struggle in sparse and complex environments, most notably in long-horiz
Tactile information holds promise for enhancing the manipulation capabilities of multi-fingered robots. In tasks such
Multi-task learning of deformable object manipulation is a challenging problem in robot manipulation. Most previous
Interactive grasping from clutter, akin to human dexterity, is one of the longest-standing problems in robot learning.
We propose a new policy class, Composable Interaction Primitives (CIPs), specialized for learning sustained-contac
Tracking and catching moving objects is an important ability for robots in a dynamic world. Whilst some objects hav
In contact-rich tasks, the hybrid, multi-modal nature of contact dynamics poses great challenges in model represent
Multi-task learning has expanded the boundaries of robotic manipulation, enabling the execution of increasingly com
Furniture assembly remains an unsolved problem in robotic manipulation due to its long task horizon and nongener
Using tactile sensors for manipulation remains one of the most challenging problems in robotics. At the heart of the
Synchronized dual-arm rearrangement is widely studied as a common scenario in industrial applications. It often fac
If a robot masters folding a kitchen towel, we would also expect it to master folding a beach towel. However, existin
Deep Reinforcement Learning (DRL) has become a popular approach for training robots due to its generalization pr
Robotic manipulation of deformable linear objects (DLOs) has great potential for applications in diverse fields such
We present BioSLAM, a lifelong SLAM framework for learning various new appearances incrementally and maintain
Solving reinforcement learning-based mapless navigation tasks is challenging due to their sparse reward and long d
Despite many successful applications of data-driven control in robotics, extracting meaningful diverse behaviors rem
Along with the advancement of robot skin technology, there has been notable progress in the development of snake
Localization is paramount for autoatnomous robots. While camera and LiDAR-based approaches have been exten-
NeRF can reconstruct incredibly realistic environmental maps in dense simultaneous localization and mapping, prov
The growing focus on indoor robot navigation utilizing wireless signals has stemmed from the capability of these sig
Navigation methods based on deep reinforcement learning (RL) have recently exhibited superior performance, part
2D top-down maps are commonly used for the navigation and exploration of mobile robots through unknown areas.
Deep reinforcement learning (RL) has shown promising results in robot motion planning with first attempts in human
Deploying a safe mobile robot policy in scenarios with human pedestrians is challenging due to their unpredictable m
Reinforcement Learning (RL) has been widely used to create generalizable autonomous vehicles. However, they re
While reinforcement learning (RL) attracts increasing research attention, maximizing the return while keeping the ag
Symmetry is a fundamental aspect of many real-world robotic tasks. However, current deep reinforcement learning
This work focuses on the dual-arm object rearrangement problem abstracted from a realistic industrial scenario of C
Safe Reinforcement Learning (RL) aims to find a policy that achieves high rewards while satisfying cost constraints
Motivated by the challenge of achieving rapid learning in physical environments, this paper presents the developme
This paper investigates the multi-agent cooperative exploration problem, which requires multiple agents to explore a
Mobile ground robots require perceiving and understanding their surrounding support surface to move around auton
For locomotion, is an arm on a legged robot a liability or an asset for locomotion? Biological systems evolved additi
Controlled execution of dynamic motions in quadrupedal robots, especially those with articulated soft bodies, prese
In this paper, we propose a method to automatically and efficiently tune high-dimensional vectors of controller param
We propose the expert composer policy, a framework to reliably expand the skill repertoire of quadruped agents. Th
Can a quadrupedal robot perform bipedal motions like humans? Although developing human-like behaviors is more
Abstract— We aim to control a robot to physically behave in the real world following any high-level language comm
Quadrupedal robots have played a crucial role in various environments, from structured environments to complex h
Electric quadruped robots used in outdoor exploration are susceptible to leg-related electrical or mechanical failures
Real-world autonomous missions often require rich interaction with nearby objects, such as doors or switches, alon
Control policies trained using deep reinforcement learning often generate stiff, high-frequency motions in response
Control of legged robots is a challenging problem that has been investigated by different approaches, such as mode
In the context of legged robots, adaptive behavior involves adaptive balancing and adaptive swing foot reflection. W
Humans can perform parkour by traversing obstacles in a highly dynamic fashion requiring precise eye-muscle coo
Deployment in hazardous environments requires robots to understand the risks associated with their actions and m
The robustness of legged locomotion is crucial for quadrupedal robots in challenging terrains. Recently, Rein- force
Reinforcement Learning (RL) has been a powerful tool for training robots to acquire agile locomotion skills. To learn
Recent advances of locomotion controllers utilizing deep reinforcement learning (RL) have yielded impressive resul
This study proposes a novel planning framework based on a model predictive control formulation that incorporates
This paper presents a reactive locomotion method for bipedal robots enhancing robustness and external disturbanc
Successfully achieving bipedal locomotion remains challenging due to real-world factors such as model uncertainty
For humans, fast, efficient walking over flat ground represents the vast majority of locomotion that an individual exp
Running and jumping are locomotion modes that allow legged robots to rapidly traverse great distances and overco
This paper presents a novel framework for learning robust bipedal walking by combining a data-driven state represe
High-frequency and accurate state estimation is crucial for biped robots. This paper presents a tightly-coupled LiDA
This paper introduces Barry, a dynamically balancing quadruped robot optimized for high payload capabilities and e
We present a novel controller design on a robotic locomotor that combines an aerial vehicle with a spring-loaded le
We present a minimal phase oscillator model for learning quadrupedal locomotion. Each of the four oscillators is co

Balancing oneself using the spine is a physiological alignment of the body posture in the most efficient manner by th
We present SLoMo: a first-of-its-kind framework for transferring skilled motions from casually captured “in- the-w
The capability of a quadruped robot to negotiate obstacles is tightly connected to its leg workspace and joint torque
We present SpaceHopper, a three-legged, small-scale robot designed for future mobile exploration of asteroids and
Learning a locomotion policy for quadruped robots has traditionally been constrained to a specific robot morphology
Quadruped robots have gained attention for their potential to navigate various terrains. However, the stability of the
Legged robots have the potential to become vital in maintenance, home support, and exploration scenarios. In orde
Learning-based methods have improved locomotion skills of quadruped robots through deep reinforcement learning
Model Predictive Control (MPC) relies heavily on the robot model for its control law. However, a gap always exists b
 This paper presents a layered control approach for real-time trajectory planning and control of robust cooperative lo
This letter presents a control framework that combines model-based optimal control and reinforcement learning (RL
This paper addresses the problem of realizing a novel robotic bipedal locomotion called wheel gait, which is achieve
We propose a limb trajectory planning method that considers both body and limb dynamics in robots, particularly su
Quadrupeds deployed in real-world scenarios need to be robust to unmodelled dynamic effects. In this work, we aim

Humanoid robots possess the ability to perform complex tasks in challenging environments. However, they require
In order to perform highly dynamic and agile maneuvers, legged robots typically spend time in underactuated doma
Robot skateboarding is an unexplored and challenging task for legged robots. Accurately modeling the dynamics of
This study focuses on a layered, experience-based, multi-modal contact planning framework for agile quadrupedal
This paper proposes a learning-based adaptive propulsion control (APC) method for a quadruped robot integrated w
In recent years, legged and wheeled-legged robots have gained prominence for tasks in environments predominan
In the natural world, benefited from the advantages of the spine, quadrupeds exhibiting extraordinary flexibility whic
The recent development of online static map element (a.k.a. HD Map) construction algorithms has raised a vast dem
This paper presents a robotics perception research dataset collected in Rome, featuring RGB data, 3D point clouds
Self-supervised monocular depth estimation has attracted extensive attention in recent years. Lightweight depth es
This paper adapts a general dataset representation technique to produce robust Visual Place Recognition (VPR) de
ation and Mapping I
Visual place recognition is a critical task in computer vision, especially for localization and navigation systems. Exis
In the field of robotics engineering and autonomous driving vehicles, precise estimation of positions through visual p
Visual place recognition(VPR) is a crucial technology for autonomous driving and robotic navigation. However, seve
Abstract— HD map construction is a fundamental and challenging task in autonomous driving to understand the su
As edge devices equipped with cameras and inertial measurement units (IMUs) are emerging, it holds huge implica
Visual Odometry (VO) plays a pivotal role in autonomous systems, with a principal challenge being the lack of depth
LiDAR Mapping has been a long-standing problem in robotics. Recent progress in neural implicit representation has
Self-supervised depth estimation algorithms rely heavily on frame-warping relationships, exhibiting substantial perfo
We present a neural-field-based large-scale reconstruction system that fuses lidar and vision data to generate high
Precise and long-term stable localization is essential in parking lots for tasks like autonomous driving or autonomou
In the process of urban environment mapping, the sequential accumulations of dynamic objects will leave a large nu
Accurate perception of objects in the environment is important for improving the scene understanding capability of S
A 3D occupancy map that is accurately modeled after real-world environments is essential for reliably performing ro
The detection of traversable regions on staircases and the physical modeling constitutes pivotal aspects of the mob
Reliable estimation of terrain traversability is critical for the successful deployment of autonomous systems in wild, o
In this study, we address the off-road traversability estimation problem, that predicts areas where a robot can navig
Human beings cooperatively navigate rule-constrained environments by adhering to mutually known navigational pa
Visual place recognition (VPR) plays a pivotal role in autonomous exploration and navigation of mobile robots within
This paper presents an accurate and fast 3D global localization method, 3D-BBS, that extends the existing branch-
Dynamic objects diversify the distribution of point cloud in the map, degrading the performance of the robotic downs
Visual-audio navigation (VAN) is attracting more and more attention from the robotic community due to its broad ap
We tackle the problem of learning an implicit scene representation for 3D instance segmentation from a sequence o
In urban environments for delivery robots, particularly in areas such as campuses and towns, many custom feature
Accurate localization is an essential technology for the flexible navigation of robots in large-scale environments. Bot
Robot localization using subsurface features captured by Ground-Penetrating Radar (GPR) complements and impr
State-of-the-art techniques for monocular camera reconstruction predominantly rely on the Structure from Motion (S
We present COIN-LIO, a LiDAR Inertial Odometry pipeline that tightly couples information from LiDAR intensity with
This paper presents a 6-DoF range-based Monte Carlo localization method with a GPU-accelerated Stein particle fi
This paper presents a range inertial localization algorithm for a 3D prior map. The proposed algorithm tightly couple
Recognizing places from an opposing viewpoint during a return trip is a common experience for human drivers. How
Image matching is still challenging in such scenes with large viewpoints or illumination changes or with low textures
State-of-the-art V-SLAM pipelines utilize robust cost functions and outlier rejection techniques to remove incorrect c
One of the key challenges of visual Simultaneous Localization and Mapping (SLAM) in large-scale environments is
The inherent scale ambiguity issue greatly limits the performance of monocular visual odometry. In recent years, a
Perceptual aliasing and weak textures pose significant challenges to the task of place recognition, hindering the per
LiDAR-based localization is valuable for applications like mining surveys and underground facility maintenance. How
Visual Place Recognition (VPR) is critical for navigation and loop closure in autonomous driving tasks, mitigating the
Most of the existing LiDAR-inertial navigation systems are based on frame-to-map registrations, leading to inconsis
We study the problem of aligning a video that captures a local portion of an environment to the 2D LiDAR scan of th
This paper tackles a localization problem in large-scale indoor environments with wayfinding maps. A wayfinding m
Visual relocalization aims to estimate the absolute camera pose from an image or sequential images. Recent works
Visual localization is an important sub-task in SfM and visual SLAM that involves estimating a 6-DoF camera pose f
As a promising fashion for visual localization, scene coordinate regression (SCR) has seen tremendous progress in
Visual Place Recognition aims at recognizing previously visited places by relying on visual clues, and it is used in ro
Place recognition is an essential and challenging task in loop closing and global localization for robotics and autono
State estimation for legged robots is challenging due to their highly dynamic motion and limitations imposed by sen
This paper proposes a pose-graph attentional graph neural network, called P-GAT, which compares (key)nodes be
We propose a topological mapping and localization system able to operate on real human colonoscopies, despite s
Remote magnetic navigation provides a promising approach for improving the maneuverability and safety of surgica
Place recognition is a challenging but crucial task in robotics. Current description-based methods may be limited by
Localization using a monocular camera in the pre-built LiDAR point cloud map has drawn increasing attention in the
Mapping an environment is essential for several robotic tasks, particularly for localization. In this paper, we address
Localizing ground penetrating radar (LGPR) has been proven to be a promising technology for robot localization in
Vision-based localization for autonomous driving has been of great interest among researchers. When a pre-built 3
Camera relocalization is pivotal in computer vision, with applications in AR, drones, robotics, and autonomous drivin
Robust and accurate pose estimation in unknown environments is an essential part of robotic applications. We focu
Absolute Pose Regressors (APRs) directly estimate camera poses from monocular images, but their accuracy is un
Global visual localization estimates the absolute pose of a camera using a single image, in a previously mapped are
Indoor parking lots are the GPS-denied spaces to which vision-based localization approaches have usually been ap
The availability of city-scale Lidar maps enables the potential of city-scale place recognition using mobile cameras.
Spiking Neural Networks (SNNs) are at the forefront of neuromorphic computing thanks to their potential energy-eff
17-point algorithm is a popular method in relative pose estimation of multi-cameras. However, the role of overlap in
Visual Place Recognition (VPR) is vital for robot localization. To date, the most performant VPR approaches are en
We present a lightweight ground texture based localization algorithm (L-GROUT) that improves the state of the art i
Achieving efficient and consistent localization with a prior map remains challenging in robotics. Conventional keyfra
Vision-aided localization for low-cost mobile robots in diverse environments has attracted widespread attention rece
Accurate global localization is crucial for autonomous navigation and planning. To this end, various GPS-aided Visu
Nano-sized unmanned aerial vehicles (UAVs) are well-fit for indoor applications and for close proximity to humans.
We present a synchronized multisensory dataset for accurate and robust indoor localization: the Lund University Vi
In this paper, we address the problem of relative localization of two mobile agents. Specifically, we consider the Du
Unmanned vehicles usually rely on Global Positioning System (GPS) and Light Detection and Ranging (LiDAR) sen
Aquatic robots require an accurate and reliable localization system to navigate autonomously and perform practical
We propose a fixed-lag smoother-based sensor fusion architecture to leverage the complementary benefits of rang
The majority of commercial and open-source autopilot software for uncrewed aerial vehicles rely on the tried and te
Indoor wireless ranging localization is a promising approach for low-power and high-accuracy localization of wearab
Indoor Positioning Systems (IPS) traditionally rely on odometry and building infrastructures like WiFi, often supplem
This paper describes a multi-modal data association method for global localization using object-based maps and ca
This paper presents an efficient method for updating particles in a particle filter (PF) to address the position estimat
Lifelong indoor localization in a given map is the basis for navigation of autonomous mobile robots. In this letter, we
The rapid and precise localization and prediction of a ball are critical for developing agile robots in ball sports, partic
We present a novel 3D odometry method that recovers the full motion of a vehicle only from a Doppler-capable ran
There is a current increase in the development of “4D” Doppler-capable radar and lidar range sensors that p
One-shot LiDAR localization refers to the ability to estimate the robot pose from one single point cloud, which yields
In this paper, we introduce a novel and efficient technique for consistent covariance recovery in nonlinear optimizat
This paper proposes an adaptive method for robot-to-human handovers under different scenarios. The method com
Reinforcement learning has shown great potential in improving the performance of robot navigation. In response to
This paper presents an adaptive online learning framework for systems with uncertain parameters to ensure safety-
Manipulating deformable objects is a ubiquitous task in household environments, demanding adequate representat
An open research question in robotics is how to combine the benefits of model-free reinforcement learning (RL) - kn
Significant progress has been made in training multimodal trajectory forecasting models for autonomous driving. Ho
Due to the inherent uncertainty in their deformability during motion, previous methods in deformable object manipul
In this work we propose a learning-based approach to box loco-manipulation for a humanoid robot. This is a particu
Reliable autonomous navigation requires adapting the control policy of a mobile robot in response to dynamics cha
A major challenge in robotics is to design robust policies which enable complex and agile behaviors in the real world
Autonomous off-road driving is challenging as unsafe actions may lead to catastrophic damage. As such, developin
Recent years have seen the development of many methods for robotic reinforcement learning (RL), some of which
We propose a method for improving the prediction accuracy of learned robot dynamics models on out-of-distribution
Pre-training on Internet data has proven to be a key ingredient for broad generalization in many modern ML system
Future planetary exploration missions will require reaching challenging regions such as craters and steep slopes. S
Safety is one of the main challenges in applying reinforcement learning to realistic environmental tasks. To ensure s
In practice, many types of manipulation actions (e.g., pick-n-place and push) are needed to accomplish real-world m
The ability to manipulate objects in desired configurations is a fundamental requirement for robots to complete vario
We present ForceSight, a system for text-guided mobile manipulation that predicts visual-force goals using a text-co
The potential of tactile sensing for dexterous robotic manipulation has been demonstrated by its ability to enable nu
Robotic manipulation of deformable, one-dimensional objects (DOOs) like ropes or cables has important potential a
3D articulated objects are inherently challenging for manipulation due to the varied geometries and intricate function
We are interested in enabling robots to solve difficult pick-and-place manipulation tasks in cluttered and constrained
We are interested in studying sports with robots and starting with the problem of intercepting a projectile moving tow
Autonomous driving has traditionally relied heavily on costly and labor-intensive High Definition (HD) maps, hinderin
Neural implicit surface representations are currently receiving a lot of interest as a means to achieve high-fidelity su
Large-scale dense mapping is vital in robotics, digital twins, and virtual reality. Recently, implicit neural mapping has
Camera relocalization is a crucial problem in computer vision and robotics. Recent advancements in neural radianc
This paper presents a novel framework to learn a concise geometric primitive representation for 3D point clouds. Di
Mapping plays a crucial role in location and navigation within automatic systems. However, the presence of dynami
Constructing a high-quality dense map in real-time is essential for robotics, AR/VR, and digital twins applications. A
3D object-level mapping is a fundamental problem in robotics, which is especially challenging when object CAD mo
Mapping is crucial for spatial reasoning, planning and robot navigation. Existing approaches range from metric, whi
Humans are remarkable in their ability to navigate without metric information. We can read abstract 2D maps, such
Particle-based dynamic occupancy maps were proposed in recent years to model the obstacles in dynamic environ
Mobile robots that navigate in unknown environments need to be constantly aware of the dynamic objects in their s
Light Detection and Ranging (LiDAR) technology has proven to be an important part of many robotics systems. Sur
A Colored point cloud, as a simple and efficient 3D representation, has many advantages in various fields, including
Mapping with uncertainty representation is required in many research domains, especially for localization. Although
This article presents an efficient occupancy mapping framework for high-resolution LiDAR sensors, termed D-Map.
Mobile robots navigating in outdoor environments frequently encounter the issue of undesired traces left by dynami
The joint optimization of the sensor trajectory and 3D map is a crucial characteristic of Simultaneous Localization an
Underwater robots are indispensable for aquatic exploration, yet their size and complexity often limit broader applic
In nature, fish are excellent swimmers due to their flexible and precise control of tail, which allows them to freely tra
The extreme conditions of the deep sea require the use of large and expensive diving robots designed to withstand
Underwater vehicle manipulator systems (UVMS) are increasingly popular platforms for performing subsea operatio
This paper introduces a small hybrid aerial underwater vehicle (HAUV), which we named Nezha-F, that can fly in th
Fish exhibit impressive locomotive performance and agility in complex underwater environments, using their undula
Stability, adaptability and maneuverability are the most important performance indexes for the underwater biomime
We present a novel approach to marine sediment sampling, which makes use of a hexapedal robotic platform, nam
Underwater hexapod robot, driven by six C-shaped legs and eight thrusters, has the potential to traverse diverse te
Robust underwater motion planning of autonomous underwater vehicles (AUVs) in dynamic cluttered environments
In this work, we address the problem of real-time dense depth estimation from monocular images for mobile underw
Object pose estimation underwater allows an autonomous system to perform tracking and intervention tasks. None
In this paper, we address the challenging problem of data association for underwater SLAM through a novel method
Deep learning has brought transformative advancements to object segmentation, especially in marine robotics cont
Underwater imagery often exhibits distorted coloration as a result of light-water interactions, which complicates the
In this paper, we present CaveSeg - the first visual learning pipeline for semantic segmentation and scene parsing f
We present a novel system which blends multiple distinct sensing modalities in audio-visual surveys to assist marin
Point cloud compression is an essential technology for the efficient storage and transmission of 3D data. Previous m
This paper examines the relevance of using catenary-based curves to model cables in underwater tethered robotic
This paper reports a numerical method for modeling underwater vehicle (UV) interactions with the free surface usin
 The agile and safe trans-domain in waves is a promising feature but the primary bottleneck of the hybrid aerial unde
There exists a capability gap in the design of currently available autonomous underwater vehicles (AUV). Most AUV
In autonomous robot navigation, the trajectories from path planners are considered to be safe regions, and deviatio
This paper describes a novel multimode motion robot named SHOALBOT,which with multimode operations depend
The optical effects that are observed in underwater imagery are more complex than those in-air. This is partially bec
Deep learning has shown promising results for multiple 3D point cloud registration datasets. However, in the underw
Developing autonomous intervention capabilities for lightweight underwater vehicle manipulator systems (UVMS) ha
This paper reports a new hierarchical architecture for modeling autonomous multi-robot systems (MRSs): a non-line
Robotics IV
Persistent operation of Autonomous Underwater Vehicles (AUVs) without manual interruption for recharging saves
Collision avoidance algorithms for Autonomous Surface Vehicles (ASV) that follow the Convention on the Internatio
Reliable localization is an essential capability for marine robots navigating in GPS-denied environments. SLAM, com
Autonomous Marine Vehicles (AMVs) have been widely used in many critical tasks such as surveillance, patrolling,
Autonomous underwater vehicles often perform surveys that capture multiple views of targets in order to provide m
In recent years, underwater robots have been increasingly used in the maintenance of hydraulic structures. Underw
An open-source software framework is presented that allows real-time underwater mapping with popular marine rob
Autonomous underwater docking is of the utmost importance for expanding the capabilities of Autonomous Underw
Determining the distances from the hull of the own ship to obstacles or land, i.e. water clearance, is a fundamental
Autonomous Underwater Vehicles (AUVs) conduct missions underwater without the need for human intervention. A
This paper presents an integrated navigation and control strategy for an autonomous surface vehicle (ASV) to oper
This work presents a camera model for refractive media such as water and its application in underwater visual-inert
This paper presents an extension to visual inertial odometry (VIO) by introducing tightly-coupled fusion of magnetom
In this work, we present an acoustic-visual-inertial navigation system (Acoustic-VINS) for underwater robot localizat
Lack of information and perceptual ambiguity are key problems in sonar-based mapping applications. We propose a
In this paper, we present a novel trajectory planning method for externally-actuated modular manipulators (EAMMs
This paper presents an aerial robotic platform for rapid remote elevated overhead-perching drill operations for wood
Modeling the kinematics and dynamics of robotics systems with suspended loads using dual quaternions has not be
Existing multirotor-based cargo transportation does not maintain a constant cargo attitude due to underactuation; ho
While the field of autonomous Uncrewed Aerial Vehicles (UAVs) has grown rapidly, most applications only focus on
Flapping wing air vehicles(FWAV) or ornithopters are bio-inspired aerial robots that mimic the flying principles of ins
Aerial unperching of multirotors has received little attention as opposed to perching that has been investigated to el
Low-cost autonomous Micro Aerial Vehicles (MAVs) have great potential to help humans by simplifying and speedin
This paper details the simulation and experimental validation of an autonomous perching and take-off method for a
The ability to fast climb and descend is crucial for Unmanned Aerial Vehicle (UAV) applications in the mountains. Th
This work studies how parametric uncertainties affect the cooperative manipulation of a cable-suspended beam-sha
Aerial robots show significant potential for forest canopy research and environmental monitoring by providing data c
Achieving accurate control performance of the end-effector is critical for practical applications of aerial manipulator.
This work focuses on understanding and identifying the drag forces applied to a rotary-wing Micro Aerial Vehicle (M
This work validates through flight tests a previously developed wide-envelope singularity-free aerodynamic framewo
Electroaerodynamic (EAD) propulsion, where thrust is produced by collisions between electrostatically-accelerated
In this letter, a variation of the classical Markov- Dubins problem is considered, which deals with curvature constrain
 A risk aware UAS path planning methodology is proposed using monetary value as the sole cost metric. A third par
Perching on the surface of moving objects, like vehicles, could extend the flight time and range of quadrotors. Sucti
In this paper, the dynamics of an emerging class of rotating nature-inspired micro aerial vehicles known as the Mon
Recently, the utilization of aerial manipulators for performing pushing tasks in non-destructive testing (NDT) applica
In this paper, we introduce PADUAV, a novel 5-DOF aerial platform designed to overcome the limitations of traditio
This paper presents a novel method for state estimation of rigid body attitude system evolving on the manifold, S3,
Aggressive time-optimal control of quadcopters poses a significant challenge in the field of robotics. The state-of-th
This letter reports an aerial robot--Quadrolltor, with the ability to roll and turn. Existing bimodal quadrotors feature cy
The design and control of winged aircraft and drones is an iterative process aimed at identifying a compromise of m
Quadrotors may crash and cause severe accidents under instantaneously drastic disturbances. To mitigate the effe
We present model predictive selection (MPS), a new method for selecting the stable closed-loop (CL) equilibrium a
ics and Control IV
Pushing tasks performed by aerial manipulators can be used for contact-based industrial inspections. Underactuate
Building and maintaining structures like antennas and cable-car masts in challenging terrain often involves hazardo
Aerial manipulation has received increasing research interest with wide applications of drones. To perform specific
Swarms of aerial drones have recently been considered for last-mile deliveries in urban logistics or automated cons
The standard quadrotor is one of the most popular and widely used aerial vehicle of recent decades, offering great
In this work, we present a novel actuation strategy for a suspended aerial platform. By utilizing an underactuation a
The problem of autonomous navigation for UAV inspection remains challenging as it requires effectively navigating
This letter presents a design framework and novel control strategy for a compact coaxial magnetic-gear-based actu
In many robotic systems, the holding state con- sumes power, limits operating time, and increases operating costs.
This paper presents a bionic foldable wing that imitates the hind wing of ladybirds. Based on the folding mechanism
Widespread applications for mobile robots are creating a large demand for developing new driving mechanisms tha
Variable Stiffness Mechanisms (VSM) are becoming ubiquitous in mechatronics given the benefit they provide in te
Plant cells expand and elongate. Their cumulative actuation defines organ morphing. Inspired by this modular trans
This paper presents the design and experimen- tal results of a proprioceptive, high-bandwidth quasi-direct drive (QD
With the development of society, aging population and the number of stroke patients is increasing year by year. Re
A support structure for flexible displays such as OLED or flexible LEDs was developed using the flexible omnidirect
This article presents a gripper comprising finger and suction with reconfigurable attributes. With the reconfiguration
We propose the method to adapt humanoids the ability to change the body structures that modular robots have by
Inspired by Yoshimura origami, this study presents a novel deployable modular continuum robot that achieves conf
Conventional mobile robots typically include a single locomotion mode and require additional arms to transport obje
A human foot with high degrees of freedom (DOF) that has multi-DOF toe joints and a two-DOF ankle joint provides
For researchers or administrators of relevant institutions who need to collect hydrological data of a certain water are
In this paper, we have developed a protection module for Light Detection and Ranging (LiDAR) sensors used in out
We present theoretical and experimental results regarding the development and the control of a two-link robotic arm
This paper introduces a novel origami-inspired shape-changing robot OSCaR. The objective is to enhance the adap
This paper proposes a novel flexible vibrational actuator with a structural anisotropy and its control method to divers
Soft growing robots with unique navigation (tip extension by eversion) hold great promise in rescue, medical, and in
Robot designs can take many inspirations from nature, where there are many examples of highly resilient and fault-
In recent years, bio-syncretic robots actuated by living materials have received widespread attention. Among the co
Sit-to-Stand (StS) is a fundamental daily activity that can be challenging for stroke survivors due to strength, motor
This paper presents an intraoperatively iterative Hough transform (IHT) based in-plane hybrid control of extracorpor
Magnetic microrobots can be navigated by an external magnetic field to autonomously move within living organisms
Severe chronic pulmonary endarterectomy needs a dissector to delicately remove proliferative intima located in the
Retinal surgery is a complex medical procedure that requires high precision dexterity to perform delicate instrument
For robotic transtibial prosthesis control, the global tibia kinematics can be used to monitor gait cycle progression a
The injection of therapeutic agents into the subretinal space might allow improved treatment of age-related macular
Colon endoscopic robots represent a promising screening modality for the visualization of colon cancers with high s
Telesurgical robotic systems provide a well established form of assistance in the operating theater, with evidence o
In the context of telehealth, robotic approaches have proven a valuable solution to in-person visits in remote areas,
The paper introduces a novel robotic system for transanal endoscopic microsurgery (TEM) with a master-slave ope
Ultrasound (US) imaging is widely used in diagnosing and staging abdominal diseases due to its lack of non-ionizin
In ophthalmic surgeries, such as vitreoretinal operations, surgeons rely on imaging systems, primarily microscopes,
Autonomous ultrasound (US) scanning has attracted increased attention, and it has been seen as a potential solutio
This study proposes a method for finding a parasternal long-axis view in echocardiography autonomously with a rob
Medical ultrasound (US) is widely used to evaluate and stage vascular diseases, in particular for the preliminary scr
Osteotomy holds a pivotal position among the fundamental procedures in craniomaxillofacial (CMF) surgery. Howev
Six-axis force/moment (6-A F/M) sensors make surgical robots effectively sense intraoperative force feedback and
In ophthalmic surgery, particularly in procedures involving the posterior segment, clinicians face significant challeng
Prior robot-assisted cranium-milling studies only considered controlling the force in the skull's vertical direction and
One issue of robotic microsurgery is that compared to manual surgery, the operation time tends to be longer due to
Skull base surgery is a demanding field in which surgeons operate in and around the skull while avoiding critical an
The increasing awareness of public health issues has highlighted the need for effective disinfection of crowded indo
Lymphadenectomy generally accompanies various oncology surgeries to remove infected cancer cells. However, th
Laparoscope-holding robots significantly enhance the stability and precision of visualization in minimally invasive su
In the study of minimally invasive surgical robots, a mini parallel continuum robot has shown motion advantage afte
Robot-assisted vascular interventional surgery can isolate interventionists and X-ray radiation, and improve surgica
In robotic-assisted minimally invasive surgery, the remote center of motion (RCM) achieves precision and safe man
This paper proposes a novel form of a three-dimensional coaxial bowtie-shaped mechanical amplifier. The propose
Magnetic continuum robots (MCR) have shown great potential in minimally invasive interventions because they can
This paper proposes an anti-interference targeted therapeutic delivery microrobot, where the targeted therapeutic d
Traditional rigid endoscopes have challenges in flexibly treating tumors located deep in the brain, and low operabilit
Continuum Dexterous Manipulators (CDMs) are well-suited tools for minimally invasive surgery due to their inheren
Concentric tube continuum robots utilize nested tubes, which are subject to a set of inequalities. Current approache
Uptake of capsule endoscopy in the large intestine has been very limited due to both the risk of missed lesions and
Millimeter-scale magnetic rotating swimmers have multiple potential medical applications. They could, for example,
In recent years, the area of Robot-Assisted Minimally Invasive Surgery (RAMIS) is standing on the the verge of a ne
With early sign of bowel cancer being changes in affected lesions biomechanical properties, an AI-assisted dynami
To date, endovascular surgeries are performed using the golden standard of Fluoroscopy, which uses ionising radia
This paper presents a 6-DOF hybrid robot for percutaneous needle intervention procedures. The new robot combin
Several robotic systems have been emerged in the recent past to enhance the precision of micro-surgeries such as
A surgeon's physiological hand tremor can significantly impact the outcome of delicate and precise retinal surgery,
Motion scaling is essential to empower users to conduct precise manipulation during teleoperation for robot-assiste
Capsule endoscopes, predominantly serving diagnostic functions, provide lucid internal imagery but are devoid of s
In single-incision laparoscopic surgery (SILS), magnetic anchoring and guidance system (MAGS) is a promising tec
Helicobacter pylori, a pervasive bacterial infection associated with gastrointestinal disorders such as gastritis, peptic
Research on endovascular surgery robots (ESR) is continuously developing, because ESR can protect surgeons fr
Recently, fiber optic sensors such as fiber Bragg gratings (FBGs) have been widely investigated for shape reconstr
Head and neck cancer is the seventh most common cancer worldwide, with squamous cell carcinoma being the mo
In this work, we propose a semi-autonomous scheme to synergistically share the complicated task of manipulation
The manipulation of instruments under a microscope suffers from physiological tremor and human errors, which are
Several robotic frameworks have been recently developed to assist ophthalmic surgeons in performing complex vitr
Minimally-invasive surgery for stent implantation is a complex procedure requiring specialized instruments. It often l
Magnetic resonance-guided focused ultrasound (MRg-FUS) has become attractive, accrediting to its non-invasive n
Transorbital Neuro Surgery (TNS) offers a novel treatment towards the lesion inside skull pursuing minimal invasive
In minimally invasive endovascular interventional surgery, guidewire navigation is an indispensable process. Howev
Aneurysms, balloon-like bulges in blood vessels, present a signifi cant health risk due to their potential to rupture, le
In this paper, a novel mathematical framework is introduced for modeling deformation behavior of Tendon- Driven C
Percutaneous needle insertions are commonly performed for diagnostic and therapeutic purposes as an effective a
This paper proposes a novel, more computationally efficient method for optimizing robot excitation trajectories for d
Micro-assembly is an emerging method to fabricate microrobots with multiple modules or particles. However, there
In this work, we present a thin-film shape memory alloy (NiTi) microactuator with a magnetic spring. This novel actu
Wirelessly actuated magnetic microrobots are promising tools in medical applications due to their tiny sizes and attr
In this paper, we introduce magnetic mobile micro-gripping microrobots with two independent actuation modes. By
Microswimmers have significant potential for medical applications such as long-term drug administration, precise su
Navigating biomedical instruments inside the brain remains challenging and high-risk. The delicate nature of the tiss
Selective control mechanisms of microrobots have attracted significant attention from researchers. So far, selective
Microsurgical manipulations are key experimental techniques in life science research, particularly in embryology. Th
For the development of robotic systems for micro-manipulation, it is challenging to design appropriate control strate
Knot tying is an essential task for robotic surgery, which is routinely realized by dual-arm robotic manipulation. Desp
Obtaining three-dimensional information, especially the z-axis depth information, is crucial for robotic micromanipula
Optoelectronic tweezers (OET) has shown great promise in various applications, especially in the precise manipula
Obtaining single cell from tissues is important for intersection of information and bioscience research. In this article,
In recent years, micromanipulation technology has achieved extensive applications in industry and life science. Imp
Designing insect-scale robots with high mobility is becoming an essential challenge in the field of robotics research.
Magnetic miniature robots exert forces and torques onto the environment to conduct minimally invasive diagnostic a
Unbiased measurement of sperm morphometric and motility parameters is essential for assessing fertility potential
This paper introduces an innovative multi-scale visual servoing framework for optical microscopy, engineered to au
This paper presented the three-dimensional oocyte manipulation system for the two-electrode voltage clamp (TEVC
This paper presents a functionality that has been developed for the home-made AFM-in-SEM robotic system at the
Automated positioning systems play a pivotal role in micro-scale cell manipulation. In clinical intracytoplasmic sperm
Traditional sperm morphology analysis is based on tedious manual annotation. Automated morphology analysis of
Photoacoustic Microscopy (PAM) is a relatively new imaging modality in biomedicine. However, point-by-point raste
Micromanipulation techniques that can achieve controlled fine operations at the micro scale play an important role i
Until now, the potential of untethered magnetic robots (UMRs), propelled by external time-periodic magnetic fields,
This paper presents Sim-Suction, a robust suction grasp policy for mobile manipulation platforms with dynamic cam
Real-world robot task planning is intractable in part due to partial observability. A common approach to reducing co
Mobile manipulators have recently gained significant attention in the robotics community due to their superior poten
Existing object-search approaches enable robots to search through free pathways, however, robots operating in un
We consider a nonprehensile manipulation task in which a mobile manipulator must balance objects on its end effe
Mobile robots are nowadays frequently used for interaction tasks in the real world, e.g. for opening doors or for pick
Mobile manipulation constitutes a fundamental task for robotic assistants and garners significant attention within the
Legged mobile manipulators are receiving much more attention. Mobile platforms can infinitely expand the workspa
Mobile Manipulation (MoMa) systems incorporate the benefits of mobility and dexterity, due to the enlarged space i
Model-based control requires an accurate model of the system dynamics for precisely and safely controlling the rob
Deformable object manipulation presents a unique set of challenges in robotic manipulation by exhibiting high degre
Advancements in Automated Driving Systems (ADSs) have enabled the achievement of a certain level of autonomy
High-speed autonomous driving in off-road environments has immense potential for various applications, but it also
Robotic adaptation to unanticipated operating conditions is crucial to achieving persistence and robustness in comp
This study presents a recursive algorithm for solving the regularised least squares problem for online identification o
Robotic manipulation of deformable linear objects (DLOs) is an active area of research, though emerging applicatio
Recent advances in aerial robotics have enabled the use of multirotor vehicles for autonomous payload transportati
Skid-Steer Wheeled Mobile Robots (SSWMRs) are increasingly being used for off-road autonomy applications. Wh
Generating human-like robot motions is pivotal for achieving smooth human-robot interactions. Such motions contri
Massage therapy is helpful for the rehabilitation of various diseases, such as headaches caused by migraines and s
Emergent-scene safety is the key milestone for fully autonomous driving, and reliable on-time prediction is essentia
Synthetic data generation is a proven method for augmenting training sets without the need for extensive setups, ye
The soft fingers and strategic grasping skills enable the human hands to grasp objects stably. This paper aims to m
In industrial workstations, the morphology of the worker is a key factor for the feasibility and the ergonomics of an a
Modeling and control of the human musculoskeletal system is important for understanding human motor functions,
Single-view 3D human reconstruction has been a hot topic due to the potential of wide applications. To achieve hig
Human-robot interaction based on real-time kinematics or electromyography (EMG) feedback improves rehabilitatio
Continuum robots with variable stiffness have gained wide popularity in the last decade. Layer jamming (LJ) has em
This paper presents a lumped-parameter dynamic model of a pressure driven eversion robot carrying a catheter thr
Soft robotics is an emergent and swiftly evolving field. Pneumatic actuators are suitable for driving soft robots becau
The passive compliance of soft robotic arms renders the development of accurate kinematic models and model-bas
The small size, high dexterity, and intrinsic compliance of continuum robots (CRs) make them well suited for constr
Soft robots have desirable qualities for use in underwater environments thanks to their inherent compliance and lac
Realtime shape estimation of continuum objects and manipulators is essential for developing accurate planning and
This paper proposes an adaptive radial basis function neural network (RBF NN) based scheme for the dynamics lea
Fully exploiting soft robots’ capabilities requires devising strategies that can accurately control their movements
Active exoskeletons for tremor suppression show potential for treatment of pathological tremor thanks to their non-i
Abstract— Using compliant materials leads to continuum robots undergoing large deformations. Their nonlinear be
In this paper, we present, for the first time, a soft robot control system (SofToss) capable of throwing life-size object
We show how a variety of techniques from Computer Graphics can be leveraged to intuitively control the shape (co
Soft robots are naturally designed to perform safe interactions with their environment, like locomotion and manipula
This paper presents vision-based autonomous navigation of a steerable soft growing robot. Our experimental platfo
Shape estimation is crucial for precise control of soft robots. However, soft robot shape estimation and propriocepti
It is challenging to perform system identification on soft robots due to their underactuated, high-dimensional dynam
While much work has been done recently in the realm of model-based control of soft robots and soft-rigid hybrids, m
This study presents an innovative test rig engineered to explore the kinematic and viscoelastic characteristics of hu
This work investigates the optimal cost function composition for human gait at different walking speeds. Kinematic a
Climbing robots can investigate scientifically valuable sites that conventional rovers cannot access due to steep terr
Payload capacity is a crucial factor for climbing robots, as it directly affects their ability to carry and transport heavy
To achieve high-dexterity motion planning on floating-base systems, the base dynamics induced by arm motions m
Geometric motion planning offers effective and interpretable gait analysis and optimization tools for locomoting syst
Performing autonomous navigation in cluttered and unstructured terrains still remains a challenging task for legged
Ensuring safety and meeting temporal specifications are critical challenges for long-term robotic tasks. Signal temp
Many applications of path planning in time-varying flow fields, particularly in areas such as marine robotics and ship
We propose a new method called Clustering Topological PRM (CTopPRM) for finding multiple distinct paths in 3D c
This paper presents a novel Stochastic Optimal Control (SOC) method based on Model Predictive Path Integral con
This paper introduces a novel, lightweight method to solve the visibility problem for 2D grids. The proposed method
In this paper, we propose a real-time clothoid tree-based path planning for self-driving robots. Clothoids, curves tha
Path planning for multiple non-holonomic robots in continuous domains constitutes a difficult robotics challenge with
Continuous exploration without interruption is important in scenarios such as search and rescue and precision agric
Sampling-based planning algorithms such as RRT have been proved to be efficient in solving path planning problem
Continuum robots (CR) offer excellent dexterity and compliance in contrast to rigid-link robots, making them suitable
We present progress on the problem of reconfiguring a 2D arrangement of building material by a cooperative group
Sampling-based planning algorithms like Rapidly-exploring Random Tree (RRT) are versatile in solving path plannin
Modern sampling-based motion planning algorithms typically take between hundreds of milliseconds to dozens of s
This paper proposes a formulation for the risk-aware path planning problem which utilizes multi-objective optimizatio
Reactive trajectory optimization for robotics presents formidable challenges, demanding the rapid generation of pur
In this paper, a variant of hybrid A* is developed to find the shortest path for a curvature-constrained robot, that is te
This paper introduces a differentiable representation for the optimization of boustrophedon path plans in convex po
Robotic manipulators are essential for future autonomous systems, yet limited trust in their autonomy has confined
We present VAPOR, a novel method for autonomous legged robot navigation in unstructured, densely vegetated ou
Classical motion planning for robotic manipulation includes a set of general algorithms that aim to minimize a scene
Many computations in robotics can be dramatically accelerated if the robot configuration space is described as a co
The multi-robot visibility-based pursuit-evasion problem tasks a team of robots with systematically searching an env
Paths generated by A* and other graph-search-based planners are widely used in the robotic field. Due to the restri
A complete motion planner has long been desired but is hard to achieve in high dimensions. Previous work propose
In this paper, we describe a novel roadmap construction method in unknown environments, which relies on the extr
Robotic exploration or monitoring missions require mobile robots to autonomously and safely navigate between mu
Landing safety is a challenge heavily engaging the research community recently, due to the increasing interest in a
In drone racing, the time-minimum trajectory is affected by the drone's capabilities, the layout of the race track, and
Motion planning techniques for quadrotors have advanced significantly over the past decade. Most successful plann
The robotics community is increasingly interested in autonomous aerial transportation. Unmanned aerial vehicles w
In this paper, an aggressive quadrotor Ball plAying sysTem called BAT is proposed, whose goal is to intercept a flyi
In this work, we present a nonlinear Model Predictive Control (NMPC) scheme for tracking a ground target using a m
Repeated exploration of a water surface to detect objects of interest and their subsequent monitoring is important in
Recent human-robot interaction paradigms on aerial robots unfold many potential applications, and efforts are furth
This article introduces an experimental emulation of a novel chunk-based flexible multi-DoF aerial 3D printing frame
This letter focuses on the design and control of a novel seven-degree-of-freedom (7-DOF) robotic manipulator (D-A

This study proposes a uniform passive fault-tolerant control (FTC) method for a quadcopter that does not rely on fa
This work focuses on the agile transportation of liquids with robotic manipulators. In contrast to existing methods tha
Autonomous vehicle motion control development requires testing and evaluation at all stages of the process. The d
Collision-free navigation is a critical issue in robotic systems as the environment is often dynamic and uncertain. Th
The presented work tackles the question of quantifying the pose deviations of robots subject to external disturbance
This paper proposed an iterative learning control (ILC) scheme for deformable open-frame cable-driven parallel rob
Force-free control (FFC) allows for flexible manipulator motion in response to external forces, making it a vital comp
This paper presents a safety-critical approach to the coordination of robots in dynamic environments. To this end, w
This paper presents a safety-critical control framework tailored for quadruped robots equipped with a roller arm, par
Remote center of motion (RCM) describes a robot with a rod-like end-effector operating through a hole in the interfa
Controller tuning and parameter optimization are crucial in system design to improve both the controller and underly
Retractable worm robots possess hyper-flexibility, allowing them to work in confined spaces that are difficult for hum
To achieve high-accuracy manipulation in the presence of unknown disturbances, we propose two novel efficient an
To enhance the precision of coarse long-stroke motion axes, complementary short-stroke fine positioning stages ar
Disturbance observer (DOB) is a well-known two-loop control structure that imparts robustness to a controller with a
Numerous studies have emphasized the application of autonomous intelligence in human-robot shared control to en
Controlling the contact force on various surfaces is essential in many robotic applications such as in service tasks o
In the context of interaction with unmodelled systems, it becomes imperative for a robot controller to possess the ca
In recent years, there has been a growing demand for robotic manipulators to perform tasks in various unstructured
Model-based approaches for planning and control for bipedal locomotion have a long history of success. It can prov
Nonlinear dynamics bring difficulties to controller design for control-affine systems such as tractor-trailer vehicles, e
Object shaping by grinding is a crucial industrial process in which a rotating grinding belt removes material. Object-
The hybrid nature of multi-contact robotic systems, due to making and breaking contact with the environment, creat
An accurate motion model is a fundamental component of most autonomous navigation systems. While much work
Increasing robotic systems' degrees of freedom (DoFs) makes them more versatile and flexible. This usually render
Rearrangement puzzles are variations of rearrangement problems in which the elements of a problem are potential
Sample-based motion planning approaches, such as RRT*, have been widely adopted in robotics due to their supp
Magnetic modular cubes are cube-shaped bodies with embedded permanent magnets. The cubes are uniformly co
Benefiting from its hyper-redundant structure, the biomimetic snake-like manipulator retains its remarkable flexibility
Smooth and safe speed planning is imperative for the successful deployment of autonomous vehicles. This paper p
Many applications require a robot to accurately track reference end-effector trajectories. Certain trajectories may no
 A Control Barrier Function (CBF)-based motion planning algorithm is proposed. The algorithm explores an unknown
Constrained Motion Planning (CMP) aims to find a collision-free path between the given start and goal configuration
This paper offers a practical method for certifiably safe operations of an unmanned aerial vehicle (UAV) with limited
As the adoption of robots across society increases, so does the importance of considering cybersecurity issues suc
Planning informative trajectories while considering the spatial distribution of the information over the environment, a
We address a challenging problem in motion planning where robots must navigate through narrow passages in thei
This paper introduces an optimisation-based trajectory deformation and tracking algorithm for tethered differential-d
Sampling-based motion planning methods for manipulators in crowded environments often suffer from expensive co
Mobile manipulators have been employed in many applications that are traditionally performed by either multiple fix
In this work, we explore using a transformer-based model for motion planning with task space constraints for manip
This article presents a hybrid motion planning and control approach applicable to various ground robot types and m
Unknown area exploration is a crucial but challeng- ing task for autonomous driving of unmanned ground vehicles (
How easy is it to sneak up on a robot? We examine whether we can detect where people are using only the inciden
For a service robot, it is crucial to perceive as early as possible that an approaching person intends to interact: in th
When integrating robots into human daily life, persuasive power can be essential. However, there are often group d
Robot-human object handover has been extensively studied in recent years for a wide range of applications. Howev
The socially-aware navigation system has evolved to adeptly avoid various obstacles while performing multiple task
To foster an immersive and natural human-robot interaction (HRI), the implementation of tactile perception and feed
odal Perception for HRI I
Interactive hand mesh reconstruction from single-view images poses a significant challenge with the severe occlusi
This paper presents a 2D skeleton-based action segmentation method with applications in fine-grained human activ
Employing skin-like sensors on robots enhances both the safety and usability of collaborative robots by adding the
Interactive visual grounding in Human-Robot Interaction (HRI) is challenging yet practical due to the inevitable amb
Zero-Shot Object Navigation (ZSON) enables agents to navigate towards open-vocabulary objects in unknown env
Interactive Object Grasping (IOG) is the task of identifying and grasping the desired object via human-robot natural
Electrical Impedance Tomography (EIT)-based tactile sensors offer durability, scalability, and cost-effective manufa
In this paper, we focus on inferring whether the given user command is clear, ambiguous, or infeasible in the contex
Socially assistive robots are potentially to be integrated with human daily lives in the near future, and expected to b
Industrial exoskeletons are a potential solution for reducing work-related musculoskeletal disorders during carrying
Ensuring robust tracking of controllers' movement is critical for human-robot interaction in virtual reality (VR) scenar
Integrated aerial Platforms (IAPs), comprising multiple aircrafts, are typically fully actuated and hold significant pote
Multi-robot global localization (MR-GL) with unknown initial positions in a large scale environment is a challenging ta
Collaborative Simultaneous Localization And Mapping (C-SLAM) is a vital component for successful multi-robot ope
Real-time dense reconstruction using Unmanned Aerial Vehicle (UAV) is becoming increasingly popular in large-sc
Collaborative state estimation using heterogeneous multi-sensors is a fundamental prerequisite for robotic swarms
Most distributed algorithms for robot coordination require relative location information, but how to obtain relative loc
In recent years, multiple Light Detection and Ranging (LiDAR) systems have grown in popularity due to their enhan
In the era of advancing autonomous driving and increasing reliance on geospatial information, high-precision mapp
This paper presents a collaborative quadrotor-quadruped robot system for the manipulation of a cable-towed payloa
Coordinating heterogeneous robots is essential for autonomous multi-robot teaming. To execute a set of dependen
Fleets of unmanned robots can be beneficial for the long-term monitoring of large areas, e.g., to monitor wild flocks
With increasing numbers of mobile robots arriving in real-world applications, more robots coexist in the same space
In this paper we present a fully distributed, asynchronous, and general purpose optimization algorithm for Consensu
Looking back at previous diversity work on Rein-forced learning, diversity is often achieved through the augmented
This work is focused on reinforcement learning (RL)-based navigation for drones, whose localisation is based on vis
We propose a novel approach to multi-robot collaboration that harnesses the power of pre-trained large language m
End-to-end deep reinforcement learning (DRL) for quadrotor control promises many benefits -- easy deployment, ta
Deep reinforcement learning (DRL) methods have been widely applied in distributed multi-robotic systems and succ
Rollout algorithms are renowned for their abilities to correct for the suboptimalities of offline-trained base policies. In
It is common for us to feel pressure in a competition environment, which arises from the desire to obtain success co
Communication is a key element in applying multi-agent reinforcement learning to a wide range of real-world scena
Multi-robot systems are gaining popularity over single-agent systems for their advantages. Although they have been
When agents in a multi-robot team have limited knowledge about their relative performance, their teammates, or the
We show that a distributed network of robots or other devices which make measurements of each other can collabo
Recently a line of researches has delved the use of graph neural networks (GNNs) for decentralized control in swar
Mutual localization stands as a foundational component within various domains of multi-robot systems. Nevertheles
Multi-robot collaboration in large-scale environments with limited-sized teams and without external infrastructure is
In this paper, we present an approach for coverage path planning for a team of an energy-constrained Unmanned A
Space-filling building blocks of diverse shape permeate nature at all levels of organization, from atoms to honeycom
This paper explores the optimal containment control problem for nonlinear and underactuated quadrotors with multi
While individual robots are becoming increasingly capable, the complexity of expected missions increases exponen
In this paper, we develop a control framework for the coordination of multiple robots as they navigate through crowd
In this work, we present a coordination strategy tailored for scenarios involving multiple agents and tasks. We devis
We present a multi-robot task and motion planning method that, when applied to the rearrangement of objects by m
Relative navigation methods are a critical enabling technology for the next generation of autonomous spacecraft co
Maximizing the utility of limited Earth observing satellite resources is a difficult ongoing problem. Dynamic Targeting
In the field of resource-constrained robots and the need for effective place recognition in multi-robotic systems, this
We propose a shared semantic map architecture to construct and configure Model Predictive Controllers (MPC) dyn
This paper presents a decentralized cooperative motion planning approach for surface inspection of 3D structures w
This paper proposes a decentralized trajectory planning framework for the collision avoidance problem of multiple m
We propose a decentralised multi-robot coordination algorithm that features a rich representation for encoding and
This paper presents a method to reduce computations for stochastic dynamic games with game-theoretic belief spa
Multi-agent multi-target tracking has a wide range of applications, including wildlife patrolling, security surveillance o
We propose an autonomous exploration algorithm designed for decentralized multi-robot teams, which takes into ac
This paper presents a novel approach to optimal multi-robot task allocation in heterogeneous teams of robots. Whe
Most real-world Multi-Robot Task Allocation (MRTA) problems require fast and efficient decision-making, which is o
This survey article provides a synopsis on some of the engineering problems, technological innovations, robotic dev
This paper concerns the coordination and the traffic management of a group of Automated Guided Vehicles (AGVs
Multiple preferences between robots and tasks have been largely overlooked in previous research on Multi-Robot T
This paper considers the problem of persistently monitoring a set of moving targets using a team of aerial vehicles.
The decentralization of formations using onboard sensing is important for multi-robot systems, improving the robust
This paper investigates the multi-robot patrolling (MuRP) problem in a discrete environment with the objective of ap
This paper focuses on coordinating a robot swarm orbiting a convex path without collisions among the individuals. T
Robot swarms can be tasked with a variety of automated sensing and inspection applications in aerial, aquatic, and
We develop a planner that directs robots to construct a 3D target structure composed of blocks. The robots themse
Anthropomorphic robotic hands have been widely investigated to dexterously manipulate objects because of their a
This work introduces the New Dexterity modular, dexterous, anthropomorphic, open-source, bimanual manipulation
Soft tactile sensors can enable robots to grasp objects easily and stably by simultaneously providing tactile data an
Grasping motion planning aims to find a feasible grasping trajectory in the configuration space given an input target
Dexterous manipulation through imitation learning has gained significant attention in robotics research. The collectio
A robotic hand for social interaction should be capable of comfortable touch with humans. However, it is difficult to m
Conventional industrial robots often use two-fingered grippers or suction cups to manipulate objects or interact with
This paper presents our solution for the Real Robot Challenge III, aiming to address dexterous robotic manipulation
Jig-less assembly of a shaft-shaped part with a single versatile robotic hand requires several functions of the hand
Abstract—Coordinated landing control for multiple unmanned aerial vehicles (UAVs) on appropriate multiple unma
In this work, we propose a framework for the distributed implementation of Quadratic Programs-based controllers, b
Necessary and sufficient conditions are described for stable distance-angle leader-follower formation control of first
The distributed coordination of robot teams performing complex tasks is challenging to formulate. The different aspe
The fifth generation (5G) cellular network technology is mature and increasingly utilized in many industrial and robo
We present CloudGripper, an open source cloud robotics testbed, consisting of a scalable, space and cost-efficient
Cloud service providers provide over 50,000 distinct and dynamically changing set of cloud server options. To help
In this paper we present a new approach to Mobile Infrastructure on Demand (MID) where a dedicated team of robo
This paper addresses the migration of high-frequency robot controllers to remote computing services, which are con
We introduce IFFNeRF to estimate the six degrees-of-freedom (6DoF) camera pose of a given image, building on t
Combining motion prediction in LiDAR-based 3D object detection is an effective method for improving overall accur
Environmental perception tasks such as object detection and free space detection based on 3+1D radar severely s
Recent camera-based 3D object detection is limited by the precision of transforming from image to 3D feature spac
Despite the success of deep learning-based object detection methods in recent years, it is still challenging to make
Detecting objects and estimating their 6D poses is essential for automated systems to interact safely with the enviro
In the expanding landscape of AI-enabled robotics, robust quantification of predictive uncertainties is of great impor
While showing promising results, recent RGB-D camera-based category-level object pose estimation methods have
Category-level object pose estimation demonstrates robust generalization capabilities that benefit robotics applicati
The detection of unknown traffic obstacles is vital to ensure safe autonomous driving. The standard object-detection
3D object detectors for point clouds often rely on a pooling-based PointNet to encode sparse points into grid-like vo
Event cameras are gaining popularity due to their unique properties, such as their low latency and high dynamic ran
3D object detection is an essential vision technique for various robotic systems, such as augmented reality and dom
In LiDAR-based 3D detection, history point clouds contain rich temporal information helpful for future prediction. In t
Autonomous vehicles require real-time, high- performance 3D object detectors to guarantee system robust- ness an
Robots, constrained by limited onboard computing resources, often encounter situations wherein high-resolution an
We present a new 3D point-based detector model, named Shift-SSD, for precise 3D object detection in autonomou
Panoptic segmentation is a challenging perception task, which can help robots to comprehensively perceive the sur
This paper presents a novel framework for robust 3D object detection from point clouds via cross-modal hallucinatio
Unsupervised Domain Adaptive Object Detection (UDA-OD) uses unlabelled data to improve the reliability of robotic
 Fusing the camera and LiDAR information in the unified BEV representation serves as the elegant paradigm for the
In autonomous driving and robotics, there is a growing interest in utilizing short-term historical data to enhance mult
Small object detection has been a challenging problem in the field of object detection. There has been previous wor
The classical human-robot interface in uncalibrated image-based visual servoing (UIBVS) relies on either human an
A vision-based drone-to-drone detection system offers a cost-effective solution for a range of applications, including
Accurate 3D object detection is crucial to autonomous driving. Though LiDAR-based detectors have achieved impre
Despite outstanding semantic scene segmentation in closed-worlds, deep neural networks segment novel instance
In order to successfully perform manipulation tasks in new environments, such as grasping, robots must be proficie
Fully decentralized, multiagent trajectory planners enable complex tasks like search and rescue or package delivery
Affordance detection presents intricate challenges and has a wide range of robotic applications. Previous works hav
Advancements in deep learning-based 3D object detection necessitate the availability of large-scale datasets. Howe
Various heuristic objectives for modeling hand-object interaction have been proposed in past work. However, due to
Dynamic Occupancy Grid Mapping is a technique used to generate a local map of the environment, containing both
In unmanned vehicle perception, dynamic object classification is applied to classify objects accurately and timely, p
Real-time detection of moving objects is an essential capability for robots acting autonomously in dynamic environm
3D object detection (OD) is a crucial element in scene understanding. However, most existing 3D OD models have
Range-View(RV)-based 3D point cloud segmentation is widely adopted due to its compact data form. However, RV
Robots and autonomous vehicles should be aware of what happens in their surroundings. The segmentation and tr
With the rapid development of autonomous driving systems, 3D object detection based on Bird's Eye View (BEV) in
Detection IV
This work proposes a semantic segmentation network that produces high-quality uncertainty estimates in a single fo
Depth-aware panoptic segmentation (DPS) combines image segmentation and monocular depth estimation in a sin
Collaborative perception facilitates a more comprehensive representation of the environment by leveraging complem
We propose Concavity-induced Distance (CID) as a novel way to measure the dissimilarity between a pair of points
Multi-object tracking (MOT) presents a crucial challenge in robotics. Due to limited resources embedded in robots, o
In the field of domain adaptation (DA) on 3D object detection, most of the work is dedicated to unsupervised domain
Multimodal sensor fusion methods for 3D object detection have been revolutionizing the autonomous driving re- sea
Navigating autonomous vehicles efficiently across unstructured and off-road terrains remains a formidable challeng
While LiDAR sensors have been successfully applied to 3D object detection, the affordability of radar and camera s
Semantic image segmentation plays a pivotal role in creating High-Definition (HD) maps for autonomous driving, wh
We propose LiRaFusion to tackle LiDAR-radar fusion for 3D object detection to fill the performance gap of existing
While LiDAR data acquisition is easy, labeling for semantic segmentation remains highly time consuming and must
Active learning strives to reduce the need for costly data annotation, by repeatedly querying an annotator to label th
Tracking dynamic targets poses a significant challenge for Unmanned Ground Vehicles (UGVs). Existing methods o
Navigation in agricultural fields imposes various constraints on manoeuvrability, which can be tackled by using four-
Jerk-constrained trajectories offer a wide range of advantages that collectively improve the performance of robotic s
There has been a growing interest in parallel strategies for solving trajectory optimization problems. One key step in
Nonlinear Model Predictive Control (NMPC) is a state-of-the-art approach for locomotion and manipulation which le
Agile quadrotor flight relies on rapidly planning and accurately tracking time-optimal trajectories, a technology critica
Invariant descriptors of point and rigid-body motion trajectories have been proposed in the past as representative ta
Trajectory retiming is the task of computing a feasible time parameterization to traverse a path. It is commonly used
To enhance robots' applicability in real-world scenarios, it is essential to establish a complex and multi-tasking beha
Trajectory optimization under uncertainty underpins a wide range of applications in robotics. However, existing meth
Intrinsically elastic robots surpass their rigid counterparts in a range of different characteristics. By temporarily storin
Soft robots can execute tasks with safer interactions. However, control techniques that can effectively exploit the sy
Nonlinear model-predictive control has recently shown its practicability in robotics. However it remains limited in con
Many problems in robotics are fundamentally problems of geometry, which lead to an increased research effort in g
We consider the problem of tracking a reference trajectory for dynamical systems subject to a priori unknown state-
Control barrier functions (CBFs) have become popular as a safety filter to guarantee the safety of nonlinear dynami
In this paper, we present a unified framework to ensure the stability, safety, and passivity of a multi-robot teleopera
Sums-of-squares (SOS) optimization is a promising tool to synthesize certifiable controllers for nonlinear dynamical
In this paper, we propose a model predictive control (MPC) that accomplishes interactive robotic tasks, in which mu
The control of granular materials, which are found in many industrial applications, is a challenging open research pr
This paper presents a novel algorithm for the continuous control of dynamical systems that combines Trajectory Op
This work presents a novel Learning Model Predictive Control (LMPC) strategy for autonomous racing at the handli
We study how to safely control nonlinear control-affine systems that are corrupted with bounded non-stochastic noi
This paper develops a controller synthesis method for ensuring an admissible bound of external forces on biped rob
This paper investigates the robust optimal control problem of a class of continuous-time, partially linear, interconnec
Optimal behaviours of a system to perform a specific task can be achieved by leveraging the coupling between traje
This paper presents a whole-body robot control method for exploring and probing a given region of interest. The erg
We present ReLU-QP, a GPU-accelerated solver for quadratic programs (QPs) that is capable of solving high-dime
To date, commercial industrial robots only provide multi-robot coordination for their own fleet of robots and treat rob
Guided trajectory planning involves a leader robot strategically directing a follower robot to collaboratively reach a d
In this article, we consider a multi-agent path planning problem in a partially impeded environment. The impeded en
Information-based coverage directs robots to move over an area to optimize a pre-defined objective function based
In public spaces shared with humans, ensuring multi-robot systems navigate without collisions while respecting soc
Anticipating possible future deployment of connected and automated vehicles (CAVs), cooperative autonomous driv
This paper addresses the problem of the communication of optimally compressed information for mobile robot path
In multi-drone systems such as drone light shows, drones move in formation while avoiding collisions. However, few
This paper addresses multi-robot informative path planning (IPP) for environmental monitoring. The problem involve
Coordinating a fleet of robots in unstructured, human-shared environments is challenging. Human behavior is hard
We address the challenge of efficiently controlling multi-agent systems, crucial in fields like logistics and traffic man
Quadrotor motion planning in complex environments leverage the concept of safe flight corridor (SFC) to facilitate s
This paper presents a scalable multi-robot motion planning algorithm called Conflict-Based Model Predictive Contro
This paper presents a multi-robot kinodynamic motion planner that enables a team of robots with different dynamics
The multi-agent pathfinding (MAPF) problem seeks collision-free paths for a team of agents from their current posit
We present an online centralized path planning algorithm to cover a large, complex, unknown workspace with multi
Multi-robot motion planning (MRMP) is the problem of finding collision-free paths for a set of robots in a continuous
We investigate the time-optimal Multi-Robot Coverage Path Planning (MCPP) problem for both unweighted and we
This paper introduces a Nonlinear Model Predictive Control (N-MPC) framework exploiting a Deep Neural Network
In this work, we investigate the integration of a Dynamic Vision Sensor (DVS) into an Unmanned Aerial Vehicle (UA
We present a fully autonomous self-recharging drone system capable of long-duration sustained operations near po
 In this work, we propose a reaction-based local planner for autonomous collision avoidance of quadrotor in obstacle
Autonomous exploration using unmanned aerial vehicles (UAVs) is essential for various tasks such as building insp
While instance segmentation models excel at object detection in satellite imagery, their performance drops when ap
This paper aims to solve the challenging problems in multi-modal active vision for object detection on unmanned ae
We present an accurate and GPU-accelerated Stereo Visual SLAM design called Jetson-SLAM. It exhibits frame- p
Autonomous navigation in unknown environments with obstacles remains challenging for micro aerial vehicles (MAV
Autonomous collision avoidance requires accurate environmental perception; however, flight systems often posses
This letter describes an incremental multimodal surface mapping methodology, which represents the environment a
In this paper, we address the challenge of exploring unknown indoor environments using autonomous aerial robots
3D coverage path planning for UAVs is a crucial problem in diverse practical applications. However, existing metho
This work proposes an autonomous multi-robot exploration pipeline that coordinates the behaviors of robots in an in
Efficient exploration strategies are vital in tasks such as search-and-rescue missions and disaster surveying. Unma
This work contributes a novel deep navigation policy that enables collision-free flight of aerial robots based on a mo
This paper addresses the problem of traversing through unknown, tilted, and narrow gaps for quadrotors using Dee
In this paper, we consider a Micro Aerial Vehicle (MAV) system teleoperated by a non-expert and introduce a perce
Understanding airflow around a drone is critical for performing advanced maneuvers while maintaining flight stability
Deploying autonomous robots in crowded indoor environments usually requires them to have accurate dynamic obs
Various perception-aware planning approaches have attempted to enhance the state estimation accuracy during m
To support the hypothesis that embracing robotics has the potential to address farming challenges and at the same
Exploiting thermal updrafts to gain altitude can significantly extend the endurance of fixed-wing aircraft, as has been
This paper proposes SANet, a small but accurate detector for aerial flying objects. The detector introduces an atten
Image correspondence serves as the backbone for many tasks in robotics, such as visual fusion, localization, and m
In the realm of robotic grasping, achieving accurate and reliable interactions with the environment is a pivotal challe
Robots and prostheses are increasingly designed with curvilinear surfaces for functional, aesthetic, aerodynamic, a
Nowadays, robotic arms are ubiquitously employed for object manipulation across a spectrum of applications, span
This paper focuses on the sim-to-real issue of RGB-D grasp detection and formulates it as a domain adaptation pro
Transparent objects are ubiquitous in industry, pharmaceuticals, and households. Grasping and manipulating these
Three-dimensional (3D) reconstruction of wiry structures from vision suffers from thin geometry, lack of texture, and
Manipulating liquid is widely required for many tasks, especially in cooking. A common way to address this is extrud
This paper presents a method to learn hand-object interaction prior for reconstructing a 3D hand-object scene from
Currently, truss tomato weighing and packaging require significant manual work. The main obstacle to automation l
ion for Grasping and Manipulation II
Although existing grasp detection methods have achieved encouraging performance under well-light conditions, rep
High-resolution (HR) contact surface information is essential for robotic grasping and precise manipulation tasks. H
From dishwashers to cabinets, humans interact with articulated objects every day, and for a robot to assist in comm
To gently grasp objects, robots need to balance generating enough friction yet avoiding too much force that could d
Cross-pose estimation between rigid objects is a fundamental building block for robotic applications. In this paper, w
6D object pose estimation plays a crucial role in robot grasping and manipulation. However, the prevalent methods
This work demonstrates the utility of dual-arm robots with dual-wrist force-torque sensors in manipulating a Deform
As the basis for prehensile manipulation, it is vital to enable robots to grasp as robustly as humans. Our innate gras
To address the limitations inherent to conventional automated harvesting robots specifically their suboptimal succes
We propose a novel pipeline for unknown object grasping in shared robotic autonomy scenarios. State-of-the-art m
Despite recent exponential advancements in computer vision and reinforcement learning, it remains challenging for
Modern warehouses process millions of unique objects which are often stored in densely packed containers. To au
Tendon-based underactuated hands are intended to be simple, compliant and affordable. Often, they are 3D printed
Accurate 3D object pose estimation is key to enabling many robotic applications that involve challenging object inte
The prospect of assistive robots aiding in object organization has always been compelling. In an image-goal setting
High-frequency, low-latency, 6-DoF object tracking is useful for grasping objects in motion, taking robots beyond pic
The robotic manipulation of deformable objects, such as clothes and fabric, is known as a complex task from both t
We propose a new metric for robot state estimation based on the recently introduced $text{SE}_2(3)$ Lie group def
Offline reinforcement learning has emerged as a promising technology by enhancing its practicality through the use
Autonomous and reliable robotic grasping is a desirable functionality in robotic manipulation and is still an open pro
We present a new reproducible benchmark for evaluating robot manipulation in the real world, specifically focusing
Benchmarking is a common method for evaluating trajectory prediction models for autonomous driving. Existing ben
The 3D~Average Precision (3D AP) relies on the intersection over union between predictions and ground truth obje
This work presents the Human Navigation Simulator (HuNavSim), a novel open-source tool for the simulation of diff
We introduce RobotPerf, a vendor-agnostic benchmarking suite designed to evaluate robotics computing performan
The field of robotics faces inherent challenges in manipulating deformable objects, particularly in understanding and
When interacting with unmodelled dynamic systems, a robot controller should be capable of adapting online its beh
Advances in human sensing and machine learning are paving the way for new applications of robotics in sports and
Traditional strength and conditioning training relies on the utilization of free weights, such as weighted implements,
This paper presents a novel concept to support physically impaired humans in daily object manipulation tasks with a
Haptic upper limb exoskeletons are robots that assist human operators during task execution while having the abilit
Robots have a great potential to help people with movement limitations in activities of daily living, such as dressing.
Exergames have been considered an advanced approach for enhancing the physical activities of the elderly commu
The emergence of exoskeleton technology has enabled new opportunities for gait rehabilitation, but effective metho
The Rating Scale method has been long deemed the standard for measuring subjective perceptions. However, in th
In this paper, we propose a transformer-based architecture for predicting contact forces during a physical human-ro
Vision-based human-to-robot handover is an important and challenging task in human-robot interaction. Recent wo
Collaborative manipulation task often requires negotiation using explicit or implicit communication. An important exa
The advancement and development of human modeling have greatly benefited from principles used in robotics, for
In this paper, we analyze the possibilities offered by Deep Learning State-of-the-Art architectures such as Transform
Mobile physical human-robot collaboration (pHRC) using collaborative robots (cobots) and mobile robots has attrac
Even the best wearable robots face challenges with power losses in the system, especially at the physical attachme
Preserving and encouraging mobility in the elderly and adults with chronic conditions is of paramount importance. H
Multi-agent exploration of a bounded 3D environment with unknown initial positions of agents is a challenging probl
We propose a framework for planning in unknown dynamic environments with probabilistic safety guarantees using
Safe corridor-based Trajectory Optimization (TO) presents an appealing approach for collision-free path planning o
Model Predictive Path Integral (MPPI) control is a type of sampling-based model predictive control that simulates th
Safety is a core challenge of autonomous robot motion planning, especially in the presence of dynamic and uncerta
Real-world problems often require reasoning about hybrid beliefs, over both discrete and continuous random variab
Autonomous agents that operate in the real world must often deal with partial observability, which is commonly mod
Partially observable Markov decision processes (POMDPs) have been widely used in many robotic applications for
Environments with regions of uncertain traversability can be modeled as roadmaps with probabilistic edges for effic
Motion planning under sensing uncertainty is critical for robots in unstructured environments, to guarantee safety fo
Optimal plans in Constrained Partially Observable Markov Decision Processes (CPOMDPs) maximize reward objec
Current methods based on Neural Radiance Fields (NeRF) significantly lack the capacity to quantify uncertainty in t
Robotic manipulation relies on analytical or learned models to simulate the system dynamics. These models are oft
Exploration in dynamic and uncertain real-world environments is an open problem in robotics and it constitutes a fo
We present a novel learning-based trajectory generation algorithm for outdoor robot navigation. Our goal is to comp
In this work, we propose the Informed Batch Belief Trees (IBBT) algorithm for motion planning under motion and se
Ensuring safe navigation in human-populated environments is crucial for autonomous mobile robots. Although rece
For autonomous mobile robots, uncertainties in the environment and system model can lead to failure in the motion
We focus on the problem of long-range dynamic replanning for off-road autonomous vehicles, where a robot plans
Accurately assessing the potential value of new sensor observations is a critical aspect of planning for active perce
Optimal decision-making presents a significant challenge for autonomous systems operating in uncertain, stochasti
Nearly all state-of-the-art SLAM algorithms are designed to exploit patterns in data from specific sensing modalities
This paper proposes an informative trajectory planning approach, namely, adaptive particle filter tree with sigma po
In manipulation tasks like plug insertion or assembly that have low tolerance to errors in pose estimation (errors of t
For safe operation, a robot must be able to avoid collisions in uncertain environments. Existing approaches for moti
We consider a chance-constrained multi-robot motion planning problem in the presence of Gaussian motion and se
Motion planning for a planetary rover involves robotic traversability such that the rover can safely travel without mob
Autonomous navigation requires robots to generate trajectories for collision avoidance efficiently. Although plenty o
This paper studies Reinforcement Learning (RL) techniques to enable team coordination behaviors in graph environ
This paper tackles the dynamic crane scheduling problem in a steel coil warehouse, involving tasks, such as coil sto
A novel algorithm to solve the 2D 𝑘-shortest non-homotopic path planning (𝑘-SNPP) task is proposed in this p
Parking lots and autonomous warehouses for accommodating many vehicles/robots adopt designs in which the und
Many multi-robot deployments, such as automated construction of buildings, distributed search, or cooperative map
Multi-robot systems are becoming increasingly prevalent in various real-world applications, such as manufacturing a
Robotic systems with hundreds or even thousands of robots are widely implemented in logistic and industrial applic
Many applications require the deployment of legged-robot teams to effectively and efficiently carry out missions. Th
Prosthetic legs have been used to restore function in the lower limbs lost due to amputation. Early designs including
In Total Knee Replacement Arthroplasty (TKA), surgical robotics can provide image-guided navigation to fit implant
In the field of human-robot interaction, surface electromyography (sEMG) provides a valuable tool for measuring ac
It is essential to accurately identify gait phases when active exoskeleton devices assist with the lower limbs. This w
Research in powered prosthesis control has explored the use of impedance-based control algorithms due to their b
Inspired by the bionic characteristics of ankle and calf skeletal muscles, a novel ankle-foot prosthesis (AFP) with va
Exoskeleton technologies have numerous potential applications, ranging from improving human motor skills to aidin
The vague interpretation of myoelectrical signals on the residual limb end makes restoring dexterous hand function
Compared to conventional back-support exoskeletons (BSEs) with two motors, BSEs driven by a single motor have
This paper proposes a novel prosthetic wrist that emulates the anatomical structure of the human wrist, specifically
Ankle exoskeletons can assist the ankle joint and reduce the metabolic cost of walking. However, many existing an
Traditionally, powered exoskeletons have predominantly featured a back-enveloping design due to its simplicity in b
Assistive devices, such as exoskeletons and prostheses, have revolutionized the field of rehabilitation and mobility
Balance loss is a significant challenge in lower-limb exoskeleton applications, as it can lead to potential falls, thereb
Optimized assistance patterns have produced the greatest exoskeleton benefits to energy expenditure of any strate
Powered lower-limb prostheses have shown promise in helping individuals with amputation regain functionality that
Exosuits are wearable technologies that improve physical capabilities and mobility providing support during various
Multi-view stereo omnidirectional distance estimation usually needs to build a cost volume with many hypothetical d
On the one hand, cameras of conventional field-of-view usually considered in computer vision and robotics are very
Abstract— In this work, we present RadCloud, a novel real time framework for directly obtaining higher-resolution
This paper is about 3D pose estimation on LiDAR scans with extremely minimal storage requirements to enable sca
Limited by the working principles, LiDAR-SLAM systems suffer from the degeneration phenomenon in environment
Establishing the correspondences between newly acquired points and historically accumulated data (i.e., map) thro
Depth perception is considered an invaluable source of information in the context of 3D mapping and various roboti
The precise point cloud ground segmentation is a crucial prerequisite of virtually all perception tasks for LiDAR sens
Reconstructing accurate object shapes based on single image inputs is still a critical and challenging task, mainly d
Fast and reliable obstacle avoidance is an important task for mobile robots. In this work, we propose an efficient rea
This work explores the use of artificial whiskers as tactile sensors for enhancing the perception and navigation capa
This paper presents a reactive navigation method based on model predictive path integral (MPPI) control with splin
This paper investigates the problem of effective tool manipulation for motion planning in complex human-like scena
In this paper, we propose a reactive planning system for quadruped robots based on prescribed-time control. The n
Robotic-assisted medical systems (RAMS) have gained significant attention for their advantages in alleviating surge
Visual imitation learning provides efficient and intuitive solutions for robotic systems to acquire novel manipulation s
The rise of autonomous driving in everyday life makes efficient and collision-free motion planning more important th
We propose a new method for autonomous navigation in uneven terrains by utilizing a sparse Gaussian Process (S
Robotic exoskeletons can enhance human strength and aid people with physical disabilities. However, designing th
Community-based locomotor training post-stroke has shown improvements in independent ambulation by increasin
Recently, the importance of mechanical transparency in human-assistive robots has grown. Traditionally, its primar
The development of robotic-assisted rehabilitation exercises involving physical human-robot interaction requires ext
This paper presents a method to compute smooth ankle trajectories for lower limb exoskeletons with powered ankle
Upper limb paralysis affects the quality of life. Functional Electrical Stimulation (FES) offers a solution to restore los
The hybridisation of robot-assisted gait training and functional electrical stimulation (FES) can provide numerous ph
Physical therapy (PT) is a key component of many rehabilitation regimens, such as treatments for Parkinson's disea
Offline goal-conditioned reinforcement learning (GCRL) aims at solving goal-reaching tasks with sparse rewards fro
Safe Reinforcement Learning (RL) plays an important role in applying RL algorithms to safety-critical real-world app
Distributional Reinforcement Learning (DRL) not only endeavors to optimize expected returns but also strives to acc
Ensuring safety in dynamic multi-agent systems is challenging due to limited information about the other agents. Co
In Reinforcement Learning, the trade-off between exploration and exploitation poses a complex challenge for achie
Safe reinforcement learning (Safe RL) refers to a class of techniques that aim to prevent RL algorithms from violatin
Perceptive deep reinforcement learning (DRL) has lead to many recent breakthroughs for complex AI systems leve
Multi-task reinforcement learning could enable robots to scale across a wide variety of manipulation tasks in homes
Monocular 3D lane detection has recently received increasing research attention in autonomous driving due to its a
Demand for home-like spatial functions is increasing in the traditional role of automobiles, which had value as a me
 Reinforcement learning for control over continuous spaces typically uses high-entropy stochastic policies, such as G
The prominence of embodied Artificial Intelligence (AI), which empowers robots to navigate, perceive, and engage
The sim-to-real gap poses a significant challenge in RL-based multi-agent exploration due to scene quantization an
Offline reinforcement learning (RL) provides a promising approach to avoid costly online interaction with the real en
In this work, we introduce REFORMA, a novel robust reinforcement learning (RL) approach to design controllers for
Efficiency and performance are significant challenges in applying Machine Learning (ML) to robotics, especially in e
Manipulation of deformable Linear objects (DLOs), including iron wire, rubber, silk, and nylon rope, is ubiquitous in
Reinforcement learning (RL) with linear temporal logic (LTL) objectives can allow robots to carry out symbolic event
A holistic understanding of object properties across diverse sensory modalities (e.g., visual, audio, and haptic) is es
Rearrangement planning for object retrieval tasks from confined spaces is a challenging problem, primarily due to th
This paper presents a flexible representation of neural radiance fields based on multi-plane images (MPI), for high-q
Collision detection is one of the most time-consuming operations during motion planning. Thus, there is an increasi
Neural fields, coordinate-based neural networks, have recently gained popularity for implicitly representing a scene.
The manual annotation for large-scale point clouds is still tedious and unavailable for many harsh real-world tasks.
There have recently been large advances both in pre-training visual representations for robotic control and segmen
We study the problem of hanging a wide range of grasped objects on diverse supporting items. Hanging objects is a
The basis for robotics skill learning is an adequate representation of manipulation tasks based on their physical pro
Learning multi-object dynamics from visual data using unsupervised techniques is challenging due to the need for r
We present a large empirical investigation on the use of pre-trained visual representations (PVRs) for training down
Historically, feature-based approaches have been used extensively for camera-based robot perception tasks such a
Visual control policies can encounter significant performance degradation when visual conditions like lighting or cam
A good representation of a large, complex mobile robot workspace must be space-efficient yet capable of encoding
Generic re-usable pre-trained image representation encoders have become a standard component of methods for m
entation Learning II
This paper studies the problem of extracting planar regions in uneven terrains from unordered point cloud measure
General scene reconstruction refers to the task of estimating the full 3D geometry and texture of a scene containing
To address the problem that traditional geometric registration based estimation methods only exploit the CAD mode
Accurate and dense depth estimation with stereo cameras and LiDAR is an important task for automatic driving and
With the rise in consumer depth cameras, a wealth of unlabeled RGB-D data has become available. This prompts t
Introducing MMAUD: a Multi-Modal Anti-UAV Dataset, developed in response to the evolving challenges posed by s
Object instance segmentation is a key challenge for indoor robots navigating cluttered environments with many sma
Segmenting unseen objects from images is a critical perception skill that a robot needs to acquire. In robot manipul
We introduce SLCF-Net, a novel approach for the Semantic Scene Completion (SSC) task that sequentially fuses L
Depth estimation models have shown promising performance on clear scenes but fail to generalize to adverse wea
Collaborative decision-making is an essential capability for multi-robot systems, such as connected vehicles, to coll
We investigate a variation of the 3D registration problem, named multi-model 3D registration. In the multi-model reg
3D perception of deformable linear objects (DLOs) is crucial for DLO manipulation. However, perceiving DLOs in 3D
We provide methods which recover planar scene geometry by utilizing the transient histograms captured by a class
Understanding the real world through point cloud video is a crucial aspect of robotics and autonomous driving syste
For robots to perform a wide variety of tasks, they require a 3D representation of the world that is semantically rich,
The TrackDLO algorithm estimates the shape of a Deformable Linear Object (DLO) under occlusion from a sequen
Existing 3D understanding datasets typically provide annotations for a limited number of object classes, with sufficie
The joint design of origami-inspired robots is one of the most distinctive features that distinguishes them from conve
In response to the need for sustainable energy solutions, solar panels have gained significant traction. One way to
Inspired by the necessity of morphological adaptation in animals, a growing body of work has attempted to expand
There are many instances of helical mechanisms that are used to efficiently grasp different objects with various sha
Unlike parallel robots, for which hundreds of different architectures have been proposed, the vast majority of six-deg
This paper presents a free-swimming, tetherless, cable-driven modular soft robotic fish. The body comprises a serie
This paper introduces DRAGON: Deformable Robot for Agile Guided Observation and Navigation, a free-swimming
A kinematically redundant (6+3)-degree-of-freedom (DOF) hybrid parallel robot with an axisymmetric workspace is
The tensegrity structure was initially created as a static structure, but it has gained significant attention among robo
Large, high-capacity models trained on diverse datasets have shown remarkable successes on efficiently tackling d
We present Self-Adaptive Robust Attention for Robotics Transformers (SARA-RT): a new paradigm for addressing
Dexterous manipulation, especially of small daily objects, continues to pose complex challenges in robotics. This pa
Although pre-training on a large amount of data is beneficial for robot learning, current paradigms only perform larg
We pursue the goal of developing robots that can interact zero-shot with generic unseen objects via a diverse reper
In order for a bimanual robot to manipulate an object that is held by both hands, it must construct motion plans such
Safety is a fundamental requirement of many robotic systems. Control barrier function (CBF)-based approaches ha
Ensuring safety in real-world robotic systems is often challenging due to unmodeled disturbances and noisy sensor
Much effort has been devoted in the field of reachability analysis to obtaining control-invariant sets which ensure tha
Control barrier functions (CBFs) have been widely applied to safety-critical robotic applications. However, the const
Cloth manipulation is a category of deformable object manipulation of great interest to the robotics community, from
Deploying autonomous systems in safety critical settings necessitates methods to verify their safety properties. This
Dynamic locomotion in legged robots is close to industrial collaboration, but a lack of standardized testing obstructs
Autonomous systems, such as self-driving cars and drones, have made significant strides in recent years by levera
A key source of brittleness for robotic systems is the presence of model uncertainty and external disturbances. Exis
Perception modules are integral in many modern autonomous systems, but their accuracy can be subject to the vag
Multi-robot exploration is a complex problem that involves multiple robots working in a shared unknown environmen
In recent years, advanced model-based and data-driven control methods are unlocking the potential of complex rob
Safety is often the most important requirement in robotics applications. Nonetheless, control techniques that can pr
Autonomous robots navigating in changing environments demand adaptive navigation strategies for safe long-term
Collaborative robots are more and more used in applications requiring robots and humans to work in proximity or di
This paper addresses the problem of safety-critical control of autonomous robots, considering the ubiquitous uncert
We introduce Conformal Decision Theory, a framework for producing safe autonomous decisions despite imperfect
Ensuring robot safety in complex environments is a difficult task due to actuation limits, such as torque bounds. Thi
Given its wide application in robotics, point cloud registration is a widely researched topic. Conventional methods ai
In order to navigate safely and reliably in novel environments, robots must estimate perceptual uncertainty when co
In this paper, we strongly advocate square-root covariance (instead of information) filtering for visual-inertial naviga
3D single object tracking (SOT) is an important and challenging task for the autonomous driving and mobile robotic
Neural implicit representations have emerged as a promising solution for addressing the challenges of Simultaneou
Real-world robotic tasks stretch over extended horizons and encompass multiple stages. Learning long-horizon ma
We present and evaluate a novel suction-based gripper designed for fruit picking. This work is motivated by commo
Labor shortages and the development of digital technology both impose requirements on the fruit industry. Modern
Grain Appearance Inspection (GAI) is a pre-requisite for grain quality determination, providing guidance for grain pr
Satellite imagery has traditionally been used to collect crop statistics, but its low resolution and registration accurac
Pruning weight is indicative of a vine’s ability to produce crop the following year, informing vineyard man- agem
Measuring plant traits with a high throughput allows breeders to monitor and select the best cultivars to be used in f
There is growing interest in automating agricultural tasks that require intricate and precise interaction with specialty
This paper introduces a prioritized path-planning algorithm for heterogeneous tasks performed by multiple unmanne
Crops for food, feed, fiber, and fuel are key resources for our society. Monitoring plants and measuring their traits is
This paper focuses on developing a robot-assisted system for stem water potential (SWP) measurement in orchard
Automation application in crop harvesting has increased in the past decades. Various types of harvesting robots are
Advancements in machine learning, computer vision, and robotics have paved the way for transformative solutions
Tracking of plant organs over spatio-temporal sequence of point cloud data is one of the demanding tasks of agricu
With the need to feed a growing world population, the efficiency of crop production is of paramount importance. To
Autonomous aerial manipulation requires the capability to handle inevitable dynamic changes during physical intera
This paper presents an online trajectory planning algorithm for an Unmanned Aerial Vehicle (UAV) to autonomously
In this article, we propose a control solution for the safe transfer of a quadrotor UAV between two surface robots po
3D scene graphs have recently emerged as a powerful and human-understandable way of representing complex 3D
The increasing usage of modern AI techniques represents a transforming shift in the robotics domain. Training and
The aging population and diminishing working population in agriculture motivate the development of autonomous ha
In Japan, labor shortage of agriculture is becoming increasingly severe due to the lack of farmers and aging. There
Soil is the most important building element of agriculture and its analysis is crucial for healthy plants and a high crop
Dormant pruning is an important orchard activity for maintaining tree health and producing high-quality fruit. Due to
By-tree information gathering is an essential task in precision agriculture achieved by ground mobile sensors, but it
In this paper, we present a next-best-view planning approach to autonomously size apple fruitlets. State-of-the-art v
Robots are increasingly used in tomato greenhouses to automate labour-intensive tasks such as selective harvestin
Containerized vertical farming is a type of vertical farming practice using hydroponics in which plants are grown in v
Perception datasets for agriculture are limited both in quantity and diversity which hinders effective training of super
Agricultural robots have the potential to improve the efficiency and sustainability of existing agricultural practices. M
The timber-harvesting industry is lagging its peer industries, such as mining and agriculture, with respect to deploym
In this paper, a region-determined navigation method applied for unmanned ground vehicles (UGVs) is presented. T
Forests play a crucial role in our ecosystems, functioning as carbon sinks, climate stabilizers, biodiversity hubs, and
The efficient collection of samples is an important factor in outdoor information gathering applications on account of
Herding is performed by people or trained animals to control the movement of livestock under the desired direction
This paper presents the hardware design and development of Hyblock, a modular hydraulic robot for heavy-duty ap
Plaster is commonly used in the construction industry to finish walls and ceilings, but the application is labor-intensi
We present a robotic system for the assembly of 3D discrete lattice structures in which the robots are able to self-re
We present LiSTA (LiDAR Spatio-Temporal Analysis), a system to detect probabilistic object-level change over time
We present a scalable combined localization infrastructure deployment and task planning algorithm for underwater
Loading multiple different materials with wheel loaders is a challenging task because various materials require diffe
Excavation is a frequent task in construction. In this context, automation is expected to reduce hazard risks and lab
Concrete infrastructure often develops a variety of internal flaws that cannot be detected through visual inspection a
This letter introduces ERRA, an embodied learning architecture that enables robots to jointly obtain three fundamen
Foundation models such as ChatGPT have made significant strides in robotic tasks due to their universal represent
Assistive agents performing household tasks such as making the bed or cooking breakfast often compute and exec
This paper combines two contributions. First, we introduce an extension of the Meta-World benchmark, which we ca
Designing robotic agents to perform open vocabulary tasks has been the long-standing goal in robotics and AI. Rec
Recent advances in metric, semantic, and topological mapping have equipped autonomous robots with concept gro
Extracting knowledge and reasoning from large language models (LLMs) offers a path to designing intelligent robot
Task-oriented grasping (TOG) refers to the problem of predicting grasps on an object that enable subsequent mani
We propose a new system named DOS(D eployment O perating System for RoboOps) for reliably deploying any da
This paper presents a robust image-based visual servoing design for a quad-rotor unmanned aerial vehicle perform
This letter presents the design and experimental study of an adaptive nonlinear controller for Unmanned Aerial Veh
In this paper, a new kind of adaptive controller for the problem of output feedback tracking is proposed on the basis
Model Predictive Control (MPC) has exhibited remarkable capabilities in optimizing objectives and meeting constrai
We report a novel model-based nullspace adaptive trajectory-tracking control (NS-ATTC) algorithm for fully-actuate
Autonomous racing is a challenging problem, as the vehicle needs to operate at the friction or handling limits in ord
Designing robust algorithms in the face of estimation uncertainty is a challenging task. Indeed, controllers seldom c
This paper researches a novel global terminal sliding mode control(GTSMC) on a tethered satellites system(TSS) u
Task-space quadratic programming (QP) is an elegant approach for controlling robots subject to constraints. Yet, in
Integrating humanoid service mobile robots into human environments presents numerous challenges, primarily con
In the field of Learning from Demonstration (LfD), Dynamical Systems (DSs) have gained significant attention due t
In human-robot collaboration, there has been a trade-off relationship between the speed of collaborative robots and
Handovers play a key role in human-robot interactions. However, current research focuses on visible-hand handove
Humans have a remarkable ability to fluently engage in joint collision avoidance in crowded navigation tasks despite
This paper presents a unified prediction and planning algorithm for an autonomous vehicle to interact with an uncer
Perception plays a pivotal role in enhancing the functionality of autonomous agents. However, the intricate relations
Passivity is necessary for robots to fluidly collaborate and interact with humans physically. Nevertheless, due to the
AI-enabled collaboration robots are designed to be used in close collaborations with humans, thus demanding strin
The use of autonomous underwater vehicles (AUVs) to accomplish traditionally challenging and dangerous tasks ha
Powerline robotics is slowly becoming key tools for electric utilities. Contrary to drones that are usually limited to ins
Although continuum robots have the potential to operate in narrow areas by changing their shapes and propelling th
Mission-critical operations, particularly in the context of Search-and-Rescue (SAR) and emergency response situati
In this article, we propose a novel navigation framework that leverages a two layered graph representation of the en
Search for targets in a 3D environment can be formulated as submodular maximization problems with routing const
The efficient task allocation is a challenge for multi-robot search. The multi-robot search problem is reformulated as
In the paper, we address wireless communication infrastructure building by relay placement based on approaches u
This systems paper presents the implementation and design of RB5, a wheeled robot for autonomous long-term ex
In our previous work, we designed a human-like white-box and causal generative model of perception NaivPhys4RP
Robots require a semantic understanding of their surroundings to operate in an efficient and explainable way in hum
3D visual grounding is a critical skill for household robots, enabling them to navigate, manipulate objects, and answ
Estimating the traversability of terrain should be reliable and accurate in diverse conditions for autonomous driving
Huge progress has been made in LiDAR 3D semantic segmentation, but there are still two under-explored imbalanc
Current state-of-the-art methods for panoptic segmentation require an immense amount of annotated training data
In the fields of robotic perception and computer vision, achieving accurate semantic segmentation of low-light or nig
The complementarity between camera and LiDAR data makes fusion methods a promising approach to improve 3D
Inventory monitoring in homes, factories, and retail stores relies on maintaining data despite objects being swapped
Robots performing human-scale manipulation tasks require an extensive amount of knowledge about their surround
The availability of real-time semantics greatly improves the core geometric functionality of SLAM systems, enabling
In this article, we introduce a novel strategy for robotic exploration in unknown environments using a semantic topo
Precise 3D environmental mapping with semantics is essential in robotics. Existing methods often rely on predefine
Accurately perceiving and tracking instances over time is vital for the decision-making processes of autonomous ag
Scene understanding is crucial for autonomous systems to reliably navigate in the real world. Panoptic segmentatio
Visual Semantic Navigation (VSN) aims at navigating a robot to a given target object in a previously unseen scene.
In this article, we propose the novel concept of Belief Scene Graphs, which are utility-driven extensions of partial 3D
Visual topic modeling (VTM) provides key insight into data sets based on learned semantic topic models. The Gaus
Collision-free space detection is of utmost importance for autonomous robot perception and navigation. State-of-the
In the domain of 3D point cloud scene semantic segmentation, a preponderance of methodology predominantly ado
The perception of moving objects is crucial for autonomous robots performing collision avoidance in dynamic enviro
Semantic 3D mapping, the process of fusing depth and image segmentation information between multiple views to
High-definition (HD) maps play a crucial role in autonomous driving systems. Recent methods have attempted to co
RGB-thermal semantic segmentation is one potential solution to achieve reliable semantic scene understanding in a
Maps have played an indispensable role in enabling safe and automated driving. Although there have been many a
A recent sensor fusion in a Bird's Eye View (BEV) space has shown its utility in various tasks such as 3D detection,
The exceptional mobility and long endurance of air-ground robots are raising interest in their usage to navigate com
Complementary to prevalent LiDAR and camera systems, millimeter-wave (mmWave) radar is robust to adverse we
Cameras and LiDARs are both important sensors for autonomous driving, playing critical roles in 3D object detectio
The ability to detect objects in all lighting (i.e., normal-, over-, and under-exposed) conditions is crucial for real-world
2D RGB images and 3D LIDAR point clouds provide complementary knowledge for the perception system of auton
Enhancing the generalization capacity of robotic perception systems for safety-critical applications is vital, especiall
Features from LiDAR and cameras are considered to be complementary. However, due to the sparsity of the LiDAR
High-definition (HD) semantic map generation of the environment is an essential component of autonomous driving
We introduce a novel framework of continuous-time ultra-wideband-inertial sensor fusion for online motion estimatio
Accurate navigation is essential for autonomous robots and vehicles. In recent years, the integration of the Global N
Event-based moving object detection is a challenging task, where static background and moving object are mixed t
Combining LiDAR and camera data has shown potential in enhancing short-distance object detection in autonomou
We present a novel approach for metric dense depth estimation based on the fusion of a single-view image and a s
The ability of a robot to perceive and understand its environment is crucial for its actions and behavior. Humans are
This paper presents an investigation into the estimation of optical and scene flow using RGBD information in scena
We propose an extrinsic calibration method for LiDAR-camera fusion systems using variations in intensities projecte
Monocular depth estimation is a crucial task to measure distance relative to a camera, which is important for applica
This paper considers homography estimation in a Bayesian filtering framework using rate gyro and camera measur
We propose a novel angular velocity estimation method to increase the robustness of Simultaneous Localization An
Tracking a specific person in 3D scene is gaining momentum due to its numerous applications in robotics. Currently
As intelligent systems become increasingly important in our daily lives, new ways of interaction are needed. Classic
This study proposes novel control methods that lower impact force by preemptive movement and smoothly transitio
Control barrier functions have become an increasingly popular framework for safe real-time control. In this work, we
In this paper, we discuss the stability analysis of Plane-to-Plane positioning task when the task is designed in proxim
Visual odometry system is challenged by complex illumination environments. Image quality and its consistency in th
Accurate robotic control over interactions with the environment is fundamentally grounded in understanding tactile c
We present a novel sound source localization method that leverages microphone pair training, designed to deliver r
In the last few years, mobile robots such as floor cleaners, assistive robots, and home telepresence have become a
Human-like planning skills and dexterous manipulation have long posed challenges in the fields of robotics and artif
Robots that operate in agricultural environments need a robust perception system that can deal with occlusions, wh
To achieve longer driving distances, planetary robotics missions require accurate localization to counteract position
Legged robots have the potential to traverse complex terrain and access confined spaces beyond the reach of trad
A general-purpose service robot (GPSR), which can execute diverse tasks in various environments, requires a syst
In this work, we propose a novel strategy to ensure infants, who inadvertently displace their quilts during sleep, are
Image servo is an indispensable technique in robotic applications that helps to achieve high precision positioning. T
Visual servoing (VS) is a widely used technique in industries where there are hundreds of robots, but it requires acc
Introducing robotics in the healthcare sector revolutionizes medical services by providing advanced treatments, me
In the healthcare industry, nursing robots have made great contributions, assisting in the delivery of food and medic
Fiducial markers enable reliable service robot control. In human-robot coexistence environments, efficient placemen
Over the last years, powerful methods for solving specific perception problems such as object detection, pose estim
Floor cleaning robots have been developed to cater to building maintenance needs. Complete area coverage is cru
In the pursuit of fully autonomous robotic systems capable of taking over tasks traditionally performed by humans, t
We present a novel convex formulation that models rigid and deformable bodies coupled through frictional contact.
Navigation through crowded human environments is challenging for social robots. While reinforcement learning has
Multi-robot navigation and dynamic obstacle avoidance are challenging problems in robot learning. Recent advance
The use of simulation in robotics is increasingly widespread for the purpose of testing, synthetic data generation an
Hydrodynamic interactions are crucial for understanding fish movement, particularly within the realm of robotic appl
Developing algorithms for extra-terrestrial robotic exploration has always been challenging. Along with the complexi
We consider the task of traffic scene generation. A common approach in the self-driving industry is to use manual c
We present Jade, a differentiable physics engine for articulated rigid bodies. Jade models contacts as the Linear Co
Simulation is a key technology in robotics as it enables the generation of environmental data and testing scenarios
Sparse keypoint matching based on distinct 3D feature representations can improve the efficiency and robustness o
Significant advances in robotics and machine learning have resulted in many datasets designed to support research
This paper introduces a novel sonar odometry system that estimates the relative spatial transformation between two
Maps of LiDAR Simultaneous Localisation and Mapping (SLAM) are often represented as point clouds. They usuall
Enabling autonomous robots to operate robustly in challenging environments is necessary in a future with increased
Multi-camera systems have been shown to improve the accuracy and robustness of SLAM estimates, yet state-of-t
We introduce Ground-Fusion, a low-cost sensor fusion simultaneous localization and mapping (SLAM) system for g
Simultaneous Localization and Mapping (SLAM) is a fundamental task in robotics, driving numerous applications su
Neural implicit representation has recently achieved significant advancements, especially in the field of SLAM(Simu
This paper proposes a new nonlinear factor sparsification paradigm for general feature-based long-term SLAM bac
The ability to detect loop closures plays an essential role in any SLAM system. Loop closures allow correcting the d
Local geometric information, i.e., normal and distribution of points, is crucial for LiDAR-based simultaneous localiza
Odometry and mapping play a pivotal role in the navigation of autonomous vehicles. In this paper, we address the p
In SLAM (Simultaneous localization and mapping) problems, Pose Graph Optimization (PGO) is a technique to refin
Correspondence matching plays a crucial role in numerous robotics applications. In comparison to conventional han
In recent years, object-oriented simultaneous localization and mapping (SLAM) has attracted increasing attention d
Not relying on ground-truth data for training, self-supervised semantic visual odometry (SVO) has recently gained c
Compared to polynomial splines, rational functions are known to be more efficient and well-behaved data fitting mo
Direct methods for event-based visual odometry solve the mapping and camera pose tracking sub-problems by est
In this paper, we present an evolved version of Situational Graphs, which jointly models in a single optimizable facto
Localization is an essential capability for mobile robots. A rapidly growing field of research in this area is Visual Plac
State estimation is a crucial component for the successful implementation of robotic systems, relying on sensors su
We introduce a new algorithm that utilizes semantic information to enhance feature matching in visual SLAM pipelin
Recent deep learning based visual simultaneous localization and mapping (SLAM) methods have made significant
We propose a new method for fine registering multiple point clouds simultaneously. The approach is characterized
Accurate and robust LiDAR odometry is a crucial technology for robot localization. However, motion distortion and r
This paper proposes LONER, the first real-time LiDAR SLAM algorithm that uses a neural implicit scene representa
Odometry estimation is crucial for every autonomous system requiring navigation in an unknown environment. In m
The integration of multiple LiDAR sensors has the potential to significantly enhance odometry systems by providing
Most Simultaneous localisation and mapping (SLAM) systems have traditionally assumed a static world, which does
This paper presents VoxelMap++: a voxel mapping method with plane merging which can effectively improve the ac
Simultaneous determination of sensor poses and scene geometry is a fundamental problem for robot vision that is o
3D point cloud maps are widely used in robotic tasks like localization and planning. However, dynamic objects, such
In this paper, we propose a novel LiDAR(-inertial) odometry and mapping framework to achieve the goal of simultan
The number and arrangement of sensors on mobile robot dramatically influence its perception capabilities. Ensuring
Visual SLAM is an essential tool in diverse applications such as robot perception and extended reality, where featur
In this study, we propose an object shape alignment method through a robust optimization scheme for 6-degrees-o
In this paper, we propose a trade-off optimization algorithm to compute an appropriate number of edges for measur
We consider the problem of learning error covariance matrices for robotic state estimation. The convergence of a st
LiDAR and cameras are frequently used as sensors for simultaneous localization and mapping (SLAM). However, t
Radar offers the advantage of providing additional physical properties related to observed objects. In this study, we
We present a new SLAM algorithm that utilizes an inexpensive four-point LiDAR to supplement the limitations of the
Existing methods for place recognition in 3D point clouds either ignore partial structure information by converting 3D
Versatile and adaptive semantic understanding would enable autonomous systems to comprehend and interact with
The precise estimation of camera poses within large camera networks is a foundational problem in computer vision
Autonomous navigation is one of the key requirements for every potential application of mobile robots in the real-wo
Rather than having each newly deployed robot create its own map of its surroundings, the growing availability of SL
Implicit neural representations have demonstrated significant promise for 3D scene reconstruction. Recent works h
When a mobile robot autonomously explores an indoor space to produce a localization and navigation map, it is imp

Event-based cameras asynchronously capture individual visual changes in a scene. This makes them more robust
 This letter re-visits the problem of visual-inertial navigation system (VINS) and presents a novel filter design we dub
To achieve precise localization for autonomous vehicles and mitigate the problem of accumulated drift error in odom
Autonomous mobile robots have become popular in various applications coexisting with humans, which requires rob
Consider a robot tasked with tidying a desk with a meticulously constructed Lego sports car. A human may recogniz
In this paper, we investigate how a customer-facing service robot can support decision making in shopping interacti
The choices made by autonomous robots in social settings bear consequences for humans and their presumptions
Safe and efficient navigation in dynamic environments shared with humans remains an open and challenging task f
This work explores a novel approach to empowering robots with visual perception capabilities using textual descript
The inexorable progress of technology brought forth an era where robots increasingly integrate into human life whic
Social robots for children have focused mainly on conventional education domains such as teaching language, scie
Different people have different preferences when it comes to human-robot interaction. Therefore, it is desirable for t
This paper presents a novel concept, Rotenna (Robot-Antenna), which aims to address the limitations of existing re
Continuum robots, characterized by their high flexibility and infinite degrees of freedom (DoFs), have gained promin
In-flight high-speed object capturing is crucial in nature to improve survival and adaptation to the environment, such
Soft robots based on particle jamming cannot return to the initial position and initial mechanical state due to the acc
Deep-sea research represents invaluable opportunities to unravel hidden ecosystems, uncover unknown biodiversit
The navigational abilities of tip-everting soft growing robots, known as vine robots, are compromised when tip-moun
This study tackles the representative yet challenging contact-rich peg-in-hole task of robotic assembly, using a soft
Soft continuum robots enable new application areas in contrast to standard rigid robots, such as interaction with a v
Soft robotic grippers are intrinsically delicate while grasping objects, and can rely on mechanical deformation to ada
Soft wall-climbing robots have been limited in their ability to perform complex locomotion in diverse environments du
Unstructured environments call for versatile robots with adaptable morphology that can perform multiple goal-direct
Soft robotic manipulators have inherent advantages in underwater applications, as they generate motion by deform
Compared to the traditional robot, which is rigidly structured, the soft robot, usually made of soft material, or followin
The solar array wing deployment of orbiting satellites cannot be performed due to power failure of the connector ca
Robot-assisted Natural Orifice Transluminal Endoscopic Surgery (NOTES) represents a paradigm shift in surgical p
How to efficiently traverse complex terrain remains an unresolved challenge for mobile soft robots, because their de
Global blackberry demand has been surging due to their antioxidant and nutritional value in a traditional diet. Howe
Aquatic organisms, due to soft body structure and high agility, have inspired many biomimetic robots. However, con
Robot mechanisms that exploit compliance can perform complex tasks under uncertainty using simple control strate
Research in the area of photo-actuation is growing rapidly, yet there are few examples of photo-actuators with prac
Many robotic hands currently rely on extremely dexterous robotic fingers and a thumb joint to envelop themselves a
bot Materials and Design I
The inherent elasticity of soft materials can be used to create robotic grippers that deform and comply to a variety o
Origami designs and structures have been widely used in many fields, such as morphing structures, robotics, and m
Magnetic manipulation of miniature soft or liquid robots capable of deformation has gained increasing attention and
Dielectric elastomer actuators (DEAs), a type of "artificial muscles", can generate significant deformations and offer
Lattice metamaterials exhibit diverse functions and complex spatial deformations by rational structural design. Here
Abstract— Soft manipulators offer the advantages of safety and adaptability. However, due to insufficient stiffness
Soft robotic manipulators with many degrees of freedom can carry out complex tasks safely around humans. Howe
Hand extension is crucial for stroke survivors with spasticity, where their fingers become rigid and their thumb rema
Soft actuators have shown advantages of adaptiveness, large deformation, and safe human-robot interaction, maki
Soft growing robots have garnered significant research interest owing to their unique locomotion. However, real-wo
Robot performance has advanced considerably both in and out of the factory, however in tightly constrained, unkno
This study introduces a novel robotic module that integrates three spring-reinforced soft actuators for positioning the
Stability and reliable operation under a spectrum of environmental conditions is still an open challenge for soft and c
Unmanned air vehicles (UAVs) have traditionally been considered as “eyes in the sky”, that can move in thre
Sensing plays a pivotal role in robotic manipulation, dictating the accuracy and versatility with which objects are han
In this paper, the development of a method for the design of soft pneumatic actuators is described. The focus is giv
Soft robotic gloves can not only provide timely, effective, safe and cheap rehabilitation training for patients with imp
For over thirty years optical belt drive configurations are being used in food industries to automatically sort produce
This study presents the design and validation of a soft sliding stiffness structure with a soft-rigid layer sliding mecha
In neurosurgery, soft robots have the potential to introduce significant benefits over traditional metal tools for their a
In this paper, we present a fully soft miniaturized continuum robot that integrates 3D optical shape sensing through
The emergence of intelligent prostheses has facilitated the life and work of disabled patients. The interaction aspec
We tackle the problem of proprioception in soft robots, specifically soft grippers with tight packaging constraints, rel
Learning manipulation from demonstration is a key way for humans to teach complex tasks. However, this domain m
The continuous electromechanical deformation of dielectric elastomer actuators (DEAs) suffers from rate-dependen
Soft continuum robots, fabricated from elastomeric materials, offer unparalleled flexibility and adaptability, making th
Soft actuators with a function of variable stiffness are beneficial to the improvement of the adaptability of robots, ex
The fields of soft and bio-inspired robotics promise to imbue synthetic systems with capabilities found in the natural
A homogeneous pneumatic soft robot may generate complex output motions using a simple input pressure, resultin
Soft robots have shown remarkable distinct capabilities due to their high deformation. Recently increasing attention
Designing information-rich and space-efficient sensors is a key challenge for soft robotics, and crucial for the develo
MoRC is a high-performance modular robot controller based on the Functional Mock-up Interface (FMI) standard. T
In an increasingly automated world – from warehouse robots to self-driving cars – streamlining the developmen
We describe AOS, the first general-purpose system for model-based control of autonomous robots using AI plannin
Selecting an optimal robot, its base pose, and trajectory for a given task is currently mainly done by human expertis
Gas Source Localization (GSL) is a challenging field of research within the robotics community, with high-stakes se
Cooking recipes are challenging to translate to robot plans as they feature rich linguistic complexity, temporally-exte
Despite significant technological advancements, the process of programming robots for adaptive assembly remains
This paper introduces a method that enables the parallel independent execution of trajectories for multi-robot multi-
The ROS 2 Navigation Stack (Nav2) has emerged as a widely used software component providing the underlying b
Lunar exploration missions require detailed and accurate planning to ensure their safety. Remote sensing data, suc
Abstract— In the realm of space exploration, floating platforms play a crucial role in scientific investigations and te
Maneuverable tether-net systems launched from an unmanned spacecraft offer a promising solution for the active r
This work presents an Online Supervised Training (OST) method to enable robust vision-based navigation about a
In this research, we introduce a deep reinforcement learning-based control approach to address the intricate challe
In recent years, there has been a growing demand for improved autonomy for in-orbit operations such as rendezvo
In this paper, we present a local, robot-centric navigation map optimized for autonomous mobile robots operating in
Rapid growth in the number of nanosatellite deployments has heightened the need for rapid, cost-effective, and acc
Space manipulator system (SMS) maneuvers can excite flexible appendages, while fuel sloshing effects impact its
 The construction of large-scale space facilities requires the use of on-orbit construction technology. However, sever

Spherical robots are a different type of mobility platform. A spherical robot is self-contained within its shell rather tha
Exploration missions on extra terrestrial celestial bodies are to date performed by complex and heavy robotic system
Gecko-inspired adhesives have the advantage of being able to grasp and release flat surfaces in a vacuum using th
In recent years, many researchers have conducted extensive research on spherical robots due to their high flexibilit
To accurately accomplish on-orbit tasks using Space Manipulator Systems (SMS), advanced model-based controlle
In mini-invasive surgery, the laparoscopic ultrasound probe is visible in the laparoscopic image. We address the pro
Robotic manipulation of 3D soft objects remains challenging in both the industrial and medical fields. Various metho
The da Vinci surgical robot introduced remote control of instruments, providing surgeons with increased dexterity an
Real-time recognition and prediction of surgical activities are fundamental to advancing safety and autonomy in rob
We introduce a novel approach for learning a complex multi-stage needle pick-and-place manipulation task for surg
Surgical tasks such as tissue retraction, tissue exposure, and needle suturing remain challenging in autonomous su
Automation in surgical robotics has the potential to improve patient safety and surgical efficiency, but it is difficult to
Recent advances in robot-assisted surgery have resulted in progressively more precise, efficient, and minimally inv
Surgical robot task automation has been a promising research topic for improving surgical efficiency and quality. Le
Bronchoscopy, as an essential minimally invasive diagnostic and therapeutic modality, assumes a pivotal role in the
Controlling robotic manipulators via visual feedback requires a known coordinate frame transformation between the
Soft tissue tracking is crucial for computer-assisted interventions. Existing approaches mainly rely on extracting disc
Accurate deformable object manipulation (DOM) is essential for achieving autonomy in robotic surgery, where soft t
Robotic surgical subtask automation has the potential to reduce the per-patient workload of human surgeons. There
l Robotics II
Fine-grained activity recognition enables explainable analysis of procedures for skill assessment, autonomy, and er
Cataract is the leading cause of blindness worldwide with an increasing number of patients due to changing demog
Physics-based simulations have accelerated progress in robot learning for driving, manipulation, and locomotion. Y
This paper reports on a new real-time robot-centered 3D-2D vascular image alignment algorithm, which is robust to
Concentric Tube Robots (CTRs) are promising for minimally invasive interventions due to their miniature diameter,
Concentric push-pull robots (CPPR) operate through the mechanical interactions of concentrically nested, laser-cut
In minimally invasive procedures such as biopsies and prostate cancer brachytherapy, accurate needle placement
Magnetically steerable robotic catheters (MSRC) are a promising technology for percutaneous endovascular interve
Aortic valve surgery is the preferred procedure for replacing a damaged valve with an artificial one. The ValveTech
Puncture robots pave a new way for stable, accurate and safe percutaneous liver tumor puncture operation. Howev
We present distributed distance-based control (DDC), a novel approach for controlling a multi-agent system, such t
In this paper, we present a swarm robotics control and coordination approach that can be used for locating a movin
Research with swarm robotics systems can be complicated, time-consuming, and often expensive in terms of space
Various robot platforms have been developed for investigating new algorithms for swarm robotics. Morphogenetic e
FireAntV3 uses a refined version of the 3D Continuous Docks to attach to other such docks at any location at any o
Natural and robotic swarms often exhibit non-reciprocal interactions; agents do not exhibit equal and opposite force
Automatic design is a promising approach to realizing robot swarms. Given a mission to be performed by the swarm
In this work we develop a software-in-the-loop simulator platform for Crazyflie nano quadrotor drone fleets. One of t
We aim to address key challenges in long-horizon embodied exploration and navigation by proposing a long-horizo
 Task and Motion Planning (TAMP) algorithms solve long-horizon robotics tasks by integrating task planning with mo
In this work, we are presenting an extension of the cooperative dual-task space (CDTS) in conformal geometric alg
Many real-world sequential manipulation tasks involve a combination of discrete symbolic search and continuous m
In this paper, we present a simultaneous exploration and object search framework for the application of autonomou
Solving a Task-and-Motion Planning (TAMP) problem can be represented as a sequential (meta-) decision process
Given the patchy nature of gas plumes and the slow response of conventional gas sensors, the use of mobile robot
This paper presents an optimization-based solution to task and motion planning (TAMP) on mobile manipulators. Lo
We consider a set of challenging sequential manipulation puzzles, where an agent has to interact with multiple mov
In this paper, we propose a robot oriented knowledge representation system based on the use of the Prolog langua
Embodied Instruction Following (EIF) is a task of planning a long sequence of sub-goals given high-level natural lan
We present the integration of a Non-Axiomatic Reasoning System (NARS) with mobile robots for planning and deci
This paper explores general multi-robot task and motion planning, where multiple robots in close proximity manipula
This paper addresses the challenge of planning a sequence of tasks to be performed by multiple robots while minim
d Motion Planning II
Solving complex robot manipulation tasks requires a Task and Motion Planner (TAMP) that searches for a sequenc
In multi-robot informative path planning the problem is to find a route for each robot in a team to visit a set of locatio
A mobile manipulator often finds itself in an application where it needs to take a close-up view before performing a m
We introduce RAMP, an open-source robotics benchmark inspired by real-world industrial assembly tasks. RAMP c
The use of pointed or edged tools or objects is one of the most challenging aspects of today's application of physica
Long-horizon task planning is important for robot autonomy, especially as a subroutine for frameworks such as Inte
Following work on joint object-action representations, functional object-oriented networks (FOON) were introduced
Task planning for robotic cooking involves generating a sequence of actions for a robot to prepare a meal successf
This paper presents a novel stepwise multi-agent task planning method that incorporates neighborhood search to a
With the increasing demand for multi-robot exploration of unknown environments, how to accomplish this problem e
The teleoperation of complex, kinematically redundant robots with loco-manipulation capabilities represents a challe
When providing task demonstrations to a remote robot over the network via bilateral teleoperation, communication
This paper presents a bilateral teleoperation system based on smart servos for the realization of dexterous manipul
Teleoperation can be very difficult due to limited perception, high communication latency, and limited degrees of fre
Teleoperation has widely contributed to many applications. Consequently, the design of intuitive and ergonomic con
In this paper, we propose a model-mediated teleoperation (MMT) system using an octree-based model (OBM) to sp
A drawing robot avatar is a robotic system that allows for telepresence-based drawing, enabling users to remotely c
Steel mill is one of the most extreme and hazardous working environments due to molten iron erupted from blast fu
Assistive robots promise to be of great help to wheelchair users with motor impairments, for example for activities o
Teleoperating robotic arms can be a challenging task for non-experts, particularly when using complex control devic
We present a tele-operation control framework that (i) enhances the upper motion synchrony between a user and a
Pick-and-place is one of the fundamental tasks in robotics research. However, the attention has been mostly focuse
Redundant robots have the potential to perform internal joints motion without modifying the pose of the end-effector
This paper introduces iRoCo (intuitive Robot Control) - a framework for ubiquitous human-robot collaboration using
Teleoperated avatar robots allow people to transport their manipulation skills to environments that may be difficult o
In human-robot collaboration, shared control presents an opportunity to teleoperate robotic manipulation to improve
Humanoid robots have the potential to help human workers by realizing physically demanding manipulation tasks su
Manually teleoperating a flying robot can be a demanding task, especially for users with limited levels of experience
Point clouds play an important role in 3D analysis, which has broad applications in robotics and autonomous driving
Multi-modal unsupervised domain adaptation (MM-UDA) for 3D semantic segmentation is a practical solution to em
Training a robotic policy from scratch using deep reinforcement learning methods can be prohibitively expensive du
This paper presents a parameter-efficient prompt tuning method, named PPT, to adapt a large multi-modal model fo
Vision-centric bird-eye-view (BEV) perception has shown promising potential in autonomous driving. Recent works
We address the problem of robotic grasping of known and unknown objects using implicit behavior cloning. We train
Training model-free intelligent agents in complex real-world scenarios using reinforcement learning (RL) often nece
Deep reinforcement learning algorithms offer a promising method for industrial robots to tackle unstructured and co
The focus of this topic is to use robotic technology to promote inclusive dance. Inclusive dance aims to break down
Abstract— In this video, we describe a Double Arched Robotic Foot-1(DARFT-1) for a humanoid robot. To protect
A novel robotic module has been proposed based on the structure and motion pattern of Rubik's Snake unit. Each m
This video presents a negative pressure adsorption climbing robot with independent steer wheel which is designed
This video presents an essential robotic motion in assembly tasks, inserting a sash-like plate into the frame. The ins
We have developed a dual-arm manipulator integrated into a cooking robot system. This system is a departure from
We introduce Mono Wi-Fi Scene, the first wireless sensing technology, and a potential sensor module for smart citie
Our planet struggles to maintain its beauty amid escalating environmental challenges. Environmental pollution and
The increasing use of collaborative robots (cobots) has provided effective solutions for various applications; howeve
We present a universal 3D human recovery real-time framework in moving robots, stationary monitoring, and sports
Deep learning, as a prominent research area, has witnessed remarkable research growth, resulting in diverse archi
DLR's mobile humanoid Agile Justin is an advanced mechatronic system with its beginnings dating back as early as
We present an architecture for ad hoc teamwork, i.e., to enable an agent to collaborate with other agents "on the fly
Indoor Positioning System (IPS) is an emerging tech- nology that is widely applied in indoor navigation, urgent evac
We have created an open-source underwater robot named "Easys". This video introduces the hardware, vision, and
Biologging methods have been utilized to study marine life's ecology and behavior. However, this procedure is critic
Able to grasp objects of any shape and size, universal grippers using variable stiffness phenomenon such as granu
In response to the rising number of persons with dementia, requests for novel forms of care that promote the overa
Cardiovascular diseases (CVD) are a leading cause of impaired quality of life and substantial healthcare costs. App
Mobile manipulation robots that can be used as service robots are becoming more popular, but they still are expens
Improving robots' performance relies on efficiently planning their trajectories. One way to do this is by using the Lea
Buoyancy Assisted Lightweight Legged Unit (BALLU) is a human size bipedal robot offering intrinsic stability, safety
In the not-so-distant future, space agencies and private enterprises will embark on resource-harvesting missions in
Modern powered prosthetic hands are equipped with advanced algorithms that can predict high-level movement int
The project consists of creating a digital twin of a robot cell. The cell includes dynamic safety zones that are project
EEWOC (Extended-reach Enhanced Wheeled Orb for Climbing) is a small, lightweight, mobile robot designed to cli
The current work is driven by the shifting demographics in Germany and aims to address the challenges faced by th
Soft robotic gripper systems that can safely collect deep-sea biological samples, nondestructively collect artifact sam
Tensegrity robots, combining rigid struts with compliant tendons, excel in navigating complex environments, endurin
This video introduces two types of flexible devices that generate tip growing motion. The operation of these actuato
Sperm whales can control its buoyancy by changing the solid-liquid phase transition state of the spermaceti in its he
The development of surgical robots is revolutionizing the paradigm of surgery, but many complex procedures remai
Supernumerary robotic limbs (SRL) are new types of wearable robots used as the third limb to work with humans. T
Robotic exoskeletons are a promising technology for improving the physical capabilities of humans and impedance
The supernumerary robotic limb (SRL) is a new type of wearable robotic arm. It can provide the human with an extr
Planning in learned latent spaces helps to decrease the dimensionality of raw observations. In this work, we propos
Event cameras are recent sensors that measure intensity changes in each pixel asynchronously. It is being used du
Non-rigid structure-from-motion reconstructs the time-varying 3D shape of a deforming object from 2D point corresp
Robot navigation within complex environments requires precise state estimation and localization to ensure robust a
We propose a new visual hierarchical representation paradigm for multi-object tracking. It is more effective to discri
The problem of multi-object tracking (MOT) consists in detecting and tracking all the objects in a video sequence wh
In this letter, we perform a tightly-coupled fusion of a monocular camera, a 6-DoF IMU, and multiple position-unkno
We propose an accurate and robust initialization approach for stereo visual-inertial SLAM systems. Unlike the curre
Dense, volumetric maps are essential to enable robot navigation and interaction with the environment. To achieve l
In the recent progress in embodied navigation and sim-to-robot transfer, modular policies have emerged as a de fac
Robot navigation requires an autonomy pipeline that is robust to environmental changes and effective in varying co
In this paper, we present LOC-ZSON, a novel Language-driven Object-Centric image representation for object navi
For people with limited mobility, navigating in cluttered indoor environments is challenging. In this work, we propose
In this paper, we investigate a hybrid scheme that combines nonlinear model predictive control (MPC) and model-b
Metric occupancy maps are widely used in autonomous robot navigation systems. However, when a robot is deploy
In the field of robotics, tag systems play an important role in various applications, such as object identification and r
Extracting motion information from videos with optical flow estimation is vital in multiple practical robot applications.
In outdoor environments, Vision-and-Language Navigation (VLN) requires an agent to rely on multi-modal cues from
Recent results suggest that splitting topological navigation into robot-independent and robot-specific components im
Object-goal navigation is a challenging task that requires guiding an agent to specific objects based on first-person
In this work, we focus on object search tasks within unexplored environments. We introduce a framework centered
Zero-shot object navigation is a challenging task for home-assistance robots. This task emphasizes visual groundin
LiDAR place recognition approaches based on deep learning suffer from significant performance degradation when
Real-time path planning in outdoor environments still challenges modern robotic systems due to differences in terra
Autonomous navigation in the underwater environment is challenging due to limited visibility, dynamic changes, and
Autonomous mobility tasks such as last-mile delivery require reasoning about operator-indicated preferences over t
Low-light enhancement and deblurring is vital for high-level vision-related nighttime tasks. Most existing cascade an
Towards flexible object-centric visual perception, we propose a one-shot instance-aware object keypoint (OKP) extr
3D occupancy prediction holds significant promise in the fields of robot perception and autonomous driving, which q
There are increasing interests of studying the video-to-depth (V2D) problem with machine learning techniques. Whi
Dense visual correspondence plays a vital role in robotic perception. This work focuses on establishing the dense c
Robotic vision applications often necessitate a wide range of visual perception tasks, such as object detection, segm
This paper introduces a novel approach named CrossVideo, which aims to enhance self-supervised cross-modal co
Predicting accurate depth with monocular images is important for low-cost robotic applications and autonomous driv
Dynamic Vision Sensor (DVS)-based solutions have recently garnered significant interest across various computer
Cluttered bin-picking environments are challenging for pose estimation models. Despite the impressive progress en
Depth estimation is a crucial task in computer vision, applicable to various domains such as 3D reconstruction, robo
In the realm of automated robotic surgery and computer-assisted interventions, understanding robotic surgical activ
 In this work, we address the challenging task of 3D object recognition without the reliance on real-world 3D labeled
Keypoint detection and tracking in traditional image frames are often compromised by image quality issues such as
Videos capture the evolution of continuous dynamical systems over time in the form of discrete image sequences. R
Thermal images excel at distinguishing objects from the background in low-light or nighttime conditions due to therm
The image-based 3D object detection task expects that the predicted 3D bounding box has a “tightness” pro
For the task of simultaneous monocular depth and visual odometry estimation, we propose two-step learning of self
We explore Bird’s-Eye View (BEV) generation, converting a BEV map into its corresponding multi-view street im
Recent studies have highlighted the promising application of NeRF in autonomous driving contexts. However, the c
3D Neural implicit representations play a significant component in many robotic applications. However, reconstructi
The SLAM system plays a pivotal role in robotic mapping and localization, leveraging various sensor technologies t
In various applications, such as robotic navigation and remote visual assistance, expanding the field of view (FOV)
Accurate pose estimation over $SE(3)$ is fundamentally crucial for numerous perception tasks, including camera re
Diffusion models have been adopted for behavioral cloning in a sequence modeling fashion, benefiting from their ex
Visual imitation learning has achieved impressive progress in learning unimanual manipulation tasks from a small s
Accurate understanding of lunar surface topography is vital for effective decision-making and remote control of luna
UAV-based intelligent data acquisition for 3D reconstruction and monitoring of infrastructure has experienced an inc
Fiducial markers, like AprilTag and ArUco, are extensively utilized in robotics applications within industrial environm
A consistent spatial-temporal coordination across multiple agents is fundamental for collaborative perception, which
When capturing images through the glass during rainy or snowy weather conditions, the resulting images often con
Perception and Learning I
Given the image collection of an object, we aim at building a real-time image-based pose estimation method, which
Recognizing human action is one of the most critical factors in the visual perception of robots. Specifically, skeleton
Depth perception is crucial for a wide range of robotic applications. Multi-frame self-supervised depth estimation me
Dehazing and enhancing visibility in real-world hazy images pose significant challenges due to the physical comple
Visual bird's eye view (BEV) semantic segmentation helps autonomous vehicles understand the surrounding enviro
In the era of big data and large models, automatic annotating functions for multi-modal data are of great significanc
Semantic segmentation, a fundamental visual task ubiquitously employed in sectors ranging from transportation an
Robots operating in low-light conditions with conventional cameras face significant challenges due to the low signal
Neuromorphic spike data, an upcoming modality with high temporal resolution, has shown promising potential in au
In this paper, we present a novel point generation model, referred to as Pillar-based Point Generation Network (Pilla
The photogrammetry system has been widely used in industrial manufacturing applications, such as high-precision
Scoliosis diagnosis and assessment depend largely on the measurement of the Cobb angle in spine X-ray images.
We present and discuss the Synset Boulevard dataset, designed for the task of surveillance-nature vehicle make an
Pose estimation is an essential technology for industrial robots to perform precise gripping and assembly. The state
Acoustic techniques have been developed as multifunctional tools for various microscale manipulations. In prevalen
Random bin picking is a crucial task in logistic centers, which is driven by E-Commerce growth. In this paper, we pr
In industry, random bin picking is a complex and difficult task where instance segmentation and object pose estima
Understanding actions of other agents increases the efficiency of autonomous mobile robots (AMRs) since they enc
Unlike in natural images in endoscopy there is no clear notion of an up-right camera orientation. Endoscopic videos
Visual question answering (VQA) can be fundamentally crucial for promoting robotic-assisted surgical education. In
Internal limiting membrane (ILM) peeling is a vital vitreoretinal surgery procedure. However, due to the thickness of
 This paper presents an image-based visual servoing scheme that can control robotic manipulators in 3D space usin
This paper presents a novel method to control an underactuated hand by using only a monocular camera, not using
Learning-based manipulation policies from image inputs often show weak task transfer capabilities. In contrast, visu
Visual Servoing is an established approach for controlling robots using visual feedback. Most controllers in this dom
This paper presents a novel approach for tracking deformable contour targets using Unmanned Aerial Vehicles (UA
In this paper, we propose a general unified tracking-servoing approach for controlling the shape of elastic deformab
Existing nighttime unmanned aerial vehicle (UAV) trackers follow an “Enhance-then-Track” architecture - firs
Visual control enables quadrotors to adaptively navigate using real-time sensory data, bridging perception with actio
Multi-object tracking (MOT) on static platforms, such as by surveillance cameras, has achieved significant progress
Object permanence, which refers to the concept that objects continue to exist even when they are no longer perceiv
Object tracking is central to robot perception and scene understanding, allowing robots to parse a video stream in te
Point cloud-based 3D object tracking is an important task in autonomous driving. Though great advances regarding
Given the difficulty of manually annotating motion in video, the current best motion estimation methods are trained w
Modern robotic systems are required to operate in dense dynamic environments, requiring highly accurate real-time
Multi-object tracking (MOT) methods have seen a significant boost in performance recently, due to strong interest fr
To efficiently achieve complex humanoid loco-manipulation tasks in industrial contexts, we propose a combined vis
The recent advancements in transformer-based visual trackers have led to significant progress, attributed to their st
In agricultural open fields, accurate autonomous localization of robots requires long-term data correlation to reduce
Wheeled mobile robots need the ability to estimate their motion and the effect of their control actions for navigation
In this letter, we present VI-HSO, a hybrid sparse monocular visual-inertial odometry system based on two innovativ
While Global Navigation Satellite System (GNSS) is often used to provide global positioning if available, its intermitt
We propose a real-time visual-inertial dense SLAM system that utilizes the online data streams from back-to-back d
Focal-Plane Sensor-Processor Arrays (FPSP)s are an emerging technology that can execute vision algorithms dire
Robust state estimation in challenge situations is still an unsolved problem, especially achieving onboard pose feed
In this paper, we introduce JacobiGPU, a technique that uses a GPU to improve the efficiency of loop closure in vis
We present a novel optimization-based Visual-Inertial SLAM system designed for multiple partially overlapped came
Outdoor sports pose a challenge for people with impaired vision. The demand for higher-speed mobility inspired us
Wearable robotic limbs support sit-to-stand (STS) transfer with increased stability while maintaining a compact size
Neural control is paramount in maintaining upright stance of a human; however, the associated time delay affects s
While existing literature encompasses exoskeleton-assisted sit-stand tasks, the integration of energy recovery mec
le Robotics I
A typical application of upper-limb exoskeleton robots is deployment in rehabilitation training, helping patients to reg
Low back pain is a widespread issue that affects people worldwide and can lead to serious conditions such as hern
This paper presents an insole-type sensor module with a novel leaf-inspired force-sensitive resistor (FSR) array for
Preference-based optimizing methods have shown their advantages and potential in exploring individual, comfortab
Many individuals experience knee dysfunctions attributed to the natural aging process and degenerative con- dition
Foot-drop (FD) is a post-stroke gait disorder characterized by impaired foot lifting during the swing phase. This pap
This paper introduces a novel funnel-based adaptive L1 fuzzy control strategy for assisting ankle joint movement du

The leg-wheel transformable robot has the advantage of smooth, fast, and power-efficient motion on flat terrain and
The state estimation of legged robots often relies on ground contact detection. However, due to complex mechanism
Robots routinely encounter obstacles and rough terrain, but terrain curvature is seldom included in models for real-t
Recent years have seen a steady rise in the abilities of wheeled-legged balancing robots. Yet, their use is still seve
This paper introduces a new wheel-legged robot and develops motion controllers based on central pattern generato
In this paper, a novel planned trajectory classification method (PTCM) is proposed to evaluate the safety of the car-
A two-wheeled vehicle has a compact structure and high mobility in crowded and complex environments. The bicyc
This work proposes a smooth time-varying controller to address the trajectory tracking problems of car-like mobile r
Quadruped robots are machines intended for challenging and harsh environments. Despite the progress in locomot
In recent years, the field of legged robotics has seen growing interest in enhancing the capabilities of these robots t
Self-recovery is a critical capability for robust, agile robots operating in the real world. Given truly challenging terrain
In this paper, we explain the Morphable StateNet as the StateNet with pseudo-generalized behaviors for robots with
Standing-up is a task that humanoids need to be able to perform in order to be employed in real-world scenarios. T
In this work, we propose a novel approach to represent robot geometry as distance fields (RDF) that extends the pr
In this work, we propose a singularity-robust whole-body control framework that ensures smooth task transitions wh
Controlling the contact force during interactions is an inherent requirement for locomotion and manipulation tasks. C
Teleoperation has emerged as an alternative solution to fully-autonomous systems for achieving human-level capab
 is essential for any autonomous scene text spotting system. However, existing state-of-the-art methods employ pretraining an
ficant attention in 3D vision. However, the geometric variability of point cloud representations across different datasets can cau
operate with the closest individual or that only one person is present. However, in real-life situations, such as shop floor operatio
pose estimate or loop closure. Re-ranking methods propose to take into account all the top retrieval candidates and re-order the
d joint embeddings across four action modalities incorporating scene graph (SG) comprehension. A new training paradigm for A
ed place recognition has gained significant attention. Many methods try to address this problem by extracting and matching glo
manned aerial vehicle (UAV) visual tracking due to their light-weight characteristics. However, CFs are prone to generating low-
there is a demand to recognize diverse objects. The state-of-the-art 6D object pose estimation methods rely on object-specific
omation applications, such as classification automation and action recognition. However, the drawback of transformers is their q
when interacting with human wearers. Accurate control of SEA-produced torques ensures the wearer’s voluntary motion a
and Biarticular Actuation (BA), which introduce elastic elements into the structure, have the potential to reduce the electrical en
otect humans and robots from impact, researchers have developed many different soft robots, which incorporate mechanical sp
hich can provide flexible adaptability for robots, improving the safety of robots interacting with the environment or human. This p
d actuation. While Series Elastic Actuators (SEA) and Variable Stiffness Actuators (VSA) are highly sophisticated, they do not fu
and efficient gearing. We present an experimental study comparing pinwheel and non-pinwheel designs of cycloidal gearing. T
omprehensive requirements in wearable assistive application. The pneumatic actuator combines the origami structure of desig
environments such as MRI, it is developed to be used with two motion control modes: 1) a step-by-step mode with tooth-based
between factory lines are examples of cyclic tasks that require large forces but little to no net mechanical energy. Motor-driven
ting of an unmanned aerial vehicle (UAV) linked to an unmanned ground vehicle (UGV) through a non-taut tether with controlla
operating in complex mountainous areas. However, in these environments, UAVs lose their line-of-sight (LOS) communication
ed aerial vehicles (UAVs) to energy-efficiently and reliably fly. To this end, this study presents perception-and-energy-aware mo
ssociated control algorithms before on-orbit deployment. Although state-of-the-art facilities are capable of reproducing zero-G e
most compelling applications rely on onboard deep learning models for perception despite severe hardware constraints (ie sub
off and landing (VTOL) tailsitter flying wing aircraft. The algorithm differs from existing approaches for fixed-wing trajectory gene
e safety as the top priority, and gradually adapt to it based on fresh experience. We present a framework, namely EVOLVER, t

Their small size makes them agile and safe for operation close to humans and allows them to navigate through narrow spaces.
ons in art, robotics, and digital design. We present a novel approach that bridges the gap between digital and robotic sketching
ure. Despite a number of researches being devoted to developing robotic table tennis systems, most of the works have demand
n policies for playing because model-free or inaccurate policy rollouts might easily damage or wear out fragile instruments. We
over 1,400 students in autonomous navigation since 2014 at the University of Michigan and our collaborating colleges. The MB
propeller downwash. Conventional methods have fallen short of providing adequate 3D force-based models that can be incorpo
nline adaptation in changing environments. In these settings, learning is typically done during an initial (offline) design phase, w
ustrial and security sectors. These systems, however, face significant challenges when implementing multi-object detection due
nce for mobile robotic applications including robot navigation in last-mile delivery and robot-enabled automated supplies in indu
planners. Meanwhile, symbolic planners offer the advantage of interpretability. This paper proposes a new task that bridges the
d neural moving horizon estimation (NeuroMHE), which uses a portable neural network to model the MHE's weightings, has sho
oth nominal and defective surface samples are inconsistent in large-scale industrial manufacturing. The most efficient solution u
efforts focus on imitation learning from demonstrations or reinforcement learning in simulation, however, due to the prohibitive
natural language processing, recent research has commenced investigations into the application of LLMs for complex, long-hor
 s, energy efficiency, and edge processing. It necessitates a machine learning algorithm capable of managing real-world odor in
ole in interactive motion planning for autonomous robots. Although existing methods have improved prediction accuracy, few re
nts in the scene, especially for aiding tasks in autonomous driving. Many networks with large-scale point clouds as input use vox
the complexity of agents and their motions increases. However, many existing methods struggle to account for the intricate dep
otic systems that perform sequential decision making subject to ethical theories in fully observable environments. However, in r
controlling humanoid robots' motion. However, these methods typically create simulation environments and rewards for specifi
hanging terrains is challenging due to the complexity of robot dynamics and interacted environments. Recent advancements in
ns, detecting distribution shift is critical. However, most existing methods for detecting distribution shift are not well-suited to rob
lief Space Planning, is a fundamental problem in Robotics and Artificial Intelligence. Due to an abundance of plausible future un
a sensorimotor controller’s observables. The key idea is the design of policies that can take conformal quantiles as input,
Among the suite of estimation techniques, particle filters have been identified as a powerful solution due to their non-parametric
egarding their security and reliability. Adversarial attacks and coverage-based testing have been extensively studied to assess a
le solutions possibly usable from one’s own domestic environment. In this article, we propose a transformer-based approa
However, these robots must perform their tasks dexterously and safely in shared human environments, especially when handlin
nt capability in embodied AI. In this paper, we present SG-Bot, a novel rearrangement framework that utilizes a coarse-to-fine sc
be effective task planners for a variety of single-robot tasks. The planning performance of LLMs is significantly improved via pr
bservations. For a robot to perform tasks such as "pick and place", understanding both what the objects are and where they ar
rly in the realm of robot learning. We present an automated Foosball table along with its corresponding simulated counterpart, s
s into executable actions, from simple enquote{pick-and-place} to tasks requiring intent recognition and visual reasoning. Inspire
g, planning and write code for robotics tasks. However, most methods rely on pre-existing primitives, which heavily limits their s
ment learning (RL) systems within a multifidelity sim-to-real pipeline, in order to deploy reliable and adaptable RL policies on ph
gns in autonomous robot navigation. While the global planner generates a reference path from the current to goal locations bas
errors. However, VSLAM systems relying on low-level visual features often suffer from the problem of perceptual confusion in r
d and demonstrated through a real example of chairs. Our approach leverages the robot's understanding of object stability, per
se environments. Current robot training pipelines rely on humans to provide kinesthetic demonstrations or to program simulatio
Learning paradigm. Complex long-horizon planning tasks require more context(including instructions and demonstrations) to g
eading strategy, promising seamless integration from perception to control via deep learning. However, existing systems grapp
x, long-horizon tasks described by natural language. Recent advances in large language models (LLMs) have shown promise f
on tasks like mobile pick-and-place (i.e., navigating to an object, picking it, navigating to another location, and placing it). ASC c
e, and retrieve information. However, the conventional practice of transferring all data from working memory (WM) to long-term
dangerous tasks. Nevertheless, the future industry demands innovative solutions to streamline the interaction process between
ent is critical. Unfortunately, existing practices are largely manual and reactive in nature. In this paper, we present THEMIS, a n
orms can achieve online planning without considering dynamics. However, in many envisioned application areas such as aerial
hicles is developed. The optimal trajectories are determined by minimizing the uncertainty in position estimation by Extended K
Vector Field method. It allows for singularity-free vector fields with strong convergence guarantees, usable even for self-intersec
ally, UAVs rely on location sensors such as global navigation satellite system (GNSS) and external cameras to obtain location d
rone navigation. Due to the weight and computation limitations on UAVs, using a large number of sensors e.g. a large amount o
ge-scale swarms of UAVs. Inspired by biological perception utilized by schools of sardines, swarms of bees, and other large gro
y respond to changes and make efficient and safe decisions for robots. Although plenty of sampling-based works have shown
l world robotic systems. Our framework splits the reading task into distinct steps, such that we can detect potential failures at ea
and controller composition. Many existing methods require a closed-form expression that limit applicability to data-driven contro
 ms subject to complex constraints. However, MPC is computationally demanding, and is often impractical to implement on sma
ed into a robotic manipulator for manipulating oocytes. The microfluidic device has the ability to release a single cell with gap ef
ss of robotic systems. Nevertheless, a stringent condition named persistent excitation (PE) needs to be satisfied to ensure para
ous vehicles’ performance in the real world under safety-critical scenarios. Sim-on-wheels runs on a self-driving vehicle ope
mous vehicles. Traditional optimization techniques suffer from the curse of dimensionality and limit the search space to fixed pa
go-vehicle for vehicle pursuit. The car being controlled does not have a pre-defined route; rather, it adapts to follow a target veh
utonomous robots must learn to predict their behaviour, especially when that behavior is potentially dangerous to other agents
en rely on testing in simulation for the creation of reliable systems. A key element of effective simulation is the incorporation of r
ain, LLM agents, for solving robotics and planning tasks by leveraging the world knowledge and general reasoning abilities of LL
ommunication is considered crucial for enhancing traffic safety and efficiency. However, existing methods often simplify the han
ng partial observations presents significant challenges. To address this, we propose a novel trajectory prediction framework call
allenge owing to the ambiguous intentions of dynamic agents and their intricate interactions. Existing studies have attempted to
deling approach, RePed, that facilitates transferring microscopic behavior models from behavior research to higher-level trajecto
ystematic and distributed integration of local behaviors in a group. Agents sharing knowledge about the mission and environmen
n-based algorithm (DMCHBA), to multiagent task-allocation problems, in which the number of agents is smaller than the numbe
The controller decomposes the prescribed reference path, corresponding to a fixed prediction horizon, into multiple shorter path
o-Shot Object Goal Navigation (L-ZSON), where an embodied agent navigates to an uniquely described target object in a prev
environments, this becomes even more demanding. In planetary explorations, the robots are continuously challenged with the
However, handling external disturbances during tasks is still an open challenge. This paper proposes a novel method to achiev
e multi-agent trajectory prediction. Hypergraph Convolutional Networks (HGCN) have become a proper choice for capturing hig
icularly in generalization and interpretability. We introduce a unique object-level multimodal LLM architecture that merges vecto
ffic participants is a fundamental but challenging task due to various factors encompassing the agents' dynamic states, interact
asynchronously operating millimeter wave radar sensors and a combination of gyroscope and accelerometer. Two different m
the field of robotics. Although systems utilizing Light Detection and Ranging (LiDAR) and Radio Detection and Ranging (RADA
ver, due to the absence of a standardized benchmark, the effectiveness of various designs remains unclear. The newly release
plex urban environments. Currently, most autonomous vehicles highly rely on the High-Definition map (HD map) to cruise acros
on about both visible and occluded regions. This involves predicting the future motion of observed agents, inferring occluded on
he cost of sensors and limitations in local computational efficiency make it difficult to scale to large commercial applications. Tra
s. Prediction must be fast, to support multiple requests from a planner exploring a range of possible futures. The generated pred
hallenging for automated vehicles due to the stochastic nature of traffic participants' behaviors and their implicit interaction with
autonomous driving. The lane-changing trajectory planning algorithm for autonomous driving is typically formulated as a optimiz
r planar workspaces into free spaces and obstacles. Deeming obstacles as non- traversable significantly limits wheeled robotsâ
e of robots moving in human-inhabited environments. Previously, roboticists have developed geometric navigation systems with
racking hinges upon various factors, such as tracking distance, occlusions, collision avoidance, etc. However, few existing work
™s-eye view (BEV) maps commonly used in autonomous driving. While existing perception systems for autonomous driving sce
ng enhanced end-to-end autonomous driving performance. In this paper, we introduce a novel multitask imitation learning frame
online learning for real-world robots remains a challenging task. Recent work directly learned from offline dataset to achieve bro
uating and improving autonomous driving software through simulation. Our approach models all driving scenario elements in a
een measured points when a low-resolution LiDAR is given, making a high-resolution LiDAR or even a resolution-free LiDAR. R
entation for an unmanned ground vehicle (UGV) to learn traversability in various scenarios containing irregular obstacles. To pr
critical to enable safe and effective robot navigation. Such spaces remain challenging as humans do not follow strict rules of m
 ement learning (HiL-RL) paradigm mainly operate on the assumption that human policies are homogeneous and deterministic
lators and large-scale driving datasets. While offline reinforcement learning (RL) is well suited for these safety-critical tasks, it s
tal requirement for the successful execution of any movements of ground robots, especially given that terrain traversability ofte
everal advantages including reduced system cost, extended perception range, enhanced data fusion capabilities, and energy s
ant but challenging issue. Using mobile robots to autonomously search and navigate to gas source location provides a promisin
y for autonomous robots. In this paper, we propose a drift-resisting topo-metric graph to map the environment and localize the r
ation. OpenBot-Fleet uses smartphones for sensing, local compute and communication, Google Firebase for secure cloud stor
her-level abstractions such as 3D boxes and polylines. These sparse shapes are inferred through annotating the original scene
ency in manufacturing. However, their low absolute position accuracy still denies the broad substitution of machine tools by ind
s and tasks, bringing the potential of learning policies that generalize in diverse scenarios. Prior works have explored visual pre
gs is at odds with the paucity of robotics datasets. Acquiring and growing such datasets is strenuous due to manual efforts, ope
VLMs) trained on 2D data into a 3D object rearrangement pipeline. This is achieved by the robot autonomously constructing a 3
e range of domains because the use of existing data or pre-trained models on the internet enables quick and easy learning of n
stems. Onboard compute budgets typically limit the accuracy of real-time systems. In this work we propose methods of improvi
ers, seconds). Operations over different units (e.g., adding meters and seconds) may be incorrect and can lead to dangerous sy
Inspired by the takeoff process of ospreys, we propose a bio-inspired takeoff strategy, then discuss the dynamic model and the
onic underwater robots. A novel origami-based adaptive flipper is presented to address a lack of environmental adaptability and
presented. Bird-size ornithopters do not have asymmetric variables in the wing due to mechanical complexity, so they rely upon
otions, such as Cruising-straight, Cruising-turn, and various fast turns, etc. Currently, there is no single mathematical model tha
ly, is complex and challenging. To understand how mud properties and robot-mud interaction strategies affect locomotion perfo
or detection, greater hunting efficiency, more accurate environmental sensing, and energetic saving. Although the benefit
characterized by high efficiency and flexibility. However, estimating the swimming direction of these robotic fish is challenging d
by many fish-inspired robotics. However, it is unknown whether the energy saved from improved efficiency can compensate for
cape from the predator through simple individual interactions. The purpose of this research is to infuse swarm robots with “f
xteroceptive sensing and central pattern generators (CPGs), i.e. systems of coupled oscillators, into the deep reinforcement lea
ubset of a snake robot's configuration space for which the constraints imposed by the obstacles in its environment render the sy
er, making them prone to flipping over and needing to self-right. We present a multi-faceted shell that by its geometry alone ena
ning and jumping. Usually, such behaviors are either optimized and engineered offline (i.e. the behavior is designed for before it
uvrability at high speeds by stabilizing its body. This has inspired the design of several agile robots, including Dima - a wheeled
otion. The spine grants the trunk with increased degrees of freedom, thereby endowing diverse postures. This paper presents
nded in biomimicry principles. Pegasus offers two distinct motion modes, including a wheeled motion and a hybrid wheeled-leg
n of running and jumping. However, replicating this versatile locomotion in a soft robot poses technical and design complexities
ced as well as the high amount of uncertainty in this domain. Because of the difficulty in modeling forces in granular media, we p
d mapping spatially constrained environments. However, achieving untethered flight powered by a small and lightweight on-boa
distance without flipping. To solve this problem, we develop a miniature locust-inspired jumping robot, which has a body length o
energy in resilin elastic cuticular structures and releasing that energy to accelerate a tail, called a furca, propelling them in the a
ace tension to remain above the water surface. A subset of these insects, such as the stonefly and waterlily beetle, flap their w
botic swimmer that is propelled by a bioinspired anguilliform propulsor. The propulsor is excited by a single 6-mg high-work-den
cues from their nest's environment. However, the precise mechanism underlying this remarkable homing skill remains a subject
an perform distinctive flight maneuvers, such as flying backwards, which has proven to be achieved through ”force vectorin
cacies and suitability for various applications leveraging miniaturized size. Electrohydrodynamic (EHD) thrust offers advantages
gence in these simpler models and to endow robots with biological intelligence has been attracting increasing attention in the fie
h as invasive surgery, search and rescue, and inspection. However, their reach is often limited because their extension mecha
a that they can only avoid obstacles by re-taking off when encountering obstacles due to the poor performance of wheeled obs
ody undulations, is frequently observed in rattlesnakes and has been successfully implemented by limbless robotic systems for
y to enhance robots' mobility while airborne or improve robots' safety in falling. The tail, especially with two Degrees of Freedom
e of colors, numerous technical challenges still exist for contemporary 3D printers to fabricate various materials and generate co
narrow and hazardous spaces. However, traditional CDRMs have limited stiffness and load capacity due to their cable routing m
c eye platform in aiding ophthalmologists to better understand the biomechanisms of human eye movement. This letter presen
perations such as manipulation and grasping. Traditional methods for achieving this objective necessitate the careful design of
cently, automatic and target-less extrinsic calibration has become the mainstream of academic research. However, geometric fe
mance of stereo matching algorithms. In prior arts, the online self-calibration of stereo cameras has commonly been formulated
ry due to dynamic forces, thermal dilations, or minor accidents. This paper proposes two methods, OCAMO and LTO, that mon
ented with depth data. However, their limited field of view (FOV) often requires multiple cameras to cover a broader area. In mu
a pivotal concern. Among the myriad of available techniques, target-based calibration methods, which employ planar boards w
s in autonomous driving. This paper proposes a novel three-stage extrinsic calibration method between LiDAR and GNSS/INS
Existing methods either suffer from the complex offline setting of man-made targets or tend to produce suboptimal and unrobu
s precisely. However, it is also a time-consuming process that needs to be repeated for every displacement of the cameras. The
fundamental prerequisite for effective fusion is the precise calibration be- tween LiDAR and camera systems. Many existing ca
d enterprise (SME) manufacturing and production due to the development of increasingly sophisticated Human-Robot Collabora
n order to simulate them accurately. We present a Bayesian optimization approach to identifying the material property paramete
on of 3D environments. To effectively fuse information from both camera and LiDAR, extrinsic calibration is indispensable. Rec
obot to the sensor attached to it. Commonly employed techniques, such as AXXB or AXZB formulations, rely on regression me
n objects. However, such sensors can be difficult to apply to small-sized robots due to their weight, size, and cost. In this paper
r perception tasks. The performance of such fusion-based algorithms largely depends on the accuracy of sensor calibration, wh
Recently, event cameras have emerged as a promising type of neuromorphic sensors, with many potential applications in field
cloud, direct doppler velocity output, long distance ranging, low-cost, and more importantly, robust perception in all weathers. H
over moderate terrains using only proprioceptive sensing. However, such blind controllers will fail in environments where robots
ed exteroceptive perception, or perception failures. Such failures often occur when harsh environments lead to degraded sensi
nts and decide where to explore next is pivotal for developing robots capable of human-like search behaviors. We introduce a z
task-oriented navigation (i.e., reaching a goal that the robot has located), and task-agnostic exploration (i.e., searching for a go
to deep learning architectures such as CNNs, RNNs, GNNs, and Transformers. While this concept has been widely applied in
achine learning algorithms have proven to accelerate collision checking. We propose CollisionGP, a Gaussian process-based a
avoids them. Because human motion cannot be predicted exactly, future trajectories are typically inferred from real-world data
ents a comprehensive approach to collision-free navigation in dynamic indoor environments. The approach integrates multimod
blems where the controlled objects and the obstacles can be modeled as the finite union of convex polyhedra. A novel collision
stacles, including their geometries and trajectories, are fully observable. While it's easy to obtain these states accurately in simu
ogy that ensures the safety and reliability of autonomous driving systems. Conventional collision avoidance approaches struggle
panning tree algorithms to develop a new approach for solving a Multi-Goal Path Finding (MGPF) problem in high dimensional
nknown environment with uneven terrain. In this study, we present a layered and systematic pipeline. At the local level, we mai
dal locomotion, leveraging diffusion-driven techniques. Our contributions include a scalable dataset generator for map images a
s in robotics. The difficulty of the problem arises from its high dimensionality and the intricate geometry of the feasible space.Ou
 oving the performance of mobile robots. Nevertheless, most existing DRL-based navigation methods primarily focus on training
spread mobile robot deployment: predicting uncertain motion of multiple agents, and planning robot motion under uncertainty. H
lans in algebraic reparametrizations of configuration space. Such motion plans include those generated by popular randomized
surveillance systems. This study presents a novel cross view capture distributed image compression network (CVCDIC) to imp
impose time-dependent constraints. Recent planning methods focus primarily on the spatial aspects, lacking the capability to d
ms. The challenging aspect is the analytic representation of collision cost for the case when both the obstacle map and robot foo
on for robot navigation under uncertainty. We present (1) a closed-form approximation of the probability of collision that natural
eption and autonomous navigation. MIMs consist of multiple stacked layers of 2D grid maps each derived from reflected point c
robots. We propose a new geometric-based uneven terrain mapless navigation framework combining a Sparse Gaussian Proc
t's exteroceptive sensors fail, which means it becomes "blind", collision can actually be leveraged to improve localization perform
l (3D) geometries. Conventional algorithms suffer from low accuracy due to their reliance on limited representations, such as po
articularly in scenarios involving large-scale maps with sparse obstacles, as it necessitates a grid-by-grid state check. Specifica
ch as robot navigation around pedestrians motivates the development of additional safety mechanisms. Running RL-enabled sy
ation in complex outdoor environments. We incorporate multimodal sensory inputs from an RGB camera, 3D LiDAR, and the ro
ver, its excessive degrees of freedom pose challenges for path planning. In this article, an ellipsoid-shape rapidly-exporing rand
safety controller for autonomous mobile robots navigating in unknown environments. The main challenges of implementing CB
asks such as visual question answering and image captioning. Consequently, these models are now well-positioned to reason a
nstruction due to their exceptional memory efficiency and requirement for only posed RGB inputs. Previous NeRF-based view p
al in this task is to maximize the information of the object to be reconstructed, given limited on-board resources. Previous view
ormers in large-scale autonomous driving benchmarks. A transformer is a popular network architecture used for this task, in wh
chnologies within the field of robotics. Previous research has predominantly focused on supervised approaches, relying on exte
nts and thus is of great interest in the autonomous driving field. Dynamic capture is always critical in the MOS problem. While pr
ng per-frame depth maps and fusing them in 3D, recent research leverages the neural implicit surface as a unified representatio
uts of transparent objects and grasp them quickly. Conventional input modalities for vision-based grasping do not provide suffic
h dimensionality of cloth models, as well as the limited observation at self-occluded regions. We leverage the recent progress in
In this work, we address these challenges at the task level by introducing Amodal Optical Flow, which integrates optical flow w

hallenges like mis-registration, object occlusions, and non-uniform upsampling. This paper introduces a scene flow estimation fr
ing focus on fine-grained 3D semantic occupancy prediction. Effectively representing detailed three-dimensional scenes has be
the explicit construction of correlation volumes, which are expensive to compute and store, rendering them unsuitable for robo
single robot's perception ability beyond its field of view. Many prior works receive messages broadcast from all collaborators, le
utonomous mobile robots, such as scalable simulation, scene manipulation, and sparse-to-dense completion of LiDAR point clo
ntaneous decision-making and rapid response. In real-world scenarios, autonomous vehicles are continuously tasked with inter
ong network with improved performance, as opposed to using separate networks for each task. In this paper, we introduce a ne
es have a wide range of applications. However, kinetostatic analysis and contact sensing for such grippers are still challenging
of rotational slippage has been proven beneficial for improving the dexterity and stability of robotic hands. An effective strategy
tructions of uneven surfaces may only be partially possible due to non-compliant end-effectors and limitations on vision system
formation, maintaining friction limits and safety constraints. However, contact geometry and dynamics are required to be known
sor. The sensor is developed for shape reconstruction under the inspiration of sketch drawing that uses the density of sketch lin
understanding of deformation can elevate robotic simulations and have broad implications across different industries. We introd
s. Pushing is a fundamental skill that enables robots to move unwieldy objects that cannot be grasped. We propose a stable pus
of a 3-D smooth object as it rolls freely on a motion-controlled robot manipulator, or "hand." While specific examples of robotic c
curate contact modeling and grasp stability estimation. One key component is the coefficient of friction, which is typically estima
generalize to new settings. Recent deep reinforcement learning (RL) methods show promise in fully autonomous learning, but
otential visual cases is infeasible in practice. Hence, test-time adaptation (TTA) allows the model to adapt itself to novel environ
ile also preventing it from forgetting its previously acquired knowledge. Streaming lifelong learning is a challenging setting of life
ic forgetting, obviating the need to re-learn all tasks from scratch. In order to enable lifelong acquisition of skills in robots, replay
ontrol due to unknown and unpredictable events that disrupt task performance. Neural network-based reinforcement learning (R
and control, most recently by generating robot policy code that utilizes a custom library of vision and control primitive skills. How
high-level instructions into a sequence of executable instructions. However, current LLM-based planners are only able to opera
works have shown significant promise in advancing robotic tactile learning. However, due to the changing of tactile data distribu
ot to continuously and efficiently learn to solve new manipulation tasks throughout its lifespan. The core idea behind LOTUS is c
r slender shape. However, the introduction of the tilting servo motors leads to a non-minimum phase roll dynamics, which affec
ally sliding pillars integrated with tactile sensors. Functionally, ArrayBot is designed to simultaneously support, perceive, and m
hanging task requirements or environments. Modular robots, on the other hand, offer flexibility and can be easily customized to
ditional methods of cleaning windows, hanging on ropes poses significant occupational hazards to workers. Furthermore, most
nomenal advances in industrial robotics, manipulation of deformable objects remains mostly a manual task. This is because of t
portance that the collected information is efficiently shared among robots in order to support exploration and navigation tasks. P
cles. It’s challenging for a robot to navigate safely without any sensors that can sense the environment, resulting in a blind
ollow changing patterns or shapes in a optimal distributed manner. The optimal flocking control with dynamic pattern formation
n delivery into the subretinal space, emphasizing minimally invasive procedures targeting the retinal pigment epithelium (RPE)-B
e detection and tracking since nobody can exit the field of view. Despite the recent diffusion of low-cost panoramic cameras, the
ailure for reasons such as miscommunication, incorrect actions and unexpected object properties. Existing works on handover
new BOP format dataset for 6-DoF object localization, crucial for robotics. This dataset aims to evaluate the current localization
c labels for complex 3D scenes. Most existing SSC models focus on volumetric representations, which are memory-inefficient f
remains limited, primarily due to challenges in data acquisition and labeling. Previous sim2real approaches using domain rando
umans, ma- chines, robots, and various other objects such as tools, fixtures, conveyors, and assemblies. In these environment
n and Mapping (SLAM) algorithms. Nevertheless, their precision must follow the rate at which SLAM algorithms improved in rec
to real-world applications, which poses an essential challenge in bridging the sim-to-real domain gap. While mainstream works
bots' dynamic performance. To address this, we present a mobile robot oriented large-scale indoor dataset, denoted as THUD
actions accordingly to smoothly execute tasks. However, the human's intent in turn depends on actions the robot takes, creatin
diction in autonomous vehicles. Yet for the equally important planning task, the adoption of ML-based techniques is slow. We p
ased methods and converged on the topic of modeling inter-pedestrian interactions and pedestrian-robot interactions. For this, l
e used for high-level reasoning with long and medium horizons and that has 2.2x higher throughput compared to traditional nar
us vehicles (AVs). As these scenarios are often insufficiently present in naturalistic driving datasets, augmenting the data with
eralizable skills for robots. Recent progress in one-shot imitation learning and robotic foundation models have shown promise in
t go beyond simple predictive models of human behavior. By observing and understanding human interactions from past experi
disciplinary efforts to enhance safety performance. This work focuses on risk identification, the process of identifying and analy
ment in order to react to possible obstacles. Deep learning is a promising technique to detect obstacles robustly. However, supe
tions, surrounding entities, and the ego vehicle's driving behavior for safe and efficient navigation. This is particularly critical in
s and their use in autonomous driving scenarios has become widespread. However, deep learning methods heavily rely on ann
h as tracking, convoying and subsea intervention tasks. 6D pose estimation is a fundamental task that enables precise object l
 the perception and interpretation of the environment. However, the development of deep learning-based techniques for autono
ts are required to handle objects, enabling them to interact with their surroundings. Among the various tasks in robotic grasping
or better localization and navigation. It is challenging due to appearance changes at various frequencies, and difficulties of obta
ch are labor-intensive and time-consuming. We seek to demonstrate that robotics offers improvements over these techniques a
vide large scale 3D point clouds (LS3DPC) that guarantee a certain level of accuracy in common environments. However, they
metric types of primitives, which plays a vital role in robot operation and industrial automation. However, differences in object st
ent differences in modalities between sensors. To deal with this problem, we propose FF-LOGO: a cross-modality point cloud re
omputer vision. In this paper, we propose CAPT: category-level articulation estimation from a point cloud using Transformer. CA
conditions like rain, snow, and fog negatively affect these sensors, reducing their reliability by introducing unwanted noise in the
uch as multi-robot SLAM and collaborative localization. To facilitate fast detection and generate compact descriptions on edge
ased methods have achieved advanced point cloud registration performance. However, the standard attention mechanism utiliz
ntric videos can greatly improve safety and efficiency in human-robot interaction (HRI). While previous research predominantly
continual test-time adaptation (CTTA) has been proposed as a strategy for transferring deployed models to continually changing
cision of manipulations but also contributing to safety, efficiency, and adaptability. Deep learning-based slip detection algorithm
of commonsense knowledge about human--object interaction. We jointly predict human joint motion and human--object interac
can be used for motion planning tasks in the absence of a classical geometric kinematic model. In particular, when the latter is h
ns. Their combination helps the robot gain an enhanced manipulation capability, in which the generated pose can facilitate the
s using an RGB-D camera and validate its effectiveness through robot grasping experiments. Unlike existing methods, which fo
ons. Our goal is to develop methods that can robustly detect and track natural features on robotic manipulators. Such features
ions lead to a significant lack of data in adequately learning UAV-based perception models. Using various synthetic renderers in
ging problem. The utility of a robot greatly depends on its ability to perform a task in the real world, outside of a well-controlled la
enarios, including leak identification and locating thermal updrafts for extending UAV flight duration. Unfortunately, the heat sign
s. We address this task in the context of miniaturized nano-drones, i.e., 10cm in diameter, which show an ever-growing interes
fectively in unstructured 3D environments. In order to address the challenges posed by the high demand for large-scale annota
ulture. In this work we present PAg-NeRF which is a novel NeRF-based system that enables 3D panoptic scene understanding
es a photometric loss. While recent methods have made strides in indoor depth estimation, they often produce inconsistent dep
ions like robot collaboration and online hand-eye calibration. However, the introduction of unknown joint angles makes predictio
echniques into vehicle perception and decision-making. However, their application in 3D lane detection for effective driving env
rmation is extremely needed by autonomous vehicles, which helps to make the following decision-making and path planning m
significant challenge as it requires seamless integration of end-effector trajectory generation with navigation skills as well as re
Maps (SOGM), which embed future semantic information of real dynamic scenes. We present an auto-labeling process that cre
us, learning and extracting more compact representations from camera observations is a promising avenue. However, current m
mous vehicles to anticipate the motion of surrounding agents. Goal-based prediction models have gained traction in recent yea
ks, such as joint compliance, demand high-quality joint torque sensing. While external torque sensors are reliable, they come w
nguage, vision, and sequential decision-making problems. To create such models, we generally require centralized training obj
obotics. Recent work utilizing prominent deep learning models for pedestrian motion prediction makes limited a priori assumptio
dies apply Multi-Agent Reinforcement Learning (MARL) to solve MAPF in partially observable settings. The objective of coopera
wer intelligent robots with complex reasoning. Existing work focuses on harnessing their abilities to reason about the histories o
handles highly uncertain and complex scenarios involving moving agents of different types. In this paper, we propose a Multi-G
botic perception and classification of fabric textures. We formalize the active sampling problem in the context of tactile fabric re
putation and improve the state-of-the-art of AI systems. Bio-electrical synapses directly transmit neural signals, by enabling fas
ey often lack accuracy. These methods attempt to improve accuracy by introducing new modules or integrating traditional metho
autonomous driving. When building large-scale scenes, limitations in computing resources and the difficulty of accessing the en
assifying object surfaces. We train a Diffusion Model to bridge this gap using a relatively small dataset of real-world images rand
rstanding of human social dynamics. In this work, we propose to model social motion forecasting in a shared human-robot repr
multimedia continues to grow. Point clouds are widely adopted within this domain because they are one of the most intuitive re
anning problem, which finds valid trajectories up to 129x faster than current methods, while also succeeding on more difficult pr
y adapt in dynamic and uncertain environments. Classic adaptive control and safe control provide stability and safety guarantee
computationally demanding, which limits their application. Especially when considering sampling-based path planning, the mod
s based on optimization or sampling the state space, suitable for solving kinematic motion planning. However, when dynamic m
s ranging from managing warehouses to cleaning and organizing home kitchens. While existing research has primarily focused
RL) algorithms. In action-constrained RL, each action taken by the learning system must comply with certain constraints. These
e exploratory grasp learning during online adaptation to new picking scenarios, i.e., objects that are unseen or out-of-domain (O
he dynamic grasp learning that occurs during real-time adaptation to novel picking scenarios. These scenarios may involve pre
n of the camera’s pose with respect to a robot. Conventional methods for camera-to- robot pose calibration rely on fiducial m
Results (GANDER), leveraging Generative Adversarial Networks to provide online error detection in manipulation tasks for auto
ask. These systems have limited generalization capabilities to new tasks, but have shown the ability to generalize to novel objec
applications in manufacturing, maintenance, and recycling. Unlike existing research, which is limited to target segmentation, po
ertain everyday objects remain challenging for robots to effectively interact with. This discrepancy arises from the fact that huma
learning in the context of object assembly tasks, particularly the alignment and insertion of objects with geometrical extrusions,
r homogeneous settings where the robot is required to either manipulate rigid or deformable (soft) objects. However, the manip
n capabilities and performance of visual policies. However, using multiple cameras in real-world scenarios can be challenging. I
ations of policy learning to robotic manipulation typically only leverage visual feedback. Recently, symmetric neural models have
rk reliably in realistic environments. We investigate the problem of encoding object-oriented memory into a multi-object manipu
objects or perform separation manipulation when the bin contains no isolated objects. The robot must be able to flexibly perform
e and complex dynamics. Learning approaches stand out as promising for this domain as they allow us to learn behaviours dire
r and robot control, making it a task with hybrid output nature. Existing methods based on closed-loop deep reinforcement learn
e to unbalanced synergy and redundancy between both actions, leading to decreased grasp efficiency. In this paper, a novel do
romising approach to deformable object manipulation. An issue arises, however, with the reliance on the specification of a goal
uestion is difficult to approach at face value, but the environment from the perspective of a robot can often be decomposed into
ction lines, with a focus on but not limited to transparent objects, which is an important and challenging problem in robotic syste
new useful behaviors quickly. Current approaches lack either the generality to onboard new tasks without task-specific enginee
s are occluded, e.g., by the environment. Single-shot grasp planning invariably fails in such scenarios. Instead, it is necessary to
eep Reinforcement Learning) to achieve target shapes. Our proposed approach uses visual observations consisting of segmen
nipulating 3D articulated objects is challenging because they have diverse geometrical shapes, semantic categories, and kineti
r in household use-cases. In recent years, machine learning methods for the prediction of 6-DoF grasps on diverse and unknow
ng (RL) algorithms due to the inefficiency in collecting successful experiences. Recent algorithms such as Hindsight Experience
e holds importance, yet it remains unsolved within the field of robotics due to failure of recover their accurate geometry by depth
-based robotic manipulation tasks and utilize certain task properties to achieve offline learning with a small dataset. We propos
cant challenge due to its inherent complexity and constantly changing nature. In this context, RGB image and point-cloud obser
task for real robots. Existing learning-based solutions mainly focus on visual affordance learning or other pre-trained visual mod
ntrol policies in simulation using reinforcement learning. Like most co-optimization methods, MORPH relies on a model of the h
 easible for learning on real robots, due to the amount of data needed. However, if done right, agents learning from real data can
on, can aid robotic manipulation. However, learning about affordances requires expensive large annotated datasets of interacti
erent requirements and needs, there has been a lack of focus on enhancing the adaptability of robotic grasping systems. Most
orks resort to the whole observed point cloud for 6-Dof grasp generation, ignoring the guidance information excavated from glob
replacing traditional controller with differentiable neural networks, but lose the ability to servo arbitrary desired poses. This lette
Learning (DRL) that is capable of directly synthesizing continuous 6-DoF actions in cartesian space. Our proposed approach u
The goal is to predict future point cloud sequences that maintain object structures while accurately representing their temporal
work model using convolutional neural networks (CNNs) and enforce consistency of the model’s predictions over perturbat
integrity of critical infrastructure. Such visual inspections sometimes happen under low-light conditions, e.g., checking for bridg
hich has the potential to be widely applied and expanded in the fields of robotics and mobile devices. By compressing the weigh
al conditions, making it a promising solution for all-weather perception tasks, such as outdoor mobile robotics. However, the rad
stic and visual information must be processed jointly and bound together. Existing works either encode the image or the languag
ears. Visual descriptors can easily describe manipulation task objectives, they can be learned efficiently using self-supervision,
n. However, existing 3D dense reconstruction systems are limited by the requirement for pre-training, and there is a demand for
their potential for impressive performance, obtaining real-world training data is expensive, and can be impractically difficult for
ct-rich, and dexterous manipulation that humans excel at. However, relying solely on tactile sensing fails to provide adequate c
environmental fixtures has diverse industrial applications. In order to establish robust contacts with these fixtures, accurate estim
erence from surrounding items. We propose a Deep Reinforcement Learning (DRL) method that develops joint policies for gras
nvironments. One of the core problems in tactile servoing is to robustly converge the contact features to the desired ones via a
and prehensile actions such as grasping and throwing, within the domain of robotic manipulation. We introduce an innovative ap
ncluding the required amount of prior knowledge of the system and the sophisticated physics that varies across different robot
ons learning from large human manipulation data can generalize well to new manipulation tasks and environments. However, th
robots. In this paper, we present the design of feedback skills for robots that must learn to stack complex-shaped objects on top
hand in terms of dexterity and versatility. A milestone in this quest for human-level performance will be possessing the skills of
. It employs probabilistic graphical models to effectively capture helical periodic features in the latent space, integrating Bayesia
In our previous study, we proposed an insertion skill framework based on force-domain wiggle motion. The main limitation of th
deformable linear object (DLO) manipulation has primarily involved parallel jaw gripper manipulation with fixed grasping positio
rming a large number of everyday tasks. To match these requirements, we present a novel actuator and finger design, the eve
of cooperative redundant manipulators, a hierarchical method using null space and an optimization method prioritizing the end-e
odel-free controller using novel 3D global deformation features based on modal analysis. Unlike most existing controllers using
much to be understood from a model-based perspective, where key difficulties are often attributed to (i) the explosion of contact
ntricate fine motions, and ever-changing scenarios. We present a hierarchical planning framework for dexterous robotic manipu
cant challenge. This letter introduces a novel approach that utilizes a tactile sensor-equipped gripper with eight degrees of freed
sical links to self-assemble and self-disassemble in midair. For the changes in task or environment, this system can dynamically
et tracking and is challenging for Cloud or Fog robotics applications due to network congestion and failures. We introduce FogR
s and facilitating power transfer, to use by the robots as a net or partition. In this paper, we show in simulation that tethers can
e computational efficiency and scalability of Differential Dynamic Programming (DDP) and the distributed nature of the Alternati
daptive coalition strategy is essential for the online performance in dynamic and unknown environments. In this work, the proble
However, the challenge of optimizing task efficiency without compromising task success still persists, particularly in expansive, u
arise in the probabilistic inference of large-scale networked robotic systems. First, for the batch DFGO problem, we derive the
allocation algorithms (i.e., CBAA, CBBA, HIPC, PI, DHBA, and DGA) subject to non-ideal communication factors. The package
of freedom, it is challenging for them to quickly respond to the attitude requirements during underwater autonomous maneuveri
, tanks, and pipelines have flourished in recent years. Most work has focused on the mechanical feasibility of the morphing sys
specifications. This problem is particularly challenging when robots navigate in continuous workspaces due to the high comput
fications described by signal temporal logic (STL) in unknown, stochastic environments. While many existing works concentrate
highly accurate and efficient control systems. Conventional methods for dynamic model identification have proven to be effectiv
tems consisting of rigid bodies connected in arbitrary topologies by localised joints and/or soft links, possibly actuated or not. Th
linear quadratic regulator (LQR) problem using Gauss' principle of least constraint and solving it using dynamic programming.
parts. In co-simulation, the subsystems only exchange information at communication points. The time delay of information exc
ex and dynamic nature of robot terrain interactions. Existing approaches often struggle to generalize amidst the complexities of
assembly tasks frequently involve tight insertions where the clearance is less than 0.1 mm and can even be negative when dea
tic peg-in-hole in uncertain environments. In contrast to past research in this field, in which force control with relatively low com
novel method for achieving high precision 6D pose estimation by exploiting the reprojection of 3D edges onto binocular RGB im
a physically feasible sequence of actions for assembling many parts together. In this paper, we present ASAP, a physics-base
the generation of roadmaps on a given layout and the coordination of multiple Automated Guided Vehicles (AGVs). A simulatio
g-term tasks (e.g., map updating, navigation, collaboration). However, it is challenging to implement multi-session mapping in e
ns about how hostile events, from which an environment must be protected, are generated at a specific time and location. Relyi
ntext of autonomous surveillance of valuable assets. Optimizing patrolling strategies often aims at bounding the time between s
n multiple agents share a common task e.g., sorting different materials in a recycling facility. Traditionally, each agent in a sorti
copters is on the rise. This study focuses on controlling a bi-copter under rotor or servo stuck failure. A relaxed hover solution is
to exploit intentional collisions with the surrounding environment and objects, an approach referred to as impact-aware manipu
s and the quality of production. When collecting data to detect such cases, it is not ensured that data from all potentially occurri
stly without requiring human assistance. Unstructured environments are challenging for robots since the exact prediction of out
gerous conditions. Consequently, diverse severe hardware failures may occur at any time during the working cycle of the robot
uption. This work focuses on developing a method to effectively utilize a malfunctioning printer for reliable printing. This method
. In particular, the quadrotor is now invaluable across a plethora of fields and recent developments, such as the inclusion of aer
safe driving properties. Specifying and monitoring such properties, however, is challenging because of the mismatch between
fore efficiently handling contingencies is an important capability in the context of human-robot teams working on assembly (i.e.
rimitive methods within a specific sensor range or search space. However, frontiers generated within constrained spaces lead t
ns and dangerous conditions. This paper proposes a quadcopter unmanned aerial vehicle (UAV) equipped with rolling-capable
botics algorithms for reconstruction, pose estimation, and occupancy modeling using compact generative models. Compactnes
d to enhance task efficiency and facilitate extensive deployment. We present a novel hybrid trajectory optimization method aime
uffer in geometrically uninformative environments that are known to deteriorate registration performance and push optimization
ntific sectors; however, propeller-driven vehicles face limitations due to their floating nature. In Remotely Operated Vehicles (R
scenarios, e.g. leakage detection. This becomes particular challenging in realistic cluttered scenarios, i.e. in the presence of co
eration of aerodynamic and hydrodynamic interferences from the marine environment. The turning radius is initially obtained ba
cavation front to evaluate underbreak locations. Considering the inspection and measurement task's safety, cost, and efficiency
and private applications. While installing solar panels (photovoltaic cells) in a stationary configuration is simple and inexpensive
arm the environment and the production of green energy. In this work, we propose a novel methodology to inspect ball joints in
tance of energy efficiency in technological developments such as robotic manipulators. Efficient energy consumption of manipu
wever, the limited flight durations of typical multirotors and the demands on human attention in outdoor missions call for more a
while putting the human operator out of danger. Generating an accurate gas distribution map quickly is of utmost importance in
œwhere” to deploy robots based on prior data with the goal of persistently monitoring a spatio-temporal phenomenon in an e
e a key challenge for high-throughput phenotyping applications. This work implements a novel calibration method to calibrate th
f properties such as particle size, morphology, and composition, to better understand impacts on public health and the climate.
ple sensors. Such fusion ensures that the robot is better aware of its surroundings, including areas of the environment that are
ng manipulation. To capture multimodal contact information within a compact framework, we designed a novel sensor called ViT
dynamic and unstructured environments. While various tactile sensors based on different sensing principles or different sensitiv
plex inferences about our environments even when vision is occluded. The emergence of biomimetic neuromorphic hardware i
omise of providing robots with the ability to measure the rich contact information that gets transmitted through its sense of touc
re complex robotic manipulation tasks. While deep reinforcement learning (RL) methods have demonstrated impressive learnin
tion framework. Bilateral teleoperation enables kinesthetic coupling between the operator and the remote environment with hap
n of their soft elastomer from images would yield poor force measurements when the elastomer deforms uniformly or reaches d
daptive, robust, and cost-effective. However, state estimation for underactuated hands is usually more challenging. Vision-based
rmation for robotic manipulation. Curved and rounded fingers are often advantageous, but it can be difficult to derive illuminatio
to intuitively understand task dynamics and adapt our actions in real time. Transferring such tactile intelligence to robotic syste
d interact with the environment. Optical sensors are a promising solution for high-resolution contact estimation. Nevertheless, th
rform contact-rich manipulation, particularly when precise positioning and contact state of the manipulated objects are crucial fo
become popular in recent years. However, these sensors have a limited receptive field, which can hinder their ability to perceive
nd improve several manipulation tasks. However, obtaining a close-up 3D view of the location where manipulation contacts occ
ure and estimate contact position and velocity. Vibration sensors are an important facet of both human and robotic tactile sensi
ands with advanced touch-related properties; however, their use for compliance discrimination has been poorly investigated and
ever, this topic is less explored than single-arm settings, partly due to the availability of suitable hardware along with the comple
obotics tasks such as grasping and manipulation. However, efficiently exploring such objects using tactile sensors is challenging
pping over. Recognizing the type of terrain can help robot avoid risks and choose an appropriate gait. One of the critical proble
touch to keep track of the object's 6D pose. However, most existing object pose tracking systems in robotics rely exclusively on
bility to deform in unpredictable ways. Previous work has focused almost exclusively on visual feedback to solve this task. We
n across a large contact surface. However, it is difficult to extract contact forces directly from observing the bubble surface beca
se optimization and enhanced food production capabilities. The adoption of automation in vertical farming is a pivotal response
ch can not sense some modalities, like temperature. As an important part of human tactile perception, temperature sensing can
an-robot interaction and allowing robots to adapt to different environments. Despite notable advancements, there are ongoing c
surroundings. In this paper, we introduce CushSense, a fabric-based soft and stretchable tactile-sensing skin designed for phy
stems in manipulation tasks. However, the reality gap of simulators for high-resolution tactile sensors remains large. Models tra
. Through tactile interaction, vision-based tactile sensors provide the capability to capture high-resolution and multi-modal surfa
nsors to fulfill various manipulation tasks. Nevertheless, these methods remain inefficient because of the lack of methods for sim
ct manipulation. By integrating measurements from robot kinematics and an image-based tactile sensor, our framework estimat
specifically targeting the constrained, irregular, and intricate spatial shapes of flexible bronchoscopes (FB) in human bronchial t
able objects as jamming against adjacent objects and walls can inhibit progress. We propose the use of two action primitives---
high performance. However, no suitable measuring method has been presented, due to issues with crosstalk and non-negligibl
ctivities. An intermediate stage of dressing is to manipulate the garment from a crumpled initial state to an unfolded configuratio
ultiple modalities. Specifically, when it comes to interacting with tangible objects, effectively combining visual and tactile sensory
e possible method is to use full-body tactile sensors on the robot and perform safety maneuvers when dangerous stimuli are de
botic applications. The significantly reduced amount of available feedback (limited visual cues, field of view, robot motion space
 identify the unobservable directions of the estimated state which can influence estimation accuracy and help develop consisten
and/or in a vision-denied environment. For a robot to achieve similar desirable interactions, such as administering a swab test, t
complexity of their structure increases the number of manufacturing processes. As a result, the rising cost of sensors has hind
ance manipulation capabilities for medical robots, particularly in contact force control. Current intrinsic force sensing (IFS) meth
esistance to external disturbances. First, a novel proprioceptive-based disturbance estimator is proposed. Compared with othe
ed on imperfect dynamic models of the manipulator and the contact force. To handle the imperfect dynamic information of the m
h-size products. Even though state-of-the-art tactile robots may provide the versatility for increased adaptability and flexibility, t
of Event-based Signal Temporal Logic (STL) specifications, during execution. These modifications allow a user to change the r
brid systems with complex continuous dynamics under temporal and reachability constraints. We model the evolution of the hy
h anytime perception sensor. We formalize the framework and present a general procedure for safety analysis using reachable
ecification but also provides a measure of how much a formula is fulfilled or violated. The calculation of robustness is based on
ethods for processing, modeling, or synthesizing data. Often, this data comprises variables that inherently carry geometric cons
given a global temporal logic specification and formal user preferences for relaxation in case of infeasibility. The relaxation pre
straints. A heterogeneous team of agents with various capabilities is tasked with a Capability Temporal Logic (CaTL) mission, a
s this data, compute-constrained robots often require assistance from remote computation, or the cloud, that runs compute-inte
vary over time. For instance, reaching a kitchen may take more time during lunchtime since the corridors are crowded with peop
ver, most existing automaton based algorithms highly rely on the manually customized representation of states for the consider
vironment interactions increases. In these interactions, a robot needs to consider not only the effects of its own actions, but als
. However, existing methods heavily rely on the predefined hard-coded robot behaviors, e.g, a pre-coded temporal logic formul
subject to both temporal logic and inter-task constraints. The considered inter-task constraints include unrelated tasks, compatib
equires generalization across novel task specifications without violating safety constraints. Linear temporal logic (LTL) is a wide
d direction of a spray-painting gun. Our angle-specific paint deposition model focused on the change in paint deposition due to t
n real time would substantially expedite the adoption of industrial robots due to enhanced system flexibility. To facilitate these e
r ability to go out into the environment and probe it, in combination with what they already know of the world, to provide answer

, a visual-tactile fusion framework for transparent object grasping in complex backgrounds is proposed, which synergizes the a
en objects possess a wide range of different properties (size, weight, friction coefficient, etc.). The human hand is often the prim
to transfer our actions flexibly due to their shared functionality. This work addresses the problem of transferring grasp experien
been developed for manipulation of objects with varying but known dynamic properties (e.g., mass distributions and friction coe
it remains a challenging problem for anthropomorphic hands, due to the lack of object affordance understanding and high-dime
or unseen objects using one annotated object and single-view RGB-D observation via categorical priors. Unlike previous works
as they allow them to evaluate grasp candidates and select the one with highest chance of success. The widespread use of pi
obotic systems require perception and manipulation capabilities to handle a wide variety of objects. Existing systems either ope
g the correct sequence of operations are crucial for robotic manipulation. However, previous algorithms inefficiently combine glo
waves. In this innovative approach, the sound waves generated by a speaker travel in channels embedded in a soft membrane
ques on its surface. Many recent studies use data-driven approaches to address grasping, but the sparse reward nature of this
ntire grasping motion, all finger joints need to be coordinated to achieve a stable grasp. Grasp synergies provide a simplified, lo
uses Visual Feedback and an uncertainty-aware Adaptive Sampling strategy (VFAS) to close the loop. At each iteration, our m
the century-old Fractal Vise. The unusual synergistic properties allow it to passively conform to diverse objects using only one
otics. Executing a successful grasp in a cluttered environment requires multiple levels of scene understanding: First, the robot n
in learning models on large numbers of grasps. Towards the goal of more advanced grasping, we present the Grasp Reset Me
 challenges still hinder their widespread deployment. Failing to detect potential failures and learn to solve them renders state-of
asing impressive performance across diverse robotic learning tasks. While previous works predominantly concentrate on the si
ceive the world directly via onboard sensing. Vision-based learning systems aim to eliminate the need for environment instrume
s by integrating Gaussian Process Distance Fields (GPDF), SE(3) equivariant networks, and Riemannian Mixture Models. The
n recent years. However, it remains challenging to handle a large number of unfamiliar objects in real and cluttered environmen
el of flexibility and adaptability as humans. However, it remains a challenging task to generate an optimal grasping strategy for
challenge, particularly for waste printed circuit boards (PCBs). PCBs are challenging to sort out from other waste materials in pa
ad of object manipulation techniques, where dexterity plays a paramount role. It is well-established that humans exhibit extraor
be systematically evaluated. However, in most cases, the performance of a specific approach is assessed by simply counting t

k up objects in confined spaces. For this purpose, the developed gripper has three operating modes: grasping, finger-bending,
grippers have sensors mounted on their fingers’ surface that provide direct force measurements. However, multiple sensors
and degrees of hardness is a challenging problem in the field of robotics. In order to design a highly self-adaptive gripper capab
h their environment, e.g. to pick up an object, they decelerate to a near standstill. A solution involving a compliant pneumatic gri
ccessfully and quickly transport parts vertically against gravity. We show analytically that vertical vibratory transport is more diff
f conventional rigid manipulators in terms of adaptability and safety, as well as the challenges faced by soft hands regarding ca
n this paper, we present a miniature magnetic gripper and develop a datadriven kinematic model. The gripper comprises four fin
ble and functionally integrated. The gripper is optimized to compensate for gripping errors with particular consideration to poten
e also becoming ever more important as we transition to a more sustainable economy. In disassembly, we are typically only inte
ut (2^{n+1}-1) joints, where n is a design parameter that defines the depth of the fingers' tree structures. The hand is synergistic
ghly deformable and fragile items, that can be easily damaged. Moreover, ingredients for the preparation of the different dishes
ment each other to interact robustly with the environment. However, this combined search is extremely high-dimensional and in
th compliant and passively adaptive fingers. We aim to test the utility of such gripping in agile robotic maneuvering. In the prese
work from multiple perspectives including productivity, worker safety, and quality of work life. The robotics research challenge is
n robotics. Furthermore, the lack of a viable tactile sensor capable of sensing over the entire area of the hand impedes the rich,
at require precise force adjustments during operation, such as grasping an unknown object or using a grasped tool, are difficult
s for one-sided grasping of heavy, irregular items. The gripping technology combines suction with adhesion for grasping and pl
iple human-controlled demonstrations. However, existing approaches often struggle to generalize across a wide range of tasks
structured objects. Sensor arrays play an important role in the exploration, which generates rich spatio-temporal information. In
of robotic hand, and hand rehabilitation of stroke patient, as human hand plays a crucial role in our daily life. The highly under-
al feedback device completely integrated into a prosthetic socket. The core of the system includes two compact vibrotactile actu
umans via handheld electronic devices, such as smartphones, touch panels, and game controllers. However, covering the entire
be beneficial in establishing a more comprehensive and robust skillset for aspiring surgeons, improving their clinical performan
such as minimally invasive surgery, due to their inherent compliance and ability to conform to highly curved paths. Yet the kinem
nsation to the finger. The 1-3 piezocomposite device has a 1 cm^2 footprint, which is an order of magnitude smaller than other
Touch is an active sense in which tactile measurements depends on the contact properties of an interaction---e.g., velocity, for
es. However, many shape-changing displays are composed of individually actuated rigid bodies, which makes them both mech
onstruction methods usually work in a reconstruction-after-scanning manner. However, these methods suffer from insufficient d
mportance in aiding patients regain ambulatory functions. One of the challenges in making these systems effective is the imple
nd unstructured environments by extending the previously proposed Fuzzy State-Long Short-Term Memory (FS-LSTM) archite
ectromyography (sEMG) are among major research questions in robotics. Due to intrinsic complexity and inaccuracy of labelin
 rch works on human-robot handovers focus on facilitating the performance of the human partner and possibly minimising the p
plays an increasingly crucial role. In order to enhance visual recognition accuracy, the utilization of large-scale datasets is impe
ving a task, but can request help from an external expert when needed. However, knowing when to request such assistance is c
well understood, but increasingly important to employ these vehicles in serious jobs such as wildland firefighting and infrastructu
ch is often present in industrial work, is one of the main physical causes of WMSDs. It uses the same set of human joints repea
learning, by allowing humans to teach robots through preferences on pairs of desired behaviours. Nonetheless, to obtain realis
onomous task completion challenging. Though improving models and increasing dataset size can enhance a robot's performan
ounts of pedestrian and vehicle data in order to learn accurate pedestrian-vehicle interaction models. However, gathering data
s efficiency in short-distance trips and environmentally friendly characteristics compared to traditional automobiles. However, m
n in recent years. However, the requirements of high-level operation keep the novice pilots away from participating in it. This pa
effort-compensation device controlled in force. Through this control strategy, in addition to an individual's forearm weight, an ex
n the object being manipulated by the robot and unplanned obstacles. In this paper, we propose a planar compliant contact con
products. Safety is a crucial requirement of the disassembly process since wornout or damaged products could break, possibly
h household tasks like vacuuming or surveilling. While the robots can navigate on their own, users might not want them to go e
gation research. We must evaluate several factors to achieve accurate predictions, such as pedestrian interactions, the environ
ting robots through large crowds of people, but the performance of these learning-based methods tends to degrade in particula
pecification of robot plans that are consistent with humans' behaviors. As is well known, humans perceive uncertainty and risk in
spaces becomes paramount. This paper introduces the notion of real-time human presence estimation for indoor robots. Real-
demand for designing a socially-aware navigation method considering human-robot interaction has risen. When developing and
neighbors. Planning safe motions for autonomous vehicles in such interactive environments, therefore, requires reasoning abou
g. Since computing time grows greatly with environment complexity, improving exploration efficiency and picking the optimal tra
cult to understand and debug when they fail to find a solution to a problem. In this paper we propose a solution to the failure-exp
o ensure satisfactory performance, which could be laborious and impractical when the robot works in some restricted area. To t
n order to accomplish efficient cooperation, trust recognition is vital, empowering robots to predict human behaviors and make t
s through images and time-series contact data. However, these methods rely on human expertise and entail the time-consumin
ultiple human users, two challenges arise: complex tasks are often composed of multiple behaviors which can only be evaluated
ormance and interaction quality. Recent research has confirmed that perceived robot proficiency is a major antecedent of trust.
-language models, allowing robots to navigate in environments with traversable obstacles. We utilize the large language model
erfaces for controlling neurorobotic systems, particularly bionic limbs. These interfaces function by decoding signals captured n
dictions of human hand motion and intentions enhance collaboration efficiency. However, low prediction accuracy increases me
ability to attribute mental states to others. Humans infer the desires, beliefs, and intentions of others by observing their behavio
present a new approach to facilitate the social robot Haru to imitate game strategies from human players' demonstrated trajecto
. While physical activity is commonplace for healthy individuals, it presents challenges for those with visual impairments, as the
-balancing moment by changing the front wheel fork slant angle. Although this method reduces the risk of falling over, it makes
search has explored robotic assistance in meal preparation tasks like cutting and cooking, the crucial task of peeling has receiv
eural networks (MIntNet) for coupled-human robot motion intention prediction. Using only general prior knowledge about the inte
anguage (textual path) is proposed. In recent years, robotics research mainly focus on vision-and-language navigation (VLN), a
as Uncrewed Aerial Systems in search & rescue missions, require nearly continuous human supervision for vehicle guidance a
ns worldwide. Such routine inspections are largely visual in nature, which are qualitative, subjective, and not repeatable. Althou
rce II, improves their efficiency and capacity. Many such facilities use the open-source Bluesky project for experimental control
ms. To this end we utilize several standardized questionnaires administered while participants perform collaborative tasks in rob
raction (pHRI), which can be identified as biomechanical excess of passivity (EoP). This can be used as a central factor in the d
interacts efficiently with the environment via object manipulation to maximize perceptual information, following given language
devices, their controllers often lack the ability to synchronize and adapt to meet individual users' needs. This paper investigates
œrobot-room” elements that, individually or in combination, change the functionality of the rooms we live in, directly controlle
ot collaboration it would be beneficial to assess the trust level during the interaction to adjust the robot's collaboration behavior
applications. To ensure psychological comfort while collaborating, coexisting, and interacting with humans, robots need to respe
ontrollers which are defined as a Model Predictive Control (MPC) problem. The proposed framework includes the design of per
C) with a focus on model learning. In this paper, we are interested in enabling the HRC system to complete complex tasks spe
man partner, specifically in scenarios where preexisting data is sparse. Much prior work in human-robot collaboration uses obse
AI assistants capable of advising teams online during task time, due to the challenges of modeling both individual and collective
ts natural human ability and the strength of an autonomous robot system. Our approach enables both flexibility and efficiency o
le agent localization algorithm. Despite the global research effort to improve SLAM, its mapping component remains limited and
hindrance in the application of deep learning in robotics. Acquiring real-world data requires considerable time and effort, yet it m
small and medium-sized companies. However, classical industrial robots struggle to cope with product variation and dynamic e
d robots work closely together. To ensure safe human-robot collaboration (HRC), regulations have been updated introducing th
obot collaborative systems promise to address issues such as workspace constrains and laborious tasks. In particular, shared c
human-robot interaction. SIREN utilizes a surface transducer to produce sound by vibrating the frame of an underwater robot,
e co-manufacturing of furniture. For these tasks, it is essential to encode tasks from human demonstration and reproduce these
cations of autonomous robotic systems. Recent studies have reported encouraging results using deep reinforcement learning (
mally invasive surgeries, they are predominantly used by the human operator to directly teleoperate the tools. This paper aims t
zing to novel environments. Therefore, adapting to and learning from online human corrections is essential but a non-trivial end
d guidance to adjust the end-effector pose for spatial operations. The rotational operation is less intuitive to humans than transla
rk, through which new collaborative skills are extracted from a one-shot human demonstration and learned through Riemannian
ety by amalgamating human expertise with autonomous assistance. This paper proposes a user-centered shared control frame
ronments, including human-robot collaboration and teleoperation, has become increasingly evident. Building on the foundation
c system devise a shared plan with an appropriate task allocation from this abstract goal and agent condition? Classical AI task
r actions with the people around them. Part of such coordination involves ensuring that people have a good understanding of ho
rkspace. Such proximity requires special attention in ensuring safety for the human operator, which is often translated in collisio
xisting control methods for HRC consider either contactless HRC or physical HRC, hindering more efficient HRC. The proposed
ators and robots that work in close contact with each other. As a consequence, the safety regulations have been updated in ord
ompassing understanding, perception, and planning. For such a multifaceted challenge, previous VLN methods totally rely on o
an intelligence. Accordingly, carefully designing this mechanism in artificial intelligence can be a successful strategy to increase
gh exploring novel ways for humans to communicate complex goals and tasks in reinforcement learning (RL) setups. One way t
o-real gap and expanding input modalities for users, but the interaction with these systems generally exists only in the input sta
Knowing objects and materials before touching is desired by the target users but under-explored in the field of human-centered
vironments by means of a mobile manipulator. The Human-Robot system consists of three key components: a mobile base, a ro
rom demonstrations and combined into complex behaviors. However, merely equipping robots with a fixed set of innate MPs is
enting astronauts during an Extra-Vehicular Activity (EVA) in a partial-gravity environment. We investigate the effectiveness of
assistance. However, in order to assist individuals who are lying in bed, robots must be capable of dealing with a significant obs
are influenced by numerous cognitive biases. One class of cognitive bias deals with the ordering of items or interactions: inform
bots and vehicles in populated environments. It is fundamental for various applications, including motion planning, tracking, hum
nd reason on the semantic knowledge of the environments. Despite a resurgence in the development of semantic reasoning arc
HRICA). Different from previous robot painters, HRICA allows a human user and a robot to alternately draw strokes on a canva
e form of independent mobility for those who suffer with motor disabilities. Although many control interfaces exist to operate the
images and infrared images. To address this problem, we propose an RGB-IR image fusion captioning model, which can take
nipulation. We propose a semi-automatic framework that can extract expressive strokes from 3D models and draw them in orie
s to see in real-time where robots are located, even if they are not in line of sight. The operator can also visualize the map that t
for learning the Energy-exchange Dynamics (EED) of a spring-loaded biped robot. Our approach enables on-site walking acqui
amic environments. However, it is still challenging for legged robots to achieve rapid and efficient locomotion on deformable, yi
n unknown uneven terrain. The adaptive controller consists of a trajectory tracking control law developed via backstepping meth
study, we introduce a Wasserstein adversarial imitation learning system, allowing humanoid robots to replicate natural whole-b
rch challenge. Computational complexity is one of the primary restrictions that preclude motion planners from generating adapt
x (NWR) in the upper limb of a humanoid robot so that it can avoid the potential damage caused by noxious heat. We use a spi
ng the detection of potential falls while standing caused by abrupt, incipient, and intermittent faults. Leveraging a 1D convolution
es. The diversity of human bodies makes it difficult to imitate various body shapes, especially the shoulders. This paper propose
ne estimation of joint-torques of humanoid robots without joint-torque sensors. At the feature level, the proposed approach con
ti-agent cooperative perception algorithms usually select the simulated sensor data for training and validating. However, the pe
in the shared bird’s-eye view (BEV) representation space. However, our empirical findings indicate that previous methods h
vehicles and infrastructure, providing a global vantage point with rich semantic context of road conditions beyond a single vehic
systems. To simplify this task, vehicles are equipped with various illumination systems, including turn indicators, emergency ligh
degrades the visibility and results in less distinctive features for dense correspondence matching. While some previous learning
air systems. However, developing an instance segmentation method that simultaneously performs multi-class defect detection
llel LiDARs and construct a hardware prototype in our experimental platform, DAWN (Digital Artificial World for Natural). It emph
els for more comprehensive understandings of 3D scenes compared with traditional perception tasks, such as 3D object detect
ds for 3D human pose estimation (HPE) have often relied on 2D image features and sequential 2D annotations. Furthermore, th
mantic information of scenes from only 2D images. It has garnered significant attention, particularly due to its potential to enhanc
y is largely dependent on the quality of input detections. In many applications, such as autonomous driving, it is preferable to o
h company might use the identical classic neural network architecture based encoder for feature extraction. However, the data
nts and leverages their relative poses determined by GPS signals to effectively fuse information, mitigating the limitations of sin
ories of surrounding agents. To do so, it requires solving mapping, detection, tracking, and then forecasting problems, in a multi
ing additional viewpoint and expanding the sensing field. Existing cooperation paradigms are either interpretable (result cooper
ts seriously degrades the quality of point clouds. Previous research has shown that the extent of degradation is correlated with
ng communication bandwidth and detection performance simultaneously. Intermediate fusion is recently studied to reach a bet
ystem to perform perception tasks. Such a framework's reliability could be limited by occlusion or sensor failure. To address this
-to-end control from visual observations. These advanced deep models, ranging from convolutional to vision transformers, from
effectively solve long-horizon tasks and perform novel ones. Recent advances in skill-based imitation learning have shown that
to maximize the cumulative rewards after performing a sequence of actions. Existing approaches learn a value function from h
verages the image-level and pixel-level capabilities of features extracted from Vision Transformers trained with DINO. When inte
iques like kinesthetic teaching or teleoperation puts a heavy burden on human supervisors. In contrast to this paradigm, it is of
us systems how a human user wants them to perform a task. Previous works have shown that actively synthesizing preference
mplex dynamics and the infinite-dimensional configuration space. Fabric unfolding manipulation takes on critical significance in
end robot control. In the context of autonomous driving, this work explores and compares a variety of alternatives for model opt
of rigid and a simplified class of deformable 3D objects subject to energy constraints. Towards this end, we propose the notion
tached to fingertips is controlled using flexible and adaptive deformation of the cloth. First, an analytical model of the ball-on-clo
mportant research topic. In this paper, a linkage-driven underactuated three-finger hand is proposed to imitate the flexion/extens
s for manipulation tasks using Deep Reinforcement Learning methods. Currently, Asymmetric Actor-Critic approaches are used
eir actions with resulting states. In reinforcement learning (RL) contexts, policies have the burden of deciphering these complica
s external contact mode with the environment, and adjusts its in-hand pose simultaneously. The proposed manipulation task le
lexity. Nonetheless, learning these skills in simulation still exhibits poor sample-efficiency which stems from the fact these skills
ack. However, the distinct nature of these modalities poses significant challenges. In this paper, we introduce a system that leve
d. In-hand manipulation remains a demanding task as the object is never stable and task success relies on carefully synchroniz
ee translational degrees of freedom as its locomotion method. Using a robot frame having a triangular pyramid shape, the robo
ate robust cable grasping in cluttered environments. Utilizing physics simulations, we generate an extensive dataset that mimics
ustrial robots is presented and experimentally validated. A model-based controller that leverages all typically available paramete
hain in many industries, but are notoriously challenging to automate. We present RoboGrind, an integrated system for the intuit
d taxis are gradually gaining acceptance. However, their application scenarios remain significantly fragmented. Typically, an Au
ges, including diminished contrast, increased noise, loss of fine details, and unnatural color reproduction. These factors can sig
tion for visual feedback of the target region. However, the limited field-of-view (FoV) of the microscopic camera necessitates ca
ams, we present RoSSO, a Python package designed for solving Markov chain optimization problems. We exploit machine-lear
ontrolled physical interaction with the environment, the majority of robot automation nowadays is still based on fixed manipulato
ctive assumptions about the system dynamics and rely on approximations to characterize the evolution of the underlying uncert
Traditional autonomy systems rely on object detection to find the agents in the scene. However, object detection assumes a dis
s importance in robotics. In this paper, we investigate the behavior of safe receding horizon motion planners when the update in
s that economize on costs and decrease carbon emissions. Current predictive control methods depend on an accurate model o
utonomous Mobile Robot (AMR) from an initial state to a terminal state without colliding with obstacles, even when subjected to
rthermore, in fastchanging, uncertain environments, the conventional warmstart, using the optimal trajectory from the last times
ted by factors such as complex logic specifications and high-dimensional robot dynamics. This paper presents a novel reinforce
appropriate selection of the input distribution used for sampling. However, it can be challenging to select inputs that satisfy outp
with forward propagation. This algorithm allows for the simultaneous exploration of a robot's state and control spaces, thereby e
s is presented. The advantages of reactive motion planning are combined with a Policy Iteration Reinforcement Learning schem
g to regions of state-action space where the model is accurate: also known as a model precondition. Empirical real-world traject
eir unstructured nature. In the absence of high-fidelity maps, occlusions from obstacles and terrain lead to limited information av
new tasks. While in the past, robot programs were often written statically and tuned manually, the current, faster transition tim
g., via sampling and exploration or starting optimization from a straight-line path. However, most motion planners have to exper
autonomous driving, particularly in the domains of prediction and planning tasks. However, the application of vector-based bac
planning algorithms. While classical sampling based algorithms fall short of computational efficiency in high dimensional planni
under-actuated nature. Unexpected disturbances can cause the object to deviate from a predicted state, making it necessary
namics constraints or with black-box models. However, they struggle to find high-quality solutions, especially when a steering fu
ne, and poor calibration can easily lead to wasted debugging time in downstream tasks. Contact-based calibration methods let
gh levels of predictive accuracy. Such condition monitoring tools must estimate dynamic quantities, such as Remaining Useful
ill complexity is often over- or underestimated based on individual experience, and the real-world performance gap is difficult or
partial differential equation (PDE) control of a flexible manipulator. The proposed scheme considers the effects of the boundary
ng, incorporating machine vision capabilities into a custom-built sewing machine. The vision system captures an image of the f
 s for humans and is still largely performed by hand. Due to the high curvatures of complex geometries, different areas of the su
scheduling to improve productivity and quality. However, there have been few studies on this integrated control due to a lack of
ging tasks. However, on-board limitations regarding computational and battery resources are hindering factors regarding the de
nd Automated Vehicles(CAVs) enable autonomous vehicles to cooperate with each other to accomplish the lane changing task
a learning model among vehicles without sharing users' data. However, data from autonomous vehicles usually suffer from the
ification and validation (V&V) of autonomous vehicles (AV). Scenarios in the database are often labelled with semantic Operati
mains, yet enabling their safe deployment remains an open problem. This work’s motivating question is: In safety-critical set
e in evaluating autonomous vehicle control systems in a simulation environments. Our methodology leverages a hierarchical for
ing task due to the unknown characteristics of agents and their varying intentions. Besides using semantic map information and
traffic efficiency in complicated scenarios. Difficulties remain unsolved in developing an effective and efficient coordination strat
ng systems. After being deployed on public roads, autonomous vehicles need to interact with human road participants with diffe
y by issuing timely warnings to drivers. However, current ADAS do not take into account the driver's cognitive state when delive
danger pedestrians and other vulnerable road users (VRU). As raw evaluation metrics are not an adequate safety indicator, rec
l problem in autonomous driving. To achieve this goal, the existing point-based methods mainly choose to adopt Random Samp
ety while driving, the tracker must be able to reliably track objects across frames and accurately estimate their states such as v
improving road safety through collaborative perception among connected and autonomous vehicles (CAVs). However, time-va
r to realize the benefits of autonomous buses. We propose a novel algorithm for planning networks of routes for autonomous bu
stem (QSLS), which consists of a point-mass load and a rigid-body quadrotor connected by an inelastic cable. To construct the
technologies to support advanced mobility use cases. As car manufactures keep developing SAE level 3+ systems to improve t
ng moving objects in 3D world thus plays a pivotal role for applications like trajectory prediction, obstacle avoidance, and path p
have proven futile in alleviating current congestion levels leading researchers to explore ideas with robot vehicles given the incre
particularly as planners, has garnered a significant degree of attention in recent scholarly literature. However, a substantial pro
autonomous ve-hicles. However, existing methods often ignore the importance of cost learning and treat them as separate mo
ns and vehicles) to make optimal decisions for navigation. The existing methods focus on techniques to utilize the positions and
mous vehicles (AVs), it is critical to ensure that model predictions are safe and reliable. In this paper, we examine the robustne
assistance systems (ADAS) and autonomous vehicles (AVs). In this work, our pursuit was to strike a balance: a method that en
ing due to its efficiency and reliability. However, real-world autonomous driving contends with a multitude of diverse traffic scen
environment where the location and number of future ride requests are unknown a-priori, but can be estimated by an empirical
om the absence of rare and safety-critical scenarios in the long-tail naturalistic driving distribution. To tackle this challenge, a su
designing a learning model for a new scene. Transfer learning can leverage knowledge from a learned domain to a new domain
ns. Accurate prediction relies on comprehending the surrounding map, which significantly regularizes agent behaviors. Howeve
oad vehicles re- mains a huge challenge. In this work, we present a decision module that is based on set-based reachability an
is crucial for the self-driving vehicle to make safe decisions and avoid accidents. Existing deep learning-based methods have s
a key feature of an advanced driver assistance system. Existing driver warning technologies, mainly the forward collision warn
task, wherein the vehicle must identify a potential gap and coordinate with surrounding human drivers, each of whom may exhi
he field of autonomous driving, however, often very simple scenarios are being examined. Common approaches use non-interp
or maneuvers, which remains a challenging problem in self-driving. In this work, we propose a chance-aware lane-change stra
able challenge, especially in mixed autonomy environments. Traditional approaches often rely on computational methods such
eloping intelligent traffic systems to deploying autonomous driving systems in both real and simulated worlds. In this paper, we
of stroke patients in the early stage. The surface electromyography (sEMG)-based control can facilitate friendly interactions be
a few seconds of the motion's onset, and (b) leveraging this predictor to facilitate shared-control manipulation tasks, easing the
robot assistive control. We focus on predicting and distinguishing motor intentions of left/right arm movements by presenting: i)
resent a person's true emotion. However, because it has a low intensity and a short duration, it is hard to be recognized correc
f human/machine autonomy share roadways. Driver intention can be leveraged to improve road safety, such as warning surrou
e intersection of spatial planning and natural language interfaces for navigation. We focus on following complex instructions tha

diction algorithms. Based on various contextual information, we extract driving scenarios for a meaningful and systematic appro
a pivotal role in ensuring prompt decision-making, safety, and an intuitive collaboration experience. Precise and robust HMIP w
Vs) by allowing interactions with the environment. The difficulty in performing a mid-air grasp is the high probability of impact be
r the execution of complex tasks with precision and efficiency. However, current robotic shoulder designs lack the compliance a
ultiple cam mechanisms. The actuator is intended for use in walking assistance, focusing on assisting individuals with diminishe
aminated surfaces in domestic and industrial settings typically require manual scrubbing which can be costly or hazardous. To
exterity and inherent structural compliance, making them a popular research topic. However, previous studies have overlooked
inery are becoming increasingly extensive. However, the contact friction between the tendon and sheath seriously affects the tr
uttered environments. While their compliance allows them to conform safely to obstacles, it leads them also to buckle under ten
h deployability, the fixed frame is usually substituted by four masts. However, not having any rigid fixture between the masts red
ower transmission from remote actuators to specific destinations. The pretension of Bowden cables profoundly influences syste
orresponding forward and inverse kinematics are critical yet often challenging tasks, especially when dealing with highly individu
ping performance is influenced by multiple factors such as driving force, rowing trajectory and robot structure. The interaction fo
sidering jerk constraints. As a first step, we analyze the traversal of each individual path dimension. We derive a range of feasi
ns and orientations at the segment tips of a piecewise constant curvature (PCC) continuum robot with up to three segments. Fo
tial circular arc, a straight component, and a final circular arc. These are commonly called CSC paths. By modeling the start an
n tasks. Generally, these measures examine the system’s kinematic transformations from configuration to task space, but
nes. However, these highly capable robots are usually degraded to simple repetitive tasks such as pick-and-place. On the other
ctories. This paper focuses on trajectory segmentation for 3D rigid-body motions. Most segmentation approaches in the literatur
cific finite elasticity problems that arise in the context of soft robotics. Whereas continuum mechanics is a very classical area of
and controls for a robot manipulator to interact with the environment through a hitting or a striking motion. Given the task of plac
ge industry, and are used in a wide range of applications, carrying anything from lightweight tape cartridges to entire pallets of g
ace operations. However, conventional trajectory planning methods in either active joint space or Cartesian space have some s
l. We will show that their configurations may be unstable, and moreover, that assessing the stability of the robot with the Irvine's
Rs). However, it’s frequently disregarded as negligible in CDPR simulations and designs. In this paper, we propose a numer
le-platform (MP) of the cable-driven parallel robots (CDPRs) can provide a sufficient amount of wrench and twist. Traditionally,
hree-link brachiation robot equipped with passive hook-shaped grippers. Brachiation, an agile and energy-efficient mode of loco
one of the effective approaches for underactuated balance robots. By sequentially controller design, trajectory tracking of the a
obotic systems. However, all recent studies on learning-based energy function synthesis only consider the feasibility of the con
ld applications. To address this issue, control barrier function (CBF) and its variants have attracted extensive attention for safet
. But the safety of manipulation skills that robots employ is still an open problem. Considering all possible failures during skill lea
ure motion in its task and motion planning to ensure safe and effective operation. This requires symbolic reasoning about proba
rediction. This model provides interpretability about modality usage and demonstrates robustness against disturbances. We us
w-rigidity bodies of some robots. In particular, the deflection of the body changes during tool-use due to object grasping, resultin
nsferring prior experience. Priors could enable robots to handle new tasks zero-shot and help prevent failures, which can be pa
dapt online to damage in robot locomotion tasks. However, finding diverse, high-performing controllers requires expensive netw
asks in Fundamentals of Laparoscopic Surgery (FLS) via imitation learning, aimed at the development of an autonomous robot
ces required to build complex robotic solutions. Nevertheless, such data-driven approaches often lack the ability to provide guar
ch is commonly used to learn movements from human demonstrations and adapt the learned movements to new task scenes. A
tracts the task to be executed and directly allows the robot to learn task-specific rules from synthetic demonstrations given thro
on, a method for transferring such human manipulation skills to robots, requires techniques for segmenting the demonstrations
ng new physical skills to many robots. In this respect, it is a good alternative to pure reinforcement learning, which can be time-
e physical properties of the manipulated object. To enable robots to capture the underlying distribution of object properties nece
nts is challenging due to the safety constraints and complex configuration. Traditional avenues usually involve convexifying obs
ety of tasks within the vision and language domains. However, using such models for autonomous systems is challenging beca
le to generalize to complex worlds and scale with number of objects. This work postulates that extracting action primitives, such
th torque sensors, for example, can be used so that the robot can follow the movements of the human, but such methods intro
m a single demonstration. Our approach distinguishes itself from existing imitation learning methods by demanding fewer exper
a given procedure. While many studies are focused on measuring in situ forces, little attention has been devoted to relating the
sed in robotics applications and research. DMPs combine linear dynamical systems to achieve robustness to perturbations and
ing and supporting, whether robots can effectively learn and perform the same type of operations remains relatively unexplored
cies. However, explicitly specifying all constraints in an environment can be a challenging task. State-of-the-art constraint infere
emonstrations, allowing users to communicate their instructions in an intuitive manner. Recent progress in LfD often relies on k
placed in a desired pose relative to another object. Previous work has shown success in learning relative placement tasks from
mimicking expert behavior. While existing methods can accurately replicate expert demonstrations, they often exhibit unpredic
ng a stable dynamical system (DS). To increase the robots' resilience to perturbations during tasks that involve static obstacle a
core challenge: overcoming the domain gap between the training and the deployment environment. This is particularly difficult fo
plans from kinesthetic demonstrations using Neural Ordinary Differential Equations (NODE). To ensure reactivity and robustnes
the field of Learning from Demonstration (LfD). Nevertheless, they confront the formidable challenge of striking a delicate balan
en shown as an effective way for demonstration learning. However, the stability guarantee may sacrifice the demonstration lear
m demonstrations to autonomously reproduce constrained skills in new scenarios. However, ICL suffers from an ill-posed nature
ble, limited number of demonstrations. To address this issue, we proposed a novel motion planning framework referred to as ex
ncy, but still struggles with long-horizon sparse-reward tasks, especially in offline settings where the agent learns from a fixed d
ynamics of the environment. Latent variable models, based on mutual information maximization, have been successful in this t
rials separation and recycling. Waste handling applications differ from milling in manufacturing processes, as they engender co
ream tasks. Prior works have defined pre-training tasks via natural language instructions, but doing so requires tedious human
ased RL. Effective state representations are crucial for high performance in vision-based RL. However, in order to learn effective
l approaches have demonstrated robust locomotion, but the exhibited behaviors lack diversity and agility. They employ approxim
efficiency and generalization capabilities compared to model-free RL. However, effective model-based RL solutions for vision-b
eal-world constraints is extremely challenging. It requires exploiting the dynamics of the interaction, and planning through both
nce modeling problem, drawing inspiration from the success of Transformers. Offline RL is often limited by the quality of the offl
ith environments as their natural counterparts. While their hyper-redundant and high-dimensional characteristics add to this ada
d challenges with large state spaces. While RL excels in these tasks, training time remains a limitation. Reward shaping is a po
multi-robot systems. Nevertheless, the low sample efficiency of MARL poses a significant obstacle to its broader application in r
mant robots. Finding these smaller neural network architectures requires repetitive experimentation and can be time-consumin
such as legged locomotion. However, RL in the real world is complicated by constraints on efficiency, safety, and overall trainin
al biped robot in combination with Sim2Real training. Position control has been popular as it has been shown to be more sampl
 eg. quadruped locomotion). In previous works, the RL-based controller was typically implemented as a single neural network w
g tasks. In 2021 we released TartanDrive 1.0, which is one of the largest datasets for off-road terrain. As a follow up to our orig
ement (UIE). However, its practical utility for high-level vision tasks, such as underwater object detection (UOD) in Autonomous
ve vehicle mobility analysis is required. However, the computational complexity of existing mobility analysis methods, such as d
m in field robotics. Often, the need for precise, accurate traversability estimation is in direct opposition to the limited sensing an
s with LiDAR point clouds. This method includes a Transformer-based 2D-3D matching network called D-GLSNet that directly m
botics, especially when factoring in the varying capabilities of robots across different terrains. To achieve this, the robot system
path traps due to limited sensor perception range and lack of spatial awareness. Existing approaches that avoid local optimum
namics and dealing with many human-robot interaction applications. However, collecting massive and annotated real-world hum
transferring in the reinforcement learning setting. We identify several challenges towards this goal and propose a transferring a
diverse goals to learn a set of skills. Despite the excellent works proposed in various fields, reaching distant goals in temporally
forward models learn this representation through extensive supervision (e.g., object class and bounding box) although such gr
st notably in long-horizon manipulation tasks entailing a plethora of different sequences. In this work, we propose the Intrinsica
d robots. In tasks such as in-hand manipulation, where robots frequently switch between contact and non-contact states, it is im
pulation. Most previous works address this problem in a goal-conditioned way and adapt goal images to specify different tasks,
blems in robot learning. Challenges stem from the intricacies of visual perception, the demand for precise motor skills, and the c
arning sustained-contact manipulation skills like opening a drawer, pulling a lever, turning a wheel, or shifting gears. CIPs have
Whilst some objects have highly predictable state evolution e.g., the ballistic trajectory of a tennis ball, reactive targets alter their
ges in model representation, planning, and control. Recent efforts have attempted to address these challenges via data-driven
ution of increasingly complex tasks. However, policies learned through reinforcement learning frequently exhibit limited generaliz
horizon and nongeneralizable operations plan. This paper presents the Tactile Ensemble Skill Transfer (TEST) framework, a p
tics. At the heart of these challenges is generalization: How can we train a tactile-based policy that can manipulate unseen and
applications. It often faces scalability challenges due to the computational complexity of robotic arm rearrangement and the hig
towel. However, existing works for policy learning that rely on data augmentation are still limited in achieving this level of genera
e to its generalization promise, complex task capacity and minimal human intervention. Nevertheless, DRL-trained controllers ar
s in diverse fields such as agriculture or industry. However, a major challenge lies in acquiring accurate deformation models tha
rementally and maintaining accurate place recognition for previously visited areas. Unlike humans, artificial neural networks suf
parse reward and long decision horizon nature. Hierarchical reinforcement learning has the ability to leverage knowledge at diffe
ul diverse behaviors remains a challenge. Typically, task performance needs to be compromised in order to achieve diversity. In
e development of snake robots featuring body-surface tactile perception. In this study, we proposed a locomotion control frame
ches have been exten- sively investigated, they are affected by adverse illumination and weather conditions. Therefore, radar s
ation and mapping, providing robots with more comprehensive scene map information. However, NeRF often struggles with geo
e capability of these signals to capture high-resolution angular and temporal measurements. Prior heuristic-based methods, ba
erior performance, particularly for navigation in dynamic environments. However, most existing methods solely rely on deep ne
hrough unknown areas. Typically, the robot builds the navigation maps incrementally from local observations using onboard sen
first attempts in human-robot collaboration (HRC). However, a fair comparison of RL approaches in HRC under the constraint
to their unpredictable movements. Current Reinforcement Learning-based motion planners rely on a single policy to simulate p
hicles. However, they rely on fixed reward functions that struggle to balance values like safety and efficiency. How can autonom
urn while keeping the agent safe at the same time remains an open problem. Motivated to address this challenge, this work prop
reinforcement learning (DRL) approaches can seldom harness and exploit symmetry effectively. Often, the learned behaviors f
industrial scenario of Cartesian robots. The goal of this problem is to transfer all the objects from sources to targets with the m
 isfying cost constraints. When learning from scratch, safe RL agents tend to be overly conservative, which impedes exploration
presents the development and training of a robotic system designed to navigate and solve a labyrinth game using model-based
tiple agents to explore an unseen environment via sensory signals in a limited time. A popular approach to exploration tasks is t
e to move around autonomously and safely. The support surface is commonly estimated based on exteroceptive depth measur
systems evolved additional limbs beyond legs that facilitates postural control. This work shows how a manipulator can be an as
ated soft bodies, presents a unique set of challenges that traditional methods struggle to address efficiently. In this study, we ta
ctors of controller parameters. The proposed method first learns a mapping from the high-dimensional controller parameter spa
f quadruped agents. The composer policy links pair of experts via transitions to a sampled target state, allowing experts to be c
n-like behaviors is more often studied on costly bipedal robot platforms, we present a solution over a lightweight quadrupedal ro
gh-level language command like “cartwheel” or “kick”. Although human motion datasets exist, this task remains p
ronments to complex harsh terrains, thanks to their agile locomotion ability. However, these robots can easily lose their locomo
al or mechanical failures. Unexpected joint power loss and joint locking can immediately pose a falling threat. Typically, controlle
doors or switches, along with effective navigation. However, such complex behaviors are difficult to learn because they involve
y motions in response to unexpected disturbances. To promote more natural and compliant balance recovery strategies, we pr
proaches, such as model-based control and learning algorithms. This work proposes a novel Imitating and Finetuning Model Pre
swing foot reflection. While adaptive balancing counteracts perturbations to the robot, adaptive swing foot reflection helps the r
recise eye-muscle coordination and movement. Getting robots to do the same task requires overcoming similar challenges. Cla
with their actions and movements to prevent accidents. Despite its importance, these risks are not explicitly modeled by currentl
s. Recently, Rein- forcement Learning (RL) has shown promising results in legged locomotion and various methods try to integr
omotion skills. To learn locomotion, it is commonly necessary to introduce additional reward-shaping terms, such as an energy
ielded impressive results in terms of achieving rapid and robust locomotion across challenging terrain, such as rugged rocks, no
ation that incorporates signal temporal logic (STL) specifications for task completion guarantee and robustness quantification. T
and external disturbance rejection performance by seamlessly rendering several walking strategies of the ankle, hip, and footst
h as model uncertainty, random disturbances, and imperfect state estimation. In this work, we propose a novel metric for locom
n that an individual experiences on a daily basis, and for an effective, real-world humanoid robot the same will likely be the case
at distances and overcome difficult terrain. In this article, we show that the 3-D Divergent Component of Motion (3D-DCM) frame
ata-driven state representation with a Reinforcement Learning (RL) based locomotion policy. The framework utilizes an autoenc
s a tightly-coupled LiDAR-Inertial-Kinematic Odometry (LIKO) for biped robot state estimation based on an iterated extended Ka
yload capabilities and efficiency. It presents a new high-torque and low-inertia leg design, which includes custom-built high-effic
with a spring-loaded leg. The main motivation is to enable the terrestrial locomotion capability on aerial vehicles so that they ca
he four oscillators is coupled only to itself and its corresponding leg through local feedback of the ground reaction force, which

st efficient manner by the muscular forces for mammals. For this reason, we can see many disabled quadruped animals can sti
y captured “in- the-wild” video footage of humans and animals to legged robots. SLoMo works in three stages: 1) synthe
kspace and joint torque limits. When facing terrain where the height of obstacles is close to the leg length, the locomotion robus
oration of asteroids and moons. The robot weighs 5.2kg and has a body size of 245mm while using space-qualifiable compone
ecific robot morphology, mass, and size. The learning process must usually be repeated for every new robot, where hyperpara
ever, the stability of these robots in different gait sequences remains an open question. This study investigates the relationship
ation scenarios. In order to interact with and manipulate their environments, most legged robots are equipped with a dedicated
p reinforcement learning. But the sim-to-real gap and low sample efficiency still limits the skill transfer. To address this issue, we
r, a gap always exists between the reduced-order control model with uncertainties and the real robot, which degrades its perfor
 of robust cooperative locomotion by two holonomically constrained quadrupedal robots. A novel interconnected network of redu
forcement learning (RL) to achieve versatile and robust legged locomotion. Our approach enhances the RL training process by
el gait, which is achieved by rotating the stance and swing legs in the same direction. First, a model of a planar 3-DOF X-shape
n robots, particularly suitable for those with non-trivial limb mass. To simplify the complexity and computation cost of using the
cts. In this work, we aim to increase the robustness of quadrupedal periodic forward jumping (i.e., pronking) by unifying cutting-

However, they require a model of the surroundings in a representation that is sufficient enough for downstream tasks such as f
in underactuated domains (e.g. with feet off the ground) where the system has limited command of its acceleration and a const
odeling the dynamics of dual floating bases and developing effective planning and control methods present significant complexit
for agile quadrupedal locomotion over a constrained rebar environment. To this end, our hierarchical planner incorporates loco
uped robot integrated with thrusters in amphibious environments, allowing it to move efficiently in water while maintaining its gr
ironments predominantly created for humans across various domains. One significant challenge faced by many of these robots
aordinary flexibility which allowing them to move efficiently on variable terrains. The previous researches have indicated the legg
ms has raised a vast demand for data with ground truth annotations. However, available public datasets currently cannot provide
B data, 3D point clouds, IMU, and GPS data. We introduce a new benchmark targeting visual odometry and SLAM, to advance
s. Lightweight depth estimation methods are crucial for resource-constrained edge devices. However, existing lightweight metho
e Recognition (VPR) descriptors, crucial to enable real-world mobile robot localisation. Two parallel lines of work on VPR have

avigation systems. Existing methods often rely on contrastive learning: image descriptors are trained to have small distance for
ositions through visual place recognition (VPR) is crucial not only for reducing localization errors caused by visual odometry but
vigation. However, severe appearance and perspective changes often lead to degradation of algorithm performance. Current m
ng to understand the surrounding environment. Recently, Camera-LiDAR BEV feature fusion methods have attracted increasin
g, it holds huge implications to endow these mobile devices with spatial computing capability. However, ultra-efficient visual-ine
e being the lack of depth information in camera images. This paper introduces OCC-VO, a novel framework that capitalizes on r
plicit representation has brought new opportunities to robotic mapping. In this paper, we propose the multi-volume neural featur
ibiting substantial performance degradation when applied in challenging circum- stances, such as low-visibility and nighttime sc
n data to generate high-quality reconstructions that are geometrically accurate and capture photo-realistic textures. This system
s driving or autonomous valet parking, textit{etc}. Existing methods rely on a fixed and memory-inefficient map, which lacks rob
ects will leave a large number of traces in the map. These traces will usually have bad influences on the localization accuracy a
rstanding capability of SLAM systems. In robotic and augmented reality applications, object maps with semantic and metric info
or reliably performing robotic tasks. Probabilistic volumetric mapping (PVM) is a well-known environment mapping method using
otal aspects of the mobility of legged robots. This paper presents an onboard framework tailored to the detection of traversable
mous systems in wild, outdoor environments. Given the lack of large-scale annotated datasets for off-road navigation, strictly-s
where a robot can navigate in off-road environments. An off-road environment is an unstructured environment comprising a com
y known navigational patterns, which may be represented as directional pathways or road lanes. Inferring these navigational pa
n of mobile robots within complex outdoor environments. While cost-effective and easily deployed, camera sensors are sensitive
ds the existing branch-and-bound (BnB)-based 2D scan matching (BBS) algorithm. To reduce memory consumption, we utilize
ce of the robotic downstream tasks. To address this problem, we present a novel real-time dynamic instance-aware static map
nity due to its broad applications, e.g., household robots and rescue robots. In this task, an embodied agent must search for an
ation from a sequence of posed RGB images. Towards this, we introduce 3DIML, a novel framework that efficiently learns a lab
, many custom features defy standard road semantic categorizations. Addressing this challenge, our paper introduces a metho
 cale environments. Both SLAM-based and map-based localization will increase the computing load due to the increase in map
complements and improves robustness over existing common sensor modalities, as subsurface features are less sensitive to w
Structure from Motion (SfM) pipeline. However, such methods often yield reconstruction outcomes that lack crucial scale informa
om LiDAR intensity with geometry-based point cloud registration. The focus of our work is to improve the robustness of LiDAR-i
elerated Stein particle filter. To update a massive amount of particles, we propose a Gauss-Newton-based Stein variational grad
algorithm tightly couples scan-to-scan and scan-to-map point cloud registration factors along with IMU factors on a sliding wind
for human drivers. However, the analogous robotics capability, visual place recognition (VPR) with limited field of view camera
ges or with low textures. In this paper, we propose a Transformer-based pseudo 3D image matching method. It upgrades the 2D
es to remove incorrect correspondences. However, these methods are typically fine-tuned to overfit certain benchmarks and str
-scale environments is how to effectively use global localization to correct the cumulative errors from long-term tracking. This c
etry. In recent years, a variety of methods have been proposed for self-supervised learning of ego-motion and depth estimation
nition, hindering the performance of Simultaneous Localization and Mapping (SLAM) systems. This paper presents a novel mod
acility maintenance. However, existing methods can struggle when dealing with uninformative geometric structures in challengin
ing tasks, mitigating the impact of shift errors caused by dynamic changes in the environment. Due to the limited ability of back
ons, leading to inconsistency in state estimation. The newest solid-state LiDAR with a non-repetitive scanning pattern makes it p
he 2D LiDAR scan of the entire environment. We introduce a method (VioLA) that starts with building a semantic map of the loc
maps. A wayfinding map abstractly portrays the environment, and humans can localize themselves based on the map. Howev
l images. Recent works tackle this problem by exploiting deep neural networks to regress camera poses. However, spatial and
a 6-DoF camera pose for an input query image relative to a given 3D model of the environment. The most accurate approach is
remendous progress in the past decade. Most recent methods usually adopt neural networks to learn the mapping from image
ues, and it is used in robotics applications for SLAM and localization. Since typically a mobile robot has access to a continuous
for robotics and autonomous driving applications. Benefiting from the recent advances in deep learning techniques, the perform
ations imposed by sensor accuracy. By integrating Kalman filtering, optimization, and learning-based modalities, we propose a
mpares (key)nodes between sequential and non-sequential sub-graphs for place recognition tasks as opposed to a common fr
olonoscopies, despite significant shape and illumination changes. The map is a graph where each node codes a colon location
ity and safety of surgical tools, such as catheters and endoscopes, in complex anatomies. The lack of existing localization syste
hods may be limited by representation capabilities, while pairwise similarity-based methods require exhaustive searches, which
creasing attention in the field of autonomous driving and mobile robotics. However, there are still many challenges (e.g. difficulti
this paper, we address the problem of mapping the environment using LiDAR point clouds with the goal to obtain a map repres
or robot localization in various dynamic environments. However, the extreme scarcity of underground features introduces false
ers. When a pre-built 3D map is not available, the techniques of visual simultaneous localization and mapping (SLAM) are typic
and autonomous driving. It estimates 3D camera position and orientation (6-DoF) from images. Unlike traditional methods like
c applications. We focus on LiDAR-based point-to-point ICP combined with effective semantic information. This paper proposes
but their accuracy is unstable for different queries. Uncertainty-aware APRs provide uncertainty information on the estimated p
previously mapped area. Obtaining the pose from a single image enables many robotics and augmented/virtual reality applica
es have usually been applied to solve localization problems. However, due to the repetitiveness and symmetry of the spaces, vi
using mobile cameras. However, the city-scale Lidar maps generally need to be compressed for storage efficiency, which incre
heir potential energy-efficiency, low latencies, and capacity for continual learning. While these capabilities are well suited for rob
r, the role of overlap in 17-point algorithm remains unexplored. And the relaxed way in solving constrained normal equation lea
PR approaches are environment- and task-specific: while they exhibit strong performance in structured environments (predomi
ves the state of the art in performance and can be run in real-time on single board computers without GPU acceleration. Such c
cs. Conventional keyframe-based approaches often suffer from sub-optimal view- points due to limited field of view (FOV) and/o
despread attention recently. Although many current systems are applicable in daytime environments, nocturnal visual localizatio
various GPS-aided Visual-Inertial Odometry (GPS-VIO) fusion algorithms are proposed in the literature. This paper presents a n
e proximity to humans. To enable autonomy, the nano-UAV must be able to self-localize in its operating environment. This is a
the Lund University Vision, Radio, and Audio (LuViRA) Dataset. The dataset includes color images, corresponding depth maps
lly, we consider the Dual-IMU system, where each agent is equipped with one IMU, and employs relative pose observations be
d Ranging (LiDAR) sensors to achieve high-precision localization results for navigation purpose. However, this combination wit
y and perform practical missions. Kalman filters (KFs) and their variants are typically used in aquatic robots to combine sensor
entary benefits of range-based sensors and visual-inertial odometry (VIO) for localization. We use two fixed-lag smoothers (FL
rely on the tried and tested extended Kalman filter (EKF) to provide the state estimation solution for the inertial navigation syste
y localization of wearable devices. A primary challenge in this domain stems from non-line of sight propagation of radio waves.
ke WiFi, often supplemented by building floor plans for increased accuracy. However, the limitation of floor plans in terms of av
ect-based maps and camera images. In global localization, or relocalization, using object-based maps, existing methods typical
ess the position estimation problem when dealing with sharp-peaked likelihood functions derived from multiple observations. Sh
robots. In this letter, we address the problem of robust localization in cluttered indoor environments like office spaces and corrid
ots in ball sports, particularly in sports like tennis characterized by high-speed ball movements and powerful spins. The Magnus
a Doppler-capable range sensor. It leverages the radial velocities measured from the scene, estimating the sensor's velocity fr
ar range sensors that produce 3D point clouds where all points also have information about the radial velocity relative to the se
oint cloud, which yields significant advantages in initialization and relocalization processes. In the point cloud domain, the topic
y in nonlinear optimization-based Visual-Inertial Navigation Systems (VINS). Estimating uncertainty in real-time is crucial for eva
narios. The method combines Dynamic Movement Primitives (DMP) with Preference Learning (PL) to generate online trajectorie
igation. In response to the increasing deployments of mobile robots within various scenarios, a data-driven paradigm of naviga
meters to ensure safety-critical control in non-stationary environments. Our approach consists of two phases. The initial phase is
g adequate representation and accurate dynamics prediction due to the objects' infinite degrees of freedom. This work proposes
ement learning (RL) - known for its strong task performance and flexibility in optimizing general reward formulations - with the ro
autonomous driving. However, effectively integrating these models with downstream planners and model-based control approac
ormable object manipulation, such as rope and cloth, often required hundreds of real-world demonstrations to train a manipulati
robot. This is a particularly challenging problem due to the need for whole-body coordination in order to lift boxes of varying we
ponse to dynamics changes in different operational conditions. Hand-designed dynamics models may struggle to capture mode
haviors in the real world. On one end of the spectrum, we have model-free reinforcement learning (MFRL), which is incredibly f
age. As such, developing controllers in simulation is often desirable. However, robot dynamics in unstructured off-road environm
ng (RL), some of which can even operate on complex image observations, run directly in the real world, and incorporate auxiliar
els on out-of-distribution (OOD) states. We achieve this by leveraging two key sources of structure often present in robot dynam
any modern ML systems. What would it take to enable such capabilities in robotic reinforcement learning (RL)? Offline RL meth
ers and steep slopes. Such regions are ubiquitous and present science-rich targets potentially containing information regarding
ental tasks. To ensure safety during and after training process, existing methods tend to adopt overly conservative policy to avo
accomplish real-world manipulation tasks. Yet, limited research exists that explores the synergistic integration of different manip
obots to complete various practical applications. While certain goals can be achieved by picking and placing the objects of inter
ce goals using a text-conditioned vision transformer. Given a single RGBD image and a text prompt, ForceSight determines a t
y its ability to enable nuanced real-world interactions. In this study, the retrieval of unknown objects from confined spaces, whic
as important potential applications in manufacturing, agriculture, and surgery. In such environments, the task may involve threa
es and intricate functionalities associated with articulated objects. Point-level affordance, which predicts the per-point actionable
uttered and constrained environments. If the robot does not have collision-free access to the object-of-interest (OoI) which it int
a projectile moving toward a robot manipulator equipped with a shield. To successfully perform this task, the robot needs to (i)
on (HD) maps, hindering scalability. In contrast, Standard Definition (SD) maps are more affordable and have worldwide covera
achieve high-fidelity surface reconstruction at a low memory cost, compared to traditional explicit representations. However, st
licit neural mapping has shown remarkable reconstruction quality. However, incremental large-scale mapping with implicit neur
ments in neural radiance fields (NeRFs) have shown promise in synthesizing photo-realistic images. Several works have utilized
for 3D point clouds. Different from representing each type of primitive individually, we focus on the challenging problem of how
he presence of dynamic objects in 3D point cloud maps generated from scan sensors can introduce map distortion and long tra
al twins applications. As Neural Radiance Field (NeRF) greatly improves the mapping performance, in this paper, we propose a
g when object CAD models are unavailable during inference. We propose a framework that can reconstruct high-quality object-
range from metric, which require precise geometry-based optimization, to purely topological, where image-as-node based grap
bstract 2D maps, such as floor-plans or hand-drawn sketches, and use them to navigate in unseen rich 3D environments, witho
cles in dynamic environments. Current particle-based maps describe the occupancy status in discrete grid form and suffer from
namic objects in their surroundings for mapping, localization, and planning. It is key to reason about moving objects in the curre
y robotics systems. Surface normals estimated from LiDAR data are commonly used for a variety of tasks in such systems. As
various fields, including robotic navigation and scene reconstruction. This representation is now commonly used in 3D reconstr
r localization. Although there are many investigations regarding the uncertainty of the pose estimation of an ego-robot with map
ensors, termed D-Map. The framework introduces three main novelties to address the computational efficiency challenges of oc
d traces left by dynamic objects and manifested as obstacles on map, impeding robots from achieving accurate localization and
taneous Localization and Mapping (SLAM) systems. To achieve this, the gold standard is Bundle Adjustment (BA). Modern 3D
ten limit broader application. This research presents a pioneering micro autonomous underwater vehicle (µAUV) design. This
llows them to freely transform between the smooth flapping and the motion of rapid response so that they can move with dexte
designed to withstand the high pressure in these depths. In order to reduce the costs for sediment sampling in the deep sea a
orming subsea operations that require precision manipulation. While there is high demand for fully autonomous or even semi-a
zha-F, that can fly in the air, perform vertical profiling underwater, and vertically take off and land from both the water surface a
ents, using their undulating tails and pectoral fins for propulsion and maneuverability. Replicating these abilities in robotic fish is
e underwater biomimetic robot, especially when it comes to the narrow space operation. However, these aspects are contradic
al robotic platform, namely SILVER2, provided with a sediment sampling system. This approach tackles the disadvantages of st
al to traverse diverse terrains with unknown deformable properties, which can lead to unknown leg-terrain interaction forces. Ho
cluttered environments is a problem that has yet to be addressed in depth. Due to advances in technology and computational c
ages for mobile underwater vehicles. We formulate a deep learning model that fuses sparse depth measurements from triangu
tervention tasks. Nonetheless, underwater target pose estimation is remarkably challenging due to, among many factors, limite
through a novel method for sonar image correspondence using learned features. We introduce SONIC (SONar Image Corresp
in marine robotics contexts such as waste management and subaquatic infrastructure oversight. However, a central challenge
which complicates the study of benthic environments in marine biology and geography. In this research, we propose an algorith
on and scene parsing for AUV navigation inside underwater caves. We address the problem of scarce annotated training data
surveys to assist marine biologists in collecting datasets for understanding the ecological relationship of fish and other organism
n of 3D data. Previous methods usually use hierarchical tree data structures for encoding the spatial sparseness of point clouds
rwater tethered robotic applications in order to take into account the influence of hydrodynamic damping. To this end, an augme
th the free surface using a finite-dimensional dynamical plant model. Although finite-dimensional plant models of fully submerge
of the hybrid aerial underwater vehicle (HAUV). In this article, the surfing algorithm is proposed to search for the dynamic windo
hicles (AUV). Most AUVs use a set of thrusters, and optionally control surfaces, to control their depth and pose. AUVs utilizing t
e regions, and deviations could endanger vessels. Model Predictive Control (MPC) stands as a popular choice for trajectory tra
ode operations depends on only one type of propulsiondevice and can work flexibly in the amphibious environment. Robots tha
-air. This is partially because we enclose most underwater cameras in a watertight enclosure, such as a hemispheric dome win
However, in the underwater domain, most registration of multibeam echo-sounder (MBES) point cloud data are still performed
ator systems (UVMS) has garnered significant attention within recent years because of the opportunity for these systems to red
ems (MRSs): a non-linear dynamical opinion process is used to model high-level group choice, and multi-objective behavior op

n for recharging saves time and total cost for offshore monitoring and data collection applications. In order to facilitate AUVs for
ention on the International Regulations for Preventing Collisions at Sea (COLREGs) have been proposed in recent years. Howe
vironments. SLAM, commonly used to mitigate dead reckoning errors, still fails in feature-sparse environments or with limited-r
surveillance, patrolling, marine environment monitoring, and hydrographic surveying. However, most typical AMVs cannot meet
s in order to provide more information for human operators or automatic target recognition algorithms. In this work, we address
ulic structures. Underwater manipulators are essential devices that are used to carry out such maintenance tasks. For delicate
with popular marine robotics components, namely a BlueRobotics BlueROV2 with its standard Ping360 Mechanical Scanning S
of Autonomous Underwater Vehicles (AUVs). Due to a historical focus on underwater docking to only static targets, the researc
ance, is a fundamental task in navigation. This is particularly relevant during maneuvering in the harbor or navigating in confined
r human intervention. A docking station (DS) can extend mission times of an AUV by providing a location for the AUV to recharg
e vehicle (ASV) to operate in narrow waterways without relying on GPS. The proposed method uses a camera and a light detec
underwater visual-inertial odometry. The model is self-calibrating in real-time and is free of known correspondences or calibrati
pled fusion of magnetometer measurements. A sliding window of keyframes is optimized by minimizing re-projection errors, rela
derwater robot localization. Specifically, we address the problem of the global position of the underwater visual-inertial navigatio
lications. We propose a technique for mapping of underwater environments, building on the finite, positive, sonar beamwidth. O
manipulators (EAMMs), consisting of multiple rotor-actuated links with joints that can be either locked or unlocked. This joint-lo
drill operations for wood health inspection. The platform features an innovative passive prismatic-gripper mechanism affixed to t
quaternions has not been explored so far. This paper introduces a new innovative control strategy using dual quaternions for U
ue to underactuation; however, fragile payloads may require a consistent posture. The conventional method is also cumbersom
plications only focus on passive visual tasks. Aerial interaction aims to execute tasks involving physical interactions, which offer
e flying principles of insects and birds. Autonomous take-off is an important capability for FWAV to enhance its performance an
been investigated to elongate operation time. This study presents a new aerial robot capable of both perching and unperching
simplifying and speeding up complex tasks, such as construction, package delivery, and search and rescue. These systems, w
d take-off method for a multirotor unmanned aerial vehicle (UAV) using a suction cup perching mechanism on vertical surfaces.
ns in the mountains. The slower descent speed will affect the UAV's working efficiency in reaching the rescue area. However, d
e-suspended beam-shaped load by means of two aerial robots not explicitly communicating with each other. In particular, the w
ring by providing data collection capabilities at high spatial and temporal resolutions. However, limited flight endurance hinders
s of aerial manipulator. However, due to the presence of floating-base disturbance from the UAV platform and the kinematic err
Micro Aerial Vehicle (MAV). We propose a lumped drag model that concisely describes the aerodynamical forces the MAV is su
e aerodynamic framework, called phi-theory, for modeling dual-engine tail-sitting flying-wing vehicles for optimization-based con
ostatically-accelerated ions and neutral air, is a potentially transformative method for indoor flight owing to its silent and solid-st
with curvature constrained least-cost paths in a plane with prescribed initial and final configurations, different bounds for the sin
 cost metric. A third party ground risk model is used to generate a non-uniform costmap for a modified A* heuristic search. The
ge of quadrotors. Suction cups are usually adopted for surface attachment due to their durability and large adhesive force. To s
cles known as the Monocopter is proven and shown to be differentially flat. By exploiting this phenomenon, trajectory tracking c
e testing (NDT) applications has seen significant growth. Such operations entail physical interactions between the aerial robotic
he limitations of traditional tiltrotor vehicles. PADUAV features a unique mechanical design that incorporates two off-the-shelf qu
g on the manifold, S3, which is crucial in robotics and drone applications. We introduce a particle filter with stable embedding th
obotics. The state-of-the-art approach leverages reinforcement learning (RL) to train optimal neural policies. However, a critical
al quadrotors feature cylindrical rolling cages that are rotationally decoupled from the robot's main rigid body. In contrast, the pr
ying a compromise of mission-specific costs and constraints. When agility is required, shape-shifting (morphing) drones represe
es. To mitigate the effect of such disturbances, these critical issues should be considered: efficient disturbance observation and
loop (CL) equilibrium attitude-error quaternion (AEQ) of an uncrewed aerial vehicle (UAV) during the execution of high-speed y

pections. Underactuated aerial vehicles are widely employed in aerial manipulation due to their widespread availability and rela
often involves hazardous and expensive sling-loaded helicopter operations. In this work, we challenge this paradigm by propos
es. To perform specific tasks, robotic arms with various mechanical structures will be mounted on the drone. It results in sudden
stics or automated construction. At the same time, collaborative transportation of payloads by multiple drones is another importa
ecades, offering great maneuverability with mechanical simplicity. However, the under-actuation characteristic limits its applica
ng an underactuation approach, we demonstrate the successful oscillation damping of the proposed platform, modeled as a sph
s effectively navigating in close proximity to obstacles, while accounting for dynamic risk factors such as weather conditions, co
gnetic-gear-based actuation module suitable for small-to-mid-sized mechanical and robotic applications. The proposed actuatio
eases operating costs. Electrostatic clutches have the potential to improve robotic performance by generating holding torques w
the folding mechanism of the hind wing of ladybirds and the theory of origami, the motion model of the bionic foldable wing is e
driving mechanisms that can handle diverse environments. Bristle-bot-like robot designs, mainly studied over the past decade, a
enefit they provide in terms of safety and performance. Despite these assets, VSMs remain fairly complex mechanical devices l
d by this modular transformability, this study proposes a modular concept for growing robots that will be able to grow by adding
h quasi-direct drive (QDD) actuator for highly dynamic robotic applications. A comprehensive review of the mechanical design
asing year by year. Rehabilitation exoskeleton can help patients to carry out rehabilitation training and improve their activities o
the flexible omnidirectional driving gear mechanism. It is a gear mechanism having two degrees of freedom on one surface. Th
With the reconfiguration feature, the proposed gripper has a configuration suitable for different working environments of logistic o
odular robots have by using Attach-Lock-Detachable Magnetic Couplings(ALDMag) to give the ability to detach and attach the
obot that achieves configuration maintenance by utilizing active cables and passive elastic deformation of kirigami-origami. The
l arms to transport objects. To address the challenges of traversing in diverse environments and transporting objects, a novel tr
DOF ankle joint provides multiple benefits, such as increased stride length and walking speed, impact mitigation, and enhanced
ta of a certain water area, using autonomous sailboats to tow floating detection equipment is an energy-saving and convenient
AR) sensors used in outdoor unmanned vehicles. Bio-inspired wiping motion was figured to have more efficient and excellent wi
of a two-link robotic arm with remotized actuation via rolling diaphragm hydrostatic transmissions. We propose a dynamical mod
is to enhance the adaptability of vehicles engaged in ground coverage tasks, such as floor cleaning. The robot exhibits two dis
control method to diversify the vibrational behavior. First, the analytical model of the proposed actuator, which comprises a recta
rescue, medical, and industrial applications. Equipping them with grasping capability would enhance their usefulness in constra
ighly resilient and fault-tolerant locomotion strategies to navigate complex terrains by using multi-functional appendages. For ex
attention. Among the common living materials, engineered skeletal muscle tissue (eSKT) has been the focus of researchers due
 due to strength, motor control, and proprioception deficits in their lower limbs. Existing therapies involve repetitive StS exercise
d control of extracorporeal ultrasound (US) guided magnetic catheterization for arterial intervention. One uniqueness lies in tha
within living organisms with complex and unstructured environments. Potential applications include drug delivery, diagnostics,
ve intima located in the depth of the pulmonary artery. This work proposed a novel endoscopic robotized steerable dissector for
orm delicate instrument maneuvers with sub-millimeter accuracy. Minimizing the manual tremor and achieving precise and repe
ait cycle progression and command smooth and continuous actuation. In this work, these global tibia kinematics define a phase
of age-related macular degeneration. Various robotic systems have been developed to achieve the required precision, and in c
olon cancers with high sensitivity. However, current colonoscopy robots are often characterized by intricate and bulky mechanic
heater, with evidence of growing uptake in recent years. Until now, the da Vinci surgical system (Intuitive Surgical Inc, Sunnyva
visits in remote areas, with decreased costs for patients and infection risks. In particular, in ultrasonography, robots have the p
with a master-slave operated configuration. This slave manipulator features a modular distal continuum section, comprising two
o its lack of non-ionizing radiation and prevalent availability. However, significant inter-operator variability and inconsistent imag
primarily microscopes, for real-time instrument monitoring and motion planning. However, novice surgeons struggle to estimat
en as a potential solution to overcome the limitations of conventional US examinations, such as inter-operator variations. Howe
autonomously with a robotic ultrasound (US) system. In obtaining this view, it is necessary to avoid the ribs and lungs because t
r for the preliminary screening program, due to the advantage of being radiation-free. However, automatic segmentation of sma
(CMF) surgery. However, there are inherent challenges and risks associated with ensuring the recuperation of occlusion, safe
ve force feedback and drilling status information, reducing the operating challenges and psychological burden of doctors, which
ace significant challenges in maintaining precise control of hand-held instruments without damaging the fundus tissue. Typical t
s vertical direction and neglected the milling cutter's feed force. Additionally, achieving stable force control in multiple directions
nds to be longer due to high motion scaling. To address this issue, we developed a new controller that can provide the accurac
hile avoiding critical anatomical structures including nerves and vasculature. While image-guided surgical navigation is the prev
fection of crowded indoor public areas, leading to the development of automated disinfection robots. However, most of the exis
ancer cells. However, there are two limitations in robot-assisted lymphadenectomy: 1) lymph nodes are not visible during opera
n minimally invasive surgeries. Most existing robots of this kind depend on visual servo systems and struggle with efficient, rap
motion advantage after passing through a long and winding working channel. However, due to the interaction force between th
n, and improve surgical accuracy. However, the leader side outside the operating room still has problems such as incomplete c
precision and safe manipulation of surgical devices through the insertion point into the patient’s body. One of the RCM con
amplifier. The proposed model incorporates a lever mechanism into the Sarrus linkage structure. It allows the target plate to m
tions because they can be actively and remotely navigated through complex in vivo environments. However, the deformation ca
e targeted therapeutic delivery to the lesion site in human intestine can be operated by the external magnetic field. The robot is
brain, and low operability and fixed viewing angles limit its development. This study introduces a novel dual-segment flexible rob
ery due to their inherent dexterity and reachability. Nonetheless, their flexible structure and non-linear curvature pose significan
ies. Current approaches to account for inequalities rely on branching methods such as if-else statements. It can introduce disco
k of missed lesions and the often prolonged transit times, making it an unviable alternative to standard colonoscopy. In this lette
hey could, for example, navigate inside the bloodstream of a patient toward an occlusion and remove it. Magnetic rotating swim
on the the verge of a new wave of innovations. However, autonomy in RAMIS is still in a primitive stage. Therefore, most surge
an AI-assisted dynamic tissue evaluation is proposed for early bowel cancer diagnosis. Dynamic signals from a self-propelled v
hich uses ionising radiation to visualise catheters and vasculature. Prolonged Fluoroscopic exposure is harmful for the patient a
The new robot combines the advantages of both serial robots and parallel robots, featuring compactness, high accuracy, and s
micro-surgeries such as retinal procedures. Significant advancements have recently been achieved to increase the precision of
precise retinal surgery, such as retinal vein cannulation (RVC) and epiretinal membrane peeling. Robot-assisted eye surgery tec
ration for robot-assisted microsurgery (RAMS). A constant, small motion scaling ratio can enhance the precision of teleoperatio
gery but are devoid of surgical or therapeutic capabilities. Consequently, despite lesion detection, physicians frequently resort to
AGS) is a promising technique to prevent clutter in surgical workspace and provide a larger vision field. Existing camera design
such as gastritis, peptic ulcer disease, and gastric cancer, impacts approximately 50% of the global population. The efficacy of
can protect surgeons from radiation exposure. For designing an ESR manipulator, the main challenge is controlling the soft surg
ated for shape reconstruction and force estimation of flexible surgical robots. However, most existing approaches need precise
arcinoma being the most prevalent histologic type. Surgical resection is a primary treatment modality, and precisely identifying
d task of manipulation and cutting of an unknown deformable tissue (U-DT) between a remote surgeon and a surgical robot. Pa
uman errors, which are inevitable in long microsurgery interventions. Robotic systems developed in recent years for microsurg
performing complex vitreoretinal procedures such as subretinal injection. However, in order to intuitively integrate robots into th
d instruments. It often leads to prolonged patient recovery and interrupted operations due to incorrect instrument selection and
ng to its non-invasive nature. However, ultrasound beams focusing and steering is still challenging owing to aberrations induce
rsuing minimal invasiveness. Most conventional TNS tools are rigid and straight, limiting the dexterity and accessibility in passin
nsable process. However, even experienced physicians often encounter difficulties in manually manipulating the guidewire for b
r potential to rupture, leading to life-threatening internal bleeding. Current treatments often involve delivering embolic materials
ior of Tendon- Driven Continuum Manipulators (TD-CMs) featuring discontinuous cross-sectional geometries (i.e., having flexu
poses as an effective alternative to more invasive surgical procedures. However, the outcome of needle-based approaches reli
itation trajectories for dynamic parameter identification, emphasizing self-collision avoidance. This addresses the system identif
rticles. However, there is always a lack of a flexible and efficient method to freely create the desired magnetic soft microrobots.
spring. This novel actuator design utilizes two permanent magnets and 3D-printed magnet holders to effectively apply a tensile
their tiny sizes and attractive robotic properties. However, it remains a huge challenge to integrate sufficient functionalities in a
t actuation modes. By aligning two magnets with slight variations in magnetic moment orientations, we create a net magnetic m
ministration, precise surgery, and so on. The outstanding challenge is to realize power supply, propulsion, and steering mecha
elicate nature of the tissues involved requires the development of cutting- edge robotic technologies to enhance precision and s
chers. So far, selective control within multiple/swarm magnetic microrobots has been achieved with many strategies, such as u
ularly in embryology. These techniques are most often performed manually by highly skilled scientists, posing limitations on spe
propriate control strategies due to either the lack of sufficient information for feedback or the difficulty in extracting subtle yet cr
otic manipulation. Despite the well-established protocol and progress at macro scale so far, there remain challenges to further
r robotic micromanipulation. Due to the unavailability of depth sensors such as lidars in micromanipulation setups, traditional de
n the precise manipulation of microparticles and microorganisms on a micron and nanometer scale. This technology significant
esearch. In this article, a robotic framework based on micromanipulation system has been proposed, which can automatically a
ry and life science. Improving the precision and bandwidth of the micromanipulator and simultaneously reducing size, weight, a
ld of robotics research. Among the methods for fabricating the transmission mechanism of the insect-scale robot, the smart com
lly invasive diagnostic and therapeutic tasks. The orders of magnitude of forces and torques determine what functionalities thes
essing fertility potential and guiding visual feedback for microrobotic manipulation. Automated analysis of multiple sperms and s
copy, engineered to automatically reposition the microscope for high-magnification target view across multiple magnifications, th
e voltage clamp (TEVC) experiment under stereomicroscopy. We firstly developed a sequential calibration method to correlate
M robotic system at the ISIR laboratory. The method allows extending the range of an Atomic Force Microscope (AFM) and dea
intracytoplasmic sperm injection (ICSI) of in vitro fertilization (IVF) treatment, a motile sperm needs to be immobilized by glass
morphology analysis of a high number of sperm requires accurate segmentation of each sperm part and quantitative morpholog
ver, point-by-point raster scanning in PAM suffers from low imaging speed. Sparse sampling has been studied in recent years a
play an important role in biomedical fields including embryo engineering, gene engineering, drug screening, and cell analysis. H
eriodic magnetic fields, has been hindered by the limitations of wireless manipulation systems or noninvasive imaging technique
orms with dynamic camera viewpoints, designed to pick up unknown objects from cluttered environments. We address the lack
pproach to reducing complexity is introducing additional structure into the decision process, such as mixed-observability, factore
e to their superior potential in industrial and service applications. However, the high degree of freedom associated with mobile m
robots operating in unstructured human-centered environments frequently also have to manipulate the environment to their ne
objects on its end effector without grasping them---known as the waiter's problem---and move to a desired location while avoid
pening doors or for pick-and-place tasks. When used in real-world environments, adapting the robot controllers to uncertain con
cant attention within the robotics community. A critical challenge inherent in mobile manipulation is the effective observation of
ely expand the workspace of robotic arms, providing more possibilities for robot application scenarios. Compared with wheeled
to the enlarged space in which they can move and interact with their environment. However, even when equipped with onboard
afely controlling the robot in complex and dynamic environments. Moreover, in presence of variations in the operating condition
by exhibiting high degrees of freedom and severe self-occlusion. State representation for materials that exhibit plastic behavior
ertain level of autonomy while commuting in a car. However, emergency and high-speed maneuvers still arise as significant cha
applications, but it also presents challenges due to the complexity of vehicle-terrain interactions. In such environments, it is cru
and robustness in complex real world settings. For a wide range of cutting-edge robotic systems, such as micro- and nano-scale
or online identification of rigid body dynamic model parameters with emphasis on the physical consistency of estimated inertial
gh emerging applications, like automotive wire harness installation, introduce constraints that have not been considered in prio
us payload transportation. Resorting only to classical methods to reliably model a quadrotor carrying a cable-slung load poses
nomy applications. When turning at high speeds, these robots tend to undergo significant skidding and slipping. In this work, us
ns. Such motions contribute to better predictions of robot motions by humans, thus leading to more intuitive interaction and incre
used by migraines and stress. Existing robotic systems have focused on massage therapy on the torso and limbs, but performin
e prediction is essential to maintain safety in emergency scenarios. However, these emergency scenarios are long-tailed and h
for extensive setups, yet its application in human activity recognition is underexplored. This is particularly crucial for human-rob
y. This paper aims to model human grasping skills and transfer the learned skills to robots to improve grasping quality and succ
the ergonomics of an activity. Existing digital human modeling tools can simulate different morphologies at work, but hardly sca
uman motor functions, developing embodied intelligence, and optimizing human-robot interaction systems. However, current op
cations. To achieve high accuracy, existing works usually take computationally intensive models as backbone for exhaustive un
k improves rehabilitation using assist-as-needed strategies. Muscle forces are expected to provide even more comprehensive
er jamming (LJ) has emerged as a simple and efficient technique to achieve tunable stiffness for continuum robots. Despite its
t carrying a catheter through its hollow core. A simulation framework based on the model is developed in MATLAB and is used
driving soft robots because of their superior performance. However, their control is not easy due to their hysteresis characteristi
models and model-based controllers challenging. The most widely used model in soft robotic kinematics assumes Piecewise C
m well suited for constrained environments. Solving the inverse kinematics (IK), that is finding robot joint configurations that sat
ent compliance and lack of need for exposed hardware. Nevertheless, these advantages come at the cost of considerable cont
g accurate planning and control paradigms. The existing methods that create dense point clouds from camera images, and/or u
me for the dynamics learning and tracking control problems of a soft trunk robot. Specifically, a low-order approximate model de
ontrol their movements with the limited amount of control sources available. This task is challenging for reasons including the h
or thanks to their non-invasive nature. However, the active force was only used for the voluntary movement following. As a pot
ons. Their nonlinear behavior motivates the use of model-based controllers. They require state estimation as an essential step
 hrowing life-size objects toward target positions. SofToss is an open-loop controller based on deep reinforcement learning that
y control the shape (configuration) of arbitrary 3D Soft Robots in VR. Our pipeline, Virtual Reality Soft Robot Inverse Kinematic
comotion and manipulation. In the literature, there are now many concepts, often bio-inspired, to propose new modes of locomo
Our experimental platform is the previously presented MAMMOBOT, which is a small-diameter eversion growing robot with an e
mation and proprioception are challenging due to their complex deformation behaviors and infinite degrees of freedom. Their co
gh-dimensional dynamics. In this work, we present a data-driven modeling framework, based on geometric mechanics (also kn
and soft-rigid hybrids, most works examine robots that have an inherently serial structure. While these systems have been prev
tic characteristics of human specimen hands. The rig features eight force-controlled motors linked to muscle tendons, enabling
ng speeds. Kinematic and kinetic data for walking at four walking speeds were collected from five able-bodied individuals. The
access due to steep terrain features. Robots equipped with microspine grippers are particularly well-suited to ascending rocky c
ry and transport heavy loads during various climbing tasks. However, many dry adhesion-based legged robots prioritize foot de
uced by arm motions must be treated carefully. In general, it is a significant challenge to establish a fixed-base frame during tas
ools for locomoting systems. However, due to the curse of dimensionality in coordinate optimization, a key component of geome
enging task for legged and wheeled mobile robots. To accomplish such a task, online planners shall incorporate new terrain inf
otic tasks. Signal temporal logic (STL) has been widely used to systematically and rigorously specify these requirements. Howe
marine robotics and ship routing, can be modelled as instances of the time-varying shortest path (TDSP) problem. Although ther
le distinct paths in 3D cluttered environments. Finding such distinct paths is useful in many applications. Among others, using m
dictive Path Integral control (MPPI), named Stein Variational Guided MPPI (SVG-MPPI), designed to handle rapidly shifting mul
The proposed method evaluates the existence of lines-of-sight from a source point to all other grid cells in a single pass with n
s. Clothoids, curves that exhibit linear curvature profiles, play an important role in road design and path planning due to their ap
robotics challenge with many applications. Despite significant recent progress on the topic, computationally efficient and high-q
cue and precision agriculture, where consistent presence is needed to detect events over large areas. Ergodic search already d
g path planning problems for robotic systems. Various improvements to the RRT algorithm have been presented to improve the
s, making them suitable for navigating through, and interacting with, confined environments. However, the study of path plannin
by a cooperative group of robots. These robots must avoid collisions, deadlocks, and are subjected to the constraint of maintain
e in solving path planning problems. RRT* offers asymptotic optimality but requires growing the tree uniformly over the free spa
seconds to dozens of seconds to find collision-free motions for high degree-of-freedom problems. This paper presents performa
ulti-objective optimization to dynamically plan trajectories that satisfy multiple complex mission specifications. In the setting of p
rapid generation of purposeful robot motion in complex and swiftly changing dynamic environments. While much existing resea
strained robot, that is tethered at its start position, such that the tether satisfies user-defined winding angle constraints. A varian
ath plans in convex polygons, explores an additional parameter of these path plans that can be optimized, discusses the prope
utonomy has confined them to rigid, task-specific systems. The intricate configuration space of manipulators, coupled with the
d, densely vegetated outdoor environments using offline Reinforcement Learning (RL). Our method trains a novel RL policy usi
im to minimize a scene-specific cost of executing a given plan. This approach offers remarkable adaptability, as they can be di
ce is described as a collection of simple sets. For example, recently developed motion planners rely on a convex decompositio
tically searching an environment to detect (capture) an evader. Previous techniques to generate search strategies for the pursu
c field. Due to the restricted node-expansion directions, the resulting paths are usually not the shortest. Besides,unnecessary h
Previous work proposed an asymptotically complete motion planner that reports a plan or an infeasibility proof given long enou
which relies on the extraction of the Hamilton-Jacobi skeleton of the free space. This skeleton is used to construct a graph of fr
y navigate between multiple target locations in potentially challenging environments. Currently, this type of multi-goal mission of
increasing interest in applications availed by aerial vehicles. In this paper, we propose a landing safety pipeline based on state
 t of the race track, and the configurations of the gates (e.g., their shapes and sizes). However, previous studies neglect the con
. Most successful planners have two stages: a front-end that determines a path that incorporates geometric (or kinematic or inp
anned aerial vehicles with suspended payloads have advantages over other systems, including mechanical simplicity and agility
goal is to intercept a flying ball and volley it towards a designated target. Aggressive means BAT operates the quadrotor aggres
ground target using a multirotor with a cable-suspended load. The NMPC framework relies on the dynamic model of the UAV w
onitoring is important in search-and-rescue or ocean clean-up operations. Since the location of any detected object is dynamic,
ns, and efforts are further being made to explore this field. Physical interaction with aerial robots can provide an intuitive way of
aerial 3D printing framework. The experimental demonstration of the overall autonomy focuses on precise motion planning and
obotic manipulator (D-Arm) to address the issue of backlash nonlinearity coupling unknown disturbance through the dual-motor

hat does not rely on fault information subject to one, two adjacent, two opposite, or three rotor failure. The uniform control impl
to existing methods that are either computationally heavy, system/container specific or dependant on a singularity-prone pendu
s of the process. The development phase involving the instrumentation and operation of a full-size vehicle can be especially co
amic and uncertain. This paper investigates a data-stream-driven motion control problem for mobile robots to avoid randomly m
to external disturbance forces. While this question may not be central for large robots perfectly rejecting disturbances through
able-driven parallel robots (D-CDPRs). In contrast to the straightforward inverse kinematics of the rigid frame cable-driven para
s, making it a vital component of human-robot interaction (HRI). Manual intervention may cause uneven forces on the manipula
onments. To this end, we leverage control barrier functions (CBFs) with the forward reachable set to guarantee the safe coordin
ed with a roller arm, particularly when performing locomotive tasks such as autonomous robotic inspection in complex, multi-tier
ugh a hole in the interface separating the internal space from the external space. Considering that the control of RCM may be i
e controller and underlying system performance. Bayesian optimization has been established as an efficient model-free method
that are difficult for humans. However, the spatial locomotion control of these robots remains challenging due to the robots' larg
se two novel efficient and robust motion control schemes for high-dimensional robot manipulators. Both controllers incorporate
e positioning stages are usually introduced. Being mechanically attached, the motion of the combined positioning stages needs
ess to a controller with a simple implementation. As a nonlinear DOB for the robotic systems, we proposed so-called nonlinear r
bot shared control to enhance surgical convenience and efficiency. However, the neglect of human dominance may reduce sur
ch as in service tasks or industrial use cases. Mostly, classical impedance and hybrid motion-force control approaches are emp
roller to possess the capability to dynamically adjust its actions in real-time, enhancing its resilience in the face of fluctuating en
in various unstructured environments and situations requiring precision and force control. However, traditional robotic arms hav
of success. It can provide stability and safety guarantees while being effective in accomplishing many locomotion tasks. Mode
actor-trailer vehicles, especially when the parameters in the dynamics are unknown. To address this constraint, we propose a d
moves material. Object-shape transition models are essential to achieving automation by robots; however, learning such a comp
the environment, creates significant challenges for high-quality control. Existing model-based methods typically rely on either g
ems. While much work has been done on improving model formulation, no standard protocol exists for gathering empirical data
ble. This usually renders the system kinematically redundant: the main manipulation or interaction task does not fully determine
a problem are potentially logically linked together. To efficiently solve such puzzles, we develop a motion planning approach ba
botics due to their support for high-dimensional state spaces and guarantees of completeness and optimality. This paper introdu
cubes are uniformly controlled by a global time-varying magnetic field. A 2D physics simulator is used to simulate global contro
its remarkable flexibility even within confined spaces. However, its motion planning and control pose significant challenges. Thi
s vehicles. This paper presents a mathematical formulation for the optimal speed planning of autonomous driving, which has be
tain trajectories may not be tracked as single, continuous paths due to the robot's kinematic constraints or obstacles elsewhere
m explores an unknown environment to reach a target point, providing velocity commands to the robot controller module. CBFs
t and goal configurations on the kinematic constraint manifolds. These problems appear in various scenarios ranging from obje
hicle (UAV) with limited power and computation, useful for real-time operations where the UAV is exposed to significant disturb
ybersecurity issues such as vulnerability to adversarial attacks. In this paper we investigate the vulnerability of an important com
over the environment, as well as constraints such as the robot's limited battery capacity, makes the long-time horizon persistent
narrow passages in their configuration space. Our novel approach leverages optimization techniques to facilitate sliding and roll
or tethered differential-driven mobile robots. The motivation of this work is to generate self-entanglement-free (SEF) commands
uffer from expensive collision checking and high sampling complexity, which make them difficult to use in real time. To address
ed by either multiple fixed-base robots or a large robotic system. This capability is enabled by the mobility of the mobile base. H
e constraints for manipulation systems. Vector Quantized-Motion Planning Transformer (VQ-MPT) is a recent learning-based m
ound robot types and morphologies. Our two-step approach uses a sampling-based planner to compute an approximate motion
nned ground vehicles (UGV) in unknown off-road environments. However, the exploration efficiency of single UGV is low due to
e using only the incidental sounds they produce as they move, even when they try to be quiet. We collect a robotic dataset of h
ntends to interact: in this case, it can proactively enact friendly behaviors that lead to an improved user experience. We solve th
there are often group dynamics which can complicate persuasion. This study focuses on how non-verbal cues, specifically gaze
e of applications. However, it is still far from being as natural as human-human handovers, largely due to the robots’ limited
performing multiple tasks, such as point-to-point navigation, human-following, and -guiding. However, a prominent gap persists:
tile perception and feedback becomes imperative, effectively bridging the conventional sensory gap. In this paper, we propose

with the severe occlusion and depth ambiguity inherent in interactive hand gestures. Recent approaches that employ probabilis
ne-grained human activity recognition. In contrast with state-of-the-art methods which directly take sequences of 3D skeleton co
e robots by adding the capability to detect human contact. Unfortunately, simple binary tactile sensors alone cannot determine t
e to the inevitable ambiguity in natural languages. It requires robots to disambiguate the user’s input by active information g
objects in unknown environments. The existing works of ZSON mainly focus on following individual instructions to find generic o
ia human-robot natural language interaction. Current IOG systems assume that a human user initially specifies the target objec
d cost-effective manufacturing. However, simultaneously reconstructing force and shape from boundary measurements remain
infeasible in the context of interactive robotic agents utilizing large language models (LLMs). To tackle this problem, we first pr
ure, and expected to be able to improve group dynamics when interacting with groups of people in social settings. In this paper
orders during carrying or lifting tasks. Having sensors, electrical/pneumatic actuators, and control systems, active exoskeletons
tual reality (VR) scenarios. This paper proposes a robust tracking algorithm based on a novel wearable ring-shaped controller e
nd hold significant potential for aerial manipulation tasks. Differing from a multiple aerial swarm, the aircrafts within the IAP are
ment is a challenging task. The key point is the data association between different robots' viewpoints. It also makes traditional A
ccessful multi-robot operations in environments without an external positioning system, such as indoors, underground or underw
ngly popular in large-scale rescue and environmental monitoring tasks. However, due to the energy constraints of a single UAV
site for robotic swarms operating in GPS-denied environments, presenting a formidable research challenge. In this work, we pro
ow to obtain relative locations in a distributed manner is still a primary problem to address in multi-robot applications. In order to
arity due to their enhanced accuracy and stability from the increased field of view (FOV). However, integrating multiple LiDARs
n, high-precision mapping not only demands accuracy but also flexible construction. Current approaches mainly rely on expens
of a cable-towed payload. In particular, we aim to solve the challenge from the unknown dynamics of the cable-towed payload.
cute a set of dependent tasks as quickly as possible, and to complete tasks that cannot be addressed by individual robots, it is
., to monitor wild flocks, detect intruders, search and rescue. Monitoring numerous dynamic targets in a collaborative and efficie
exist in the same space, interact, and possibly collaborate. Methods to provide such systems with system size scalability are kn
algorithm for Consensus Simultaneous Localization and Mapping (CSLAM). Multi-robot teams require that agents have timely
hrough the augmented loss function, which is required in the context of reward and diversity. Usually, the diversity optimization
alisation is based on visual odometry (VO). Such drones should avoid flying into areas with poor visual features, as this can lea
ained large language models (LLMs) for both high-level communication and low-level path planning. Robots are equipped with
-- easy deployment, task generalization and real-time execution capability. Prior end-to-end DRL-based methods have showca
botic systems and successfully realized autonomous learning in many fields. In these fields, robots need to communicate and c
rained base policies. In the multiagent setting, performing online rollout can require an exponentially large number of optimizati
ire to obtain success comparing with other individuals or opponents. Although we might get anxious under the pressure, it could
nge of real-world scenarios. We focus on optical wireless communication (OWC), which is a practical solution to be used in vari
lthough they have been studied in agriculture, search and rescue, surveillance, and environmental exploration, real-world imple
their teammates, or the environment, robots must observe individual performance variations and adapt accordingly. We propos
each other can collaborate to globally localise via efficient ad-hoc peer-to-peer communication. Our Robot Web solution is bas
ntralized control in swarm robotics. However, it has been observed that relying solely on the states of immediate neighbors is in
t systems. Nevertheless, in relative pose estimation, time synchronization is usually underappreciated and rarely addressed, al
ternal infrastructure is challenging, since the software framework required to support complex tasks must be robust to unreliabl
onstrained Unmanned Aerial Vehicle (UAV) and an Unmanned Ground Vehicle (UGV). Both the UAV and the UGV have predef
om atoms to honeycombs, and have proven useful in artificial systems, from molecular containers to clay bricks. But, despite th
d quadrotors with multiple team leaders governed by nonlinear dynamics, employing the reinforcement learning. A cascade con
ons increases exponentially in comparison. To cope with this complexity, heterogeneous teams of robots have become a signif
navigate through crowded environments. Our framework comprises of a local model predictive control (MPC) for each robot an
nts and tasks. We devise a range of tasks using signal temporal logic (STL), each earmarked for specific agents. These tasks a
gement of objects by manipulators, results in solution times up to three orders of magnitude faster than existing methods and s
onomous spacecraft conducting close proximity operations. This is especially true for multi-agent inspection operations in which
em. Dynamic Targeting is an approach to this challenge that intelligently plans and executes primary sensor observations base
lti-robotic systems, this article introduces RecNet, a novel approach that concurrently addresses both challenges. The core of R
e Controllers (MPC) dynamically, that solve navigation problems for multiple robotic agents sharing parts of the same environm
ction of 3D structures which includes uncertainties like size, number, shape, position, using multi-robot systems (MRS). Given
e problem of multiple micro aerial vehicles (MAVs) in environments with static and dynamic obstacles. The framework utilizes s
ation for encoding and communicating each robot’s intent. This representation for “intent messages” enables impro
ame-theoretic belief space planning through partially propagating beliefs. Complex interactions in scenarios such as surveillance
, security surveillance or environment monitoring. Such algorithms often make restrictive assumptions: the number of targets an
ms, which takes into account map and localization uncertainties of range-sensing mobile robots. Virtual landmarks are used to
s teams of robots. When robots have heterogeneous capabilities and there are diverse objectives and constraints to comply wit
sion-making, which is often achieved using heuristics-aided methods such as genetic algorithms, auction-based methods, and b
nnovations, robotic development and automation efforts encountered in the mining industry---particularly in the Pilbara iron-ore
Guided Vehicles (AGVs) moving in a real industrial scenario, such as an automated factory or warehouse. The proposed metho
earch on Multi-Robot Task Allocation (MRTA) problems. In this paper, we propose a preference-driven approach based on hed
eam of aerial vehicles. Each agent in the network is assumed equipped with a camera with limited range and Field of View (Fo
s, improving the robustness and independence of fleet operations. Bearing measurements (obtainable from embedded camera
with the objective of approaching the uniform node coverage probability distribution by the robot team. Prevailing MuRP solutio
among the individuals. The individual robots lack braking capabilities and can only adjust their courses while maintaining their co
s in aerial, aquatic, and surface environments. In this paper, we study a simplified two-outcome surface inspection task. We tas
cks. The robots themselves are cubes of the same size as the blocks, and they may place, carry, or remove one block at a time
ects because of their anatomical similarity to the human hand. However, the large dimension of configuration space challenges
bimanual manipulation platform (OpenBMP) that is designed for research and rapid experimentation in robot grasping, dextero
roviding tactile data and mechanical compliance to robotic hands. If there are low-cost and easy-to-build robotic hands equippe
ce given an input target grasp. While optimizing grasp motion with two or three-fingered grippers has been well studied, the stud
s research. The collection of high-quality expert data holds paramount importance when using imitation learning. The existing ap
owever, it is difficult to mount skin, tactile sensors, and driving mechanism required for human contact, especially holding hands
objects or interact with the world. Because of their simplified design, they are unable to reproduce the dexterity of human hand
us robotic manipulation tasks through learning from offline data. In this competition, participants were given two types of datase
l functions of the hand to achieve a series of operations such as alignment, picking, reorientation, and positioning of the part. In
propriate multiple unmanned surface vehicles (USVs) is an urgent yet challenging mission with the tremendous development of
ms-based controllers, building upon and rectifying a significant limitation in a previously presented approach. The proposed fram
ormation control of first- and second-order holonomic and non-holonomic mobile robots. The distance-angle leader-follower form
ulate. The different aspects of a complete task such as local planning for obstacle avoidance, global goal coordination and colla
any industrial and robotics applications, while an important functionality is the advanced Quality of Service (QoS) features. Des
pace and cost-efficient design constructed as a rack of 32 small robot arm work cells. Each robot work cell is fully enclosed and
server options. To help roboticists make cost-effective decisions, we present FogROS2-Config, an open toolkit that takes ROS2
dedicated team of robots creates and sustains a wireless network that satisfies the communication requirements of a different
services, which are connected via a communication channel prone to delays and packet loss. The stability of the networked sy
en image, building on the Neural Radiance Fields (NeRF) formulation. IFFNeRF is specifically designed to operate in real-time
mproving overall accuracy, especially the downstream autonomous driving tasks. The recent development of low-cost LiDARs
3+1D radar severely suffer from the disorder and sparsity of point cloud. To tackle this problem, we propose a novel Multi-Task
age to 3D feature spaces, as well as the accuracy of object localization within the 3D space. This paper aims to address such a
ill challenging to make the object detector reliable in adverse weather conditions such as rain and snow. For the robust perform
ct safely with the environment. Most 6D pose estimators, however, rely on a single camera frame and suffer from occlusions an
ainties is of great importance. Three-dimensional (3D) object detection, a critical robotics operation, has seen significant advanc
stimation methods have restricted applications due to the heavy reliance on depth sensors. RGB-only methods provide an alter
enefit robotics applications. However, exclusive reliance on RGB images without leveraging any 3D information introduces amb
andard object-detection methods cannot identify unknown objects that are not included under predefined categories. This is be
e points into grid-like voxels or pillars. In this paper, we identify that the common PointNet design introduces an information bott
y and high dynamic range. One task where these benefits can be crucial is real-time object detection. However, RGB detectors
mented reality and domestic robots. Transformers as versatile network architectures have recently seen great success in 3D p
or future prediction. In the same way, history detections should contribute to future detections. In this paper, we propose a dete
system robust- ness and safety. Recent point cloud-based 3D object detectors are mainly categorized into three types based o
erein high-resolution and high-bit-rate videos captured by their cameras necessitate compression before further analysis. In this
detection in autonomous driving. Traditional point-based 3D object detectors often employ architectures that rely on a progressi
nsively perceive the surrounding environment. In the task, we notice that semantic, instance, and panoptic have rich relations, h
cross-modal hallucination. Our proposed approach is agnostic to either hallucination direction between LiDAR and 4D radar. We
e the reliability of robotic vision systems in open-world environments. Previous approaches to UDA-OD based on self-training ha
 egant paradigm for the 3D detection tasks. Current multi-modal fusion methods in BEV can be categorized into LSS-based and
al data to enhance multi-camera 3D object detection, leveraging the continuous and correlated nature of input video streams. R
has been previous work to make improvements specifically for small objects, such as adding several attention blocks or chang
lies on either human annotations or semantic segmentation with categorical labels. Both methods fail to match natural human c
f applications, including collision avoidance, countering hostile drones, and enhancing search-and-rescue operations. However
rs have achieved impressive performance, the high cost of LiDAR sensors precludes their widespread adoption in affordable ve
egment novel instances poorly, which is required for autonomous agents acting in an open world. To improve out-of-distribution
robots must be proficient in segmenting unseen objects from the background and/or other objects. Previous works perform uns
cue or package delivery by ensuring safe navigation in unknown environments. However, deconflicting trajectories with other ag
ons. Previous works have faced limitations such as the complexities of 3D object shapes, the wide range of potential affordance
ge-scale datasets. However, this requirement introduces the challenge of manual annotation, which is often both burdensome a
t work. However, due to the lack of a cohesive framework, these objectives often possess a narrow scope of applicability and a
onment, containing both static and dynamic information. Typically, these maps are primarily generated using lidar measurement
ccurately and timely, providing decision-making for obstacle avoidance and planning. Low-resolution LiDAR is one of the most
sly in dynamic environments. We thus propose Dynablox, a novel online mapping-based approach for robust moving object det
g 3D OD models have been tailored to work with light detection and ranging (LiDAR) and RGB-D point cloud data, leaving their
ata form. However, RV-based methods fall short in providing robust segmentation for the occluded points and suffer from disto
he segmentation and tracking of moving objects are essential for reliable path planning, including collision avoidance. We inves
ird's Eye View (BEV) in road scenes has witnessed great progress over the past few years.As a road scene exhibits a part-who

estimates in a single forward pass. We exploit general representations from foundation models and unlabelled datasets throug
epth estimation in a single model to achieve semantic and geometry perception simultaneously. DPS task has important applica
t by leveraging complementary information shared among various agents and sensors. However, practical applications often en
between a pair of points in an unoriented point cloud. CID indicates the likelihood of two points or two sets of points belonging to
s embedded in robots, one time step per processing time for algorithms can be considerably large. This scenario necessitates t
to unsupervised domain adaptation (UDA). Yet, without any target annotations, the performance gap between the UDA approa
onomous driving re- search field. Nevertheless, most of these methods heavily rely on dense LiDAR data and accurately calibra
s a formidable challenge, often requiring intricate mapping or multi-step pipelines. However, these conventional approaches str
y of radar and camera sensors has led to a growing interest in fusing radars and cameras for 3D object detection. However, pre
autonomous driving, where every pixel in an image is assigned a label from a specific semantic class. However, obtaining dens
mance gap of existing LiDAR-radar detectors. To improve the feature extraction capabilities from these two modalities, we desi
e consuming and must therefore be done selectively. Active learning (AL) provides a solution that can iteratively and intelligentl
an annotator to label the most informative samples from a pool of unlabeled data, and then training a model from these sample
Vs). Existing methods often lack research on multi-disturbed conditions. To address this issue, we propose a trajectory-predicti
e tackled by using four-wheel steering (4WS) vehicles which are capable of switching between multiple steering mechanisms w
erformance of robotic systems, including increased energy efficiency, durability, and safety. In this paper, we present a novel a
oblems. One key step in many algorithmic approaches to trajectory optimization is the solution of moderately-large and sparse l
d manipulation which leverages trajectory optimization at each control step. While the performance of this approach is computa
ies, a technology critical to their application in the wild. However, the computational burden of computing time-optimal trajectori
ast as representative task models for motion recognition and generalization. Currently, no invariant descriptor exists for represe
th. It is commonly used in the decoupled approach to trajectory optimization whereby a path is first found, then a retiming algor
 and multi-tasking behaviour, inspired by human nature. To this purpose, from a hardware perspective, a high number of degre
However, existing methods are limited in terms of reasoning about sources of epistemic and aleatoric uncertainty, space and ti
cs. By temporarily storing potential energy and subsequently converting it to kinetic energy, elastic robots are capable of highly
effectively exploit the systems’ capabilities are still missing. Differential dynamic programming (DDP) has emerged as a pro
it remains limited in contact interaction tasks due to its inability to leverage sensed efforts. In this work, we propose a novel mo
sed research effort in geometric methods for robotics in recent years. The results were algorithms using the various framework
a priori unknown state-dependent disturbance behavior. We propose a formulation that embeds the uncertain system into a hig
ety of nonlinear dynamical systems for arbitrary inputs. However, it is difficult to construct functions that satisfy the CBF constra
a multi-robot teleoperation system in a holistic fashion. The proposed approach consists of encoding these three properties as
or nonlinear dynamical systems. Building upon prior works, we demonstrate that SOS can synthesize dynamic controllers with
botic tasks, in which multiple contacts may occur at unknown locations. To address such scenarios, we made an explicit contac
nging open research problem. Granular material systems are complex-behavior (as they could have solid-, fluid-, and gas-like b
ombines Trajectory Optimization (TO) and Reinforcement Learning (RL) in a single framework. The motivations behind this alg
ous racing at the handling limit that can iteratively explore and learn unknown dynamics in high-speed operational domains. We
ded non-stochastic noise, i.e., noise that is unknown a priori and that is not necessarily governed by a stochastic model.We foc
rnal forces on biped robots in a desired level. We first introduce the authors` preceding results on the norm-based stability crite
tially linear, interconnected systems. In addition to the dynamic uncertainties resulted from the interconnected dynamic system,
e coupling between trajectory optimization, stabilization and design optimization. This approach proves particularly advantageou
gion of interest. The ergodic control formalism behind such an exploration behavior consists of matching the time-averaged stat
ble of solving high-dimensional control problems at real-time rates. ReLU-QP is derived by exactly reformulating the Alternating
of robots and treat robots from other vendors as general obstacles. The ability to enable robots from different vendors to co-ex
ollaboratively reach a designated destination. However, this task becomes notably challenging when the leader lacks complete
nment. The impeded environment is represented by a graph with select road segments(edges) in disrepair impeding vehicular m
bjective function based on some measure of information. Our prior work determined that the spectral decomposition of an inform
ns while respecting social norms is challenging, particularly with limited communication. Although current robot social navigation
erative autonomous driving at intersections has been studied by many works in control theory and intelligent transportation acro
n for mobile robot path-planning. In this context, mobile robots compress their current local maps to assist another robot in reac
collisions. However, few existing formation planning algorithms consider the wind fields of drones during planning. Since the win
ng. The problem involves determining informative regions in the environment that should be visited by robots to gather the mos
uman behavior is hard to predict, and its uncertainty impacts the performance of the robotic fleet. Various multi-robot planning a
ogistics and traffic management. We propose a novel approach that combines learning-based techniques with search-based m
dor (SFC) to facilitate static obstacle avoidance. Typically, SFCs are constructed through convex decomposition of the environm
Model Predictive Control (CB-MPC). Inspired by Conflict-Based Search (CBS), the planner leverages a modified high-level conf
with different dynamics, actuation limits, and shapes to reach their goals in challenging environments. We solve this problem b
from their current positions to their pre-set goals in a known environment, and is an essential problem found at the core of man
n workspace with multiple homogeneous mobile robots. Our algorithm is horizon-based, synchronous, and on-demand. The rec
robots in a continuous state space. The difficulty of MRMP increases with the number of robots and is exacerbated in environm
oth unweighted and weighted terrains, which aims to minimize the coverage time, defined as the maximum travel time of all rob
Deep Neural Network for processing onboard-captured depth images for collision avoidance in trajectory-tracking tasks with U
nned Aerial Vehicle (UAV) with a cable-suspended load in order to achieve a robust and fast estimation of the cable's state duri
ined operations near powerlines. The drone is equipped with a robust onboard perception and navigation system that enables i
 of quadrotor in obstacle-cluttered environment without relying on an explicit map. Our approach searches for feasible trajectory
ks such as building inspections, rescue operations, deliveries, and warehousing. However, there are two main limitations to prev
ormance drops when applied to slant-angled aerial images due to occlusion and scale variation. This is mainly caused by a lack
ection on unmanned aerial vehicles (UAVs) with a monocular camera and a limited Field of View (FoV) LiDAR. The point cloud
AM. It exhibits frame- processing rates above 60FPS on NVIDIA’s low-powered 10W Jetson-NX embedded computer and
cro aerial vehicles (MAVs) due to their limited onboard computing and sensing resources. Although various collision avoidance
systems often possess limited sensing capabilities with field-of-view (FOV) restrictions. To navigate this challenge, we present
ents the environment as a continuous probabilistic model. This model enables high-resolution reconstruction while simultaneou
onomous aerial robots with Size Weight and Power (SWaP) constraints. The SWaP constraints induce limits on mission time re
owever, existing methods have shown unsatisfactory system simplicity, computation efficiency, and path quality in large and co
aviors of robots in an indoor environment composed of multiple rooms. Contrary to simple frontier-based exploration approache
aster surveying. Unmanned Aerial Vehicles (UAVs) have become particularly popular in such applications, promising to cover l
l robots based on a modular approach exploiting deep collision encoding and reinforcement learning. The proposed solution bu
r quadrotors using Deep Reinforcement Learning (DRL). Previous learning-based methods relied on accurate knowledge of the
t and introduce a perceptive safety filter that leverages Control Barrier Functions (CBFs) in conjunction with Visual-Inertial Simu
maintaining flight stability. Recent research has worked to understand this flow by employing 2D and 3D flow sensors to measure
e accurate dynamic obstacle perception. Although plenty of previous works in the autonomous driving field have investigated th
tion accuracy during maneuvers, while the feature matchability among frames, a crucial factor influencing estimation accuracy,
enges and at the same time replace large and complex farm machinery, this paper proposes designing a farm around a hetero
ng aircraft, as has been demonstrated by human glider pilots for decades. In this work, we present a novel end-to-end deep lea
ctor introduces an attention module into the feature extraction module (FEM) for enhancing the accuracy. This FEM with fewer c
sion, localization, and mapping. However, existing correspondence methods do not scale to large multi-robot systems, and they
ment is a pivotal challenge. Traditional methods of grasp planning methods utilizing partial point clouds derived from depth ima
sthetic, aerodynamic, and safety reasons. Electronic skins (e-skins) capable of sensing contact location and pressure across co
m of applications, spanning from production lines to warehouses, and encompassing both stationary and mobile robotic system
domain adaptation problem. In this case, we present a global-to-local method to address hybrid domain gaps in RGB and dept
and manipulating these objects is a significant challenge for robots. Existing methods have difficulty reconstructing complete de
try, lack of texture, and severe self-occlusions. We propose an online 3D edge reconstruction framework that uses monocular
o address this is extruding viscous liquid from a squeeze bottle. In this work, our goal is to create a sauce plating robot, which r
and-object scene from a single RGB image. The inference as well as training-data generation for 3D hand-object scene recons
bstacle to automation lies in the difficulty of developing a reliable robotic grasping system for already harvested trusses. We pro

well-light conditions, repetitive experiments have found that the detection performance would deteriorate drastically under low-lig
e manipulation tasks. However, it remains a challenge for current taxel-based sensors to obtain HR tactile information. In this pa
robot to assist in common manipulation tasks, it must learn a representation of articulation. Recent deep learning methods can
much force that could damage the object. In practice, the force regulation is challenging to implement since it requires knowledg
cations. In this paper, we propose a new cross-pose estimation method that predicts correspondences on a set level as oppose
the prevalent methods for 6D object pose estimation heavily rely on 6D annotated data to train deep neural networks, which po
manipulating a Deformable Linear Object (DLO) within an unknown environment that imposes constraints on the DLO's movem
umans. Our innate grasping system is prompt, accurate, flexible, and continuous across spatial and temporal domains. Few exi
their suboptimal success rates and risk of crop damage, we design a novel bot named AHPPEBot which is capable of autonom
 rios. State-of-the-art methods for fully autonomous scenarios are typically learning-based approaches optimised for a specific e
emains challenging for robots to interact with liquids. These challenges are particularly pronounced due to the limitations impos
cked containers. To automate tasks in this environment, a robot must be able to pick diverse objects from highly cluttered scen
ten, they are 3D printed and do not include tactile sensors. Hence, performing in-hand object recognition with direct touch sens
challenging object interactions. In this work, we show that the density field created by a state-of-the-art efficient radiance field r
n an image-goal setting, the robot rearranges the current scene to match the single image captured from the goal scene. The ke
aking robots beyond pick-and-place tasks. We propose using an event-camera for tracking the objects to leverage the low-laten
omplex task from both the perception and planning perspectives. Indeed, the stochastic nature of the underlying environment dy
SE}_2(3)$ Lie group definition. Our metric is related to prior metrics for SLAM but explicitly takes into account the linear velocity
cticality through the use of pre-collected large datasets. Despite its practical benefits, most algorithm development research in o
and is still an open problem. Standardized benchmarks are important tools for evaluating and comparing robotic grasping and m
d, specifically focusing on a pick-and-place task. Our benchmark uses the YCB object set, a commonly used dataset in the rob
ous driving. Existing benchmarks rely on datasets, which are biased towards more common scenarios, such as cruising, and dis
s and ground truth objects. However, camera-only detectors have limited depth accuracy, which may cause otherwise reasonab
for the simulation of different human-agent navigation behaviors in scenarios with mobile robots. The tool, the first programmed
cs computing performance across a diverse range of hardware platforms using ROS 2 as its common baseline. The suite encom
ly in understanding and standardising fabric properties like elasticity, stiffness, and friction. While the significance of these prop
adapting online its behavior, in order to be robust to the changing environmental conditions. In the paradigm of passivity-based
of robotics in sports and fitness, making skill coaching smarter, easier and more accessible. Physical and social human robot in
weighted implements, to elicit external stimuli. However, this approach poses a significant challenge when attempting to modify
manipulation tasks with a robot. Given a user’s manipulation sequence, we propose a predictive model that uniquely casts th
n while having the ability to render virtual or remote environments. Therefore, ensuring the stability of such robots in physical hu
ving, such as dressing. A common problem in almost all dressing tasks is the insertion of a garment’s opening around a pa
es of the elderly community. Advancement includes greater immersive and enjoyment factors, which compute performance vali
ion, but effective methods to restore healthy gait patterns with exoskeletons are not yet clearly understood. Early research in ro
ceptions. However, in the field of physical human-robot collaboration (pHRC), its aptness should be put under scrutiny due to in
ng a physical human-robot interaction. Our Neural Network is composed of two main parts: a Multi-Layer Perceptron called Tra
interaction. Recent work has attempted to train robot policies by interacting with dynamic virtual humans in simulated environm
ation. An important example is determining where to move when the goal destination is not uniquely specified, and who should
es used in robotics, for instance, multibody dynamics laid the foundations for physics engines of human movement simulation,
tures such as Transformers and Visual Transformers in generating a prediction of the human’s force in a Human-Robot col
mobile robots has attracted much research attention. Many researchers have focused on improving the control performance to c
at the physical attachment interface. While some sources for power loss are inherent to the system, such as human soft tissue o
aramount importance. However, existing walking aids are either inadequate to provide sufficient support to users' stability or too
s is a challenging problem. It requires quickly exploring the environments and robustly merging the sub-maps built by the agent
afety guarantees using conformal prediction. Particularly, we design a model predictive controller (MPC) that uses i) trajectory p
on-free path planning of autonomous robots, offering global optimality through its convex formulation. The safe corridor is const
ontrol that simulates thousands of trajectories and uses these trajectories to synthesize optimal controls on-the-fly. In practice,
of dynamic and uncertain obstacles. Many recent results use learning and deep learning-based motion planners and prediction
ntinuous random variables. Yet, such a setting has hardly been investigated in the context of planning. Moreover, existing onlin
which is commonly modeled as partially observable Markov decision processes (POMDPs). However, traditional POMDP mode
 robotic applications for sequential decision-making under uncertainty. POMDP online planning algorithms such as Partially Obs
babilistic edges for efficient planning under uncertainty. We would like to generate roadmaps that enable planners to efficiently f
, to guarantee safety for both the robot and any nearby humans. Most work on planning under uncertainty does not scale to hig
maximize reward objectives while satisfying hard cost constraints, generalizing safe planning under state and transition uncerta
quantify uncertainty in their predictions, particularly on the unseen space including the occluded and outside scene content. Thi
s. These models are often inaccurate and based on offline information, so that the robot planner is unable to cope with mismatc
s and it constitutes a foundational capability of autonomous systems operating in most of the real-world. While 3D exploration p
on. Our goal is to compute collision-free paths that also satisfy the environment-specific traversability constraints. Our approach
ng under motion and sensing uncertainties. The original stochastic motion planning problem is divided into a deterministic motio
e robots. Although recent advances in machine learning offer promising methods to predict human trajectories in crowded areas
to failure in the motion planning pipeline, resulting in potential collisions. In order to achieve a high level of robust autonomy, th
s, where a robot plans paths through a previously unobserved environment while continuously receiving noisy local observation
anning for active perception. This task is particularly challenging when reasoning about high-level scene understanding using m
in uncertain, stochastic and time-varying environments. Environmental variability over time can significantly impact the system
cific sensing modalities, such as time-of-flight and structured light depth sensors, or RGB cameras. This specialization increase
ilter tree with sigma point-based mutual information reward approximation (ASPIRe), for mobile target search and tracking (SA
e estimation (errors of the order of 2mm cause task failure), the utilization of touch/contact modality can aid in accurately localiz
ng approaches for motion planning under uncertainties often assume parametric obstacle representations and Gaussian uncert
Gaussian motion and sensor noise. Our proposed algorithm, CC-K-CBS, leverages the scalability of kinodynamic conflict-based
afely travel without mobility hazards. While conventional planners have primarily assessed rover traversability index with predet
ently. Although plenty of previous works have proven successful in generating smooth and spatially collision-free trajectories, th
haviors in graph environments with support actions among teammates to reduce the costs of traversing certain risky edges in a
g tasks, such as coil storage, retrieval, and shuffling. Tasks arrive dynamically with precedence relations, while multiple cranes
ask is proposed in this paper. The task is of practical significance as a sub-module for higher level planning and scheduling task
esigns in which the underlying graphs are well-connected to simplify planning and reduce congestion. In this study, we formula
ch, or cooperative mapping, often require agents to intelligently coordinate their trajectories and form coalition over a large dom
uch as manufacturing and warehouse logistics. These systems face complex challenges in 1) task allocation due to factors like
tic and industrial applications. In such systems, cooperative path planning is of great importance, as local congestion and motio
carry out missions. The use of multiple robots allows tasks to be executed concurrently, expediting mission completion. It also
Early designs including prosthetic legs with a passive joint or without any joint as well as the Energy Storing and Releasing (ES
navigation to fit implants with high precision. Its tracking approach highly relies on inserting bone pins into the bones tracked by
e tool for measuring active muscular effort. While numerous studies have investigated real-time control of upper extremity exos
the lower limbs. This work focuses on IMU-based phase detection for stair ambulation. In order to enhance the detection sensit
gorithms due to their biomimetic capabilities and intuitive structure. Modern impedance controllers feature parameters that smo
rosthesis (AFP) with variable stiffness mechanisms (VSMs) is proposed to assist transtibial amputees to restore ankle plantarfle
man motor skills to aiding individuals in their daily activities. While exoskeletons are increasingly viewed, for example, as promis
exterous hand function in amputees still impossible. Understanding motor control between human motion intention and synaptic
by a single motor have the advantage of light weight. However, current single-motor BSEs have problems with accommodating
uman wrist, specifically the wrist bones and muscles responsible for wrist movements. To achieve a range of motion (ROM) and
ever, many existing ankle exoskeletons constrain the natural 3 degrees of freedom (DoF) of the ankle to limit the exoskeleton's
due to its simplicity in both implementation and user donning. However, this design results in a backward shift of the center of m
 abilitation and mobility assistance. Efficiently detecting transitions between different activities, such as walking, stair ascending
o potential falls, thereby impacting user safety and confidence. We introduce a control framework for omnidirectional recovery
xpenditure of any strategy to date. This strategy may be effective due to the customization of the applied torque profiles to the u
egain functionality that passive prostheses cannot provide. However, the best method for controlling these devices in coordinat
support during various activities. Although primarily intended for medical rehabilitation, there is growing interest in utilizing exos
ith many hypothetical distance candidates. The cost volume building process is often computationally heavy considering the lim
n and robotics are very often modeled as a pinhole plus possibly a distortion model. On the other hand, there is a large variety
ning higher-resolution lidar-like 2D point clouds from low-resolution radar frames on resource-constrained platforms commonly
uirements to enable scalable mapping and localisation. We achieve this by clustering all points of segmented scans into seman
menon in environments such as long corridors and tunnels, due to the lack of sufficient geometric features for frame-to-frame m
ed data (i.e., map) through nearest neighbors search is crucial in numerous robotic applications. However, static tree data struc
ping and various robotics applications. However, point cloud maps acquired using consumer-level light detection and ranging s
on tasks for LiDAR sensors in autonomous vehicles. Especially the clustering and extraction of objects from a point cloud usual
allenging task, mainly due to the potential shape ambiguity and occlusion. Most existing single image 3D reconstruction approac
propose an efficient reactive system that provides high-quality obstacle avoidance while running at hundreds of hertz with minim
on and navigation capabilities of mobile robots in challenging settings such as caves and underground mines. These environme
MPPI) control with spline interpolation of control input sequence and Stein variational gradient descent (SVGD). MPPI formulate
plex human-like scenarios. Vector-field-based real-time strategies, although widely used, usually do not account for unwieldy to
bed-time control. The navigation of the quadruped robot is fundamentally depicted as omnidirectional movements, while a feed
ages in alleviating surgeons' fatigue and improving patients' outcomes. These systems comprise a range of human-computer in
re novel manipulation skills. However, simultaneously learning geometric task constraints and control policies from visual inputs
nning more important than ever. However, multi robot applications in highly dynamic environments still pose hard challenges for
e Gaussian Process (SGP) based local perception model. The SGP local perception model is trained on local ranging observat
However, designing them to ensure safety and optimal performance presents significant challenges. Developing exoskeletons
ambulation by increasing dose, intensity, and specificity of walking practice. Robotic ankle exoskeletons hold the potential to fac
Traditionally, its primary goal was minimizing interaction forces during assistance. However, under this conventional definition,
interaction requires extreme care since an injured limb may be in physical contact with the robot, so compliant behavior is impe
ons with powered ankle joints. The proposed approach defines ankle trajectories using four polynomial functions, each represe
a solution to restore lost motor functions. Yet, there remain challenges in controlling FES to induce arbitrary arm movements. R
n provide numerous physiological benefits to neurological patients. However, the design of an effective hybrid controller poses
nts for Parkinson's disease (PD). However, there are shortages of physical therapists and adherence to self-guided PT is low. R
with sparse rewards from an offline dataset. While prior work has demonstrated various approaches for agents to learn near-op
y-critical real-world applications, addressing the trade-off between maximizing rewards and adhering to safety constraints. This
s but also strives to accurately characterize the full distribution of these returns, a key aspect in enhancing risk-aware decision-
out the other agents. Control Barrier Functions (CBFs) are showing promise for safety assurance but current methods make stro
lex challenge for achieving efficient learning from limited samples. While recent works have been effective in leveraging past ex
algorithms from violating constraints in the process of decision-making and exploration during trial and error. In this paper, a no
mplex AI systems leveraging image-based input data. Applications of these results range from super-human level video game
pulation tasks in homes and workplaces. However, generalizing from one task to another and mitigating negative task interferen
ous driving due to its application effectiveness and simplicity. However, depending solely on the limited semantic information fr
hich had value as a means of transportation in the past. Advances in technology are making this possible, and as a result, the t
 astic policies, such as Gaussian distributions, for local exploration and estimating policy gradient to optimize performance. Man
perceive, and engage within virtual environments, has attracted significant attention, owing to the remarkable advances in com
o scene quantization and action discretization. Existing platforms suffer from the inefficiency in sampling and the lack of diversity
raction with the real environment in robotics. However, the performance of offline RL highly depends on the quality of the datas
to design controllers for unmanned aerial vehicles (UAVs) robust to unknown disturbances during flights. These disturbances, t
robotics, especially in energy-constrained real-world scenarios. In this context, Hyperdimensional Computing offers an energy-e
n rope, is ubiquitous in daily life. These objects exhibit diverse physical properties, such as Young's modulus and bending stiffne
arry out symbolic event plans in unknown environments. Most existing methods assume that the event detector can accurately
audio, and haptic) is essential for tasks ranging from object categorization to complex manipulation. Drawing inspiration from co
blem, primarily due to the lack of open space for robot motion and limited perception. Several traditional methods exist to solve
mages (MPI), for high-quality view synthesis of complex scenes. MPI with Normalized Device Coordinate (NDC) parameterizat
us, there is an increasing interest in exploring machine learning techniques to speed up collision detection and sampling-based
y representing a scene. In contrast to classical methods that are based on explicit representations such as point clouds, neural
harsh real-world tasks. Self-supervised learning, which is used on raw and unlabeled data to pre-train deep neural networks, is
tic control and segmenting unknown category objects in general images. To leverage these for improved robot learning, we pro
ms. Hanging objects is a ubiquitous task that is encountered in numerous aspects of our everyday lives. However, both the obje
ed on their physical properties. As manipulation tasks are inherently invariant to the choice of reference frame, an ideal task rep
ng due to the need for robust, object representations that can be learned through robot interactions. This paper presents a nove
PVRs) for training downstream policies that execute real-world tasks. Our study involves five different PVRs, each trained for fiv
perception tasks such as localization, mapping, tracking, and others. Several of these approaches also combine other sensors
ions like lighting or camera position differ from those seen during training -- often exhibiting sharp declines in capability even fo
et capable of encoding relevant geometric details. When exploring unknown environments, it needs to be updatable incrementa
ponent of methods for many computer vision tasks. As visual representations for robots however, their utility has been limited, l

ed point cloud measurements. Such a problem is critical in various robotic applications such as robotic perceptive locomotion. W
e of a scene containing previously unseen objects. In many practical applications such as AR/VR, autonomous navigation, and
y exploit the CAD model implicitly, which leads to their dependence on observation quality and deficiency to occlusion, this pap
or automatic driving and robotic perception. While sparse hints from LiDAR points have improved cost aggregation in stereo ma
vailable. This prompts the question of how to utilize this data for geometric reasoning of scenes. While many RGB-D registration
g challenges posed by small unmanned aerial vehicles (UAVs). These UAVs have the potential to transport harmful payloads o
onments with many small objects. Limitations in 3D sensing capabilities often make it difficult to detect every possible object. W
quire. In robot manipulation, it can facilitate a robot to grasp and manipulate unseen objects. Mean shift clustering is a widely u
hat sequentially fuses LiDAR and camera data. It jointly estimates missing geometry and semantics in a scene from sequences
eralize to adverse weather conditions due to illumination variations, weather particles, etc. In this paper, we propose WeatherD
nected vehicles, to collaboratively control autonomous vehicles in accident-prone scenarios. Under limited communication band
. In the multi-model registration problem, we are given two point clouds picturing a set of objects at different poses (and possib
, perceiving DLOs in 3D from a single RGBD image is challenging. Previous DLO perception methods fail to extract a decent 3D
ms captured by a class of close-range time-of-flight (ToF) distance sensor. A transient histogram is a one dimensional tempora
tonomous driving systems. However, prevailing methods for 4D point cloud recognition have limitations due to sensor resolutio
hat is semantically rich, yet compact and efficient for task-driven perception and planning. Recent approaches have attempted t
cclusion from a sequence of RGB-D images. TrackDLO is vision-only and runs in real-time. It requires no external state informa
ect classes, with sufficient examples per class. However, real-world object classes are not equally represented in practical settin
uishes them from conventional robots. A joint design using material’s compliance enables origami robots to implement com
nt traction. One way to increase the energy capture of solar systems is through solar tracking, a means of reorienting solar pane
s attempted to expand robot training to encompass physical aspects of a robot's design. However, reinforcement learning meth
objects with various shapes and sizes in nature. Inspired by the helical grasping in the nature, we proposed a helical bistable so
vast majority of six-degree-of-freedom (DOF) serial robots have one of two simple architectures. In both architectures, the inve
body comprises a series of 3D-printed wave spring structures that create a flexible biologically inspired shape that is capable o
gation, a free-swimming deformable impeller-powered vectored underwater vehicle (VUV). A 3D-printed wave spring structure d
mmetric workspace is proposed. By arranging the first revolute joint of each leg such that they have the same rotation axis, this
t attention among robotics researchers due to its benefits, including high payload capability, shock resistance, and resiliency. H
on efficiently tackling downstream applications. In domains from NLP to Computer Vision, this has led to a consolidation of pre
radigm for addressing the emerging challenge of scaling up Robotics Transformers (RT) for on-robot deployment. SARA-RT re
ges in robotics. This paper introduces the DenseTact-Mini, an optical tactile sensor with a soft, rounded, smooth gel surface an
igms only perform large-scale pretraining for visual representations, whereas representations for other modalities are trained fr
ects via a diverse repertoire of manipulation skills and show howpassive human videos can serve as a rich source of data for le
struct motion plans such that the transformation between its end effectors remains fixed. This amounts to complicated nonlinear
-based approaches have been proposed to guarantee the safety of robotic systems. However, the effectiveness of these appro
nces and noisy sensor measurements. To account for such stochastic uncertainties, many robotic systems leverage probabilis
nt sets which ensure that a system inside of these sets does not have to leave these sets, and are thus essential for guaranteei
ns. However, the construction of control barrier functions for robotic systems remains a challenging task. Recently, collision det
botics community, from applications of automated laundry-folding and home organizing to textiles and flexible manufacturing. D
r safety properties. This is challenging because real-world systems may be subject to disturbances that affect their performance
rdized testing obstructs commercialization. The issues are not merely political, theoretical, or algorithmic but also physical, indic
recent years by leveraging visual inputs and machine learning for decision-making and control. Despite their impressive perfor
ernal disturbances. Existing approaches to robust control either seek to bound the worst-case disturbance (which results in cons
an be subject to the vagaries of the environment. In this paper, we propose a learning-based approach that can automatically c
d unknown environment. In such scenarios, the safety of the robots is of paramount importance alongside the completion of the
potential of complex robotics systems, and we can expect this trend to continue at an exponential rate in the near future. Howev
techniques that can provide safety guarantees are still extremely rare for nonlinear systems, such as robot manipulators. A we
gies for safe long-term operation. While many modern control paradigms offer theoretical guarantees, they often assume know
o work in proximity or direct contact. However, conventional collaborative robots powered by servo-geared actuators are intrinsi
g the ubiquitous uncertainties arising from unmodeled dynamics and noisy sensors. To take into account these uncertainties, p
sions despite imperfect machine learning predictions. Examples of such decisions are ubiquitous, from robot planning algorithm
as torque bounds. This paper presents a safety-critical control framework that leverages learning-based switching between mu
onventional methods aim to find a rotation and translation that align two point clouds in 6 degrees of freedom (DoF). However,
ual uncertainty when confronted with out-of-distribution (OOD) obstacles not seen in training data. We present a method to accu
or visual-inertial navigation, in particular on resource-constrained edge devices, because of its superior efficiency and numerica
ving and mobile robotics. Most existing methods perform tracking between two consecutive frames while ignoring the motion pa
llenges of Simultaneous Localization and Mapping (SLAM) problems in indoor scenes. This paper presents NGEL-SLAM, a low
arning long-horizon manipulation tasks, however, is a long-standing challenge, and demands decomposing the overarching tas
is motivated by common problems observed in field trials of robotic harvesting: Calibration/perception errors, workspace obstac
e fruit industry. Modern agricultural competition has shifted from competition between products to competition between supply c
g guidance for grain processing, storage and trade. GAI is routinely performed by trained inspectors who are required to visual
nd registration accuracy limit agricultural analytics to plant stand levels and large areas. Precision agriculture seeks analytic too
ng vineyard man- agement. Current methods for estimating pruning weight are costly, laborious, and/or require specialized know
cultivars to be used in following breeding generations. This in turn can enable farmers to improve yield to produce more food, fe
teraction with specialty crops, such as trees and vines. However, developing robotic solutions for crop manipulation remains a d
ed by multiple unmanned ground vehicles (UGVs) in agricultural environments. The algorithm considers varying robot priorities,
measuring their traits is an important task in agriculture often referred to as plant phenotyping. Traditionally, this task is done m
easurement in orchards. SWP is a metric frequently used by agronomists and growers to optimize irrigation schedules for crop
of harvesting robots are emerging in both commercial and research areas. One of the main challenges is the precision alignme
ansformative solutions in various domains, particularly in agriculture. For example, accurate identification and segmentation of
manding tasks of agricultural robotics for automated plant monitoring and growth analysis. Due to the complex geometry of plan
mount importance. To support breeding and field management, various characteristics of the plant phenotype need to be meas
s during physical interaction. Previously, very few studies have addressed the stability and transient performance of the scenar
(UAV) to autonomously scan warehouse racks for inventory management. Our main motivation is to make small-sized UAVs w
n two surface robots positioning itself only using the visual features on the surface robots, which enforces safety constraints for
epresenting complex 3D environments. These describes environments through a layered or hierarchical graph where nodes rep
s domain. Training and accessing new models requires substantial amounts of application-specific data, but the limited resourc
ment of autonomous harvesting robots. Although autonomous harvesting is expanding rapidly, the commercial application of sw
mers and aging. Therefore, the development of automation of vegetable production such as transplanting, harvesting and trans
y plants and a high crop yield. But apart from its importance, soil analysis is a tedious and time-consuming task. This paper pre
gh-quality fruit. Due to decreasing worker availability, pruning is a prime candidate for robotics. However, pruning also represen
d mobile sensors, but it can be time- and labor-intensive. In this paper we present an algorithmic framework to perform real-time
uitlets. State-of-the-art viewpoint planners in agriculture are designed to size large and more sparsely populated fruit. They rely
h as selective harvesting and de-leafing. To perform these tasks, robots must be able to accurately and efficiently perceive the
ch plants are grown in vertical layers within a mobile shipping container. Space limitations within shipping containers make the a
fective training of supervised learning approaches. Self-supervised learning techniques alleviate this problem, however, existing
agricultural practices. Most autonomous agricultural robots rely on machine vision systems. Such systems,however, often perfo
with respect to deployment of robotic, AI and autonomous technologies. In this paper, we tackle automation of a critical task th
(UGVs) is presented. The method aims to solve GNSS-denied localization problem using pole-like feature such as trees or stre
, biodiversity hubs, and sources of wood. By the very nature of their scale, monitoring and maintaining forests is a challenging
plications on account of high sampling costs such as time, energy, and potential destruction to the environment. Utilization of av
er the desired direction of a operator. This paper presents a novel decentralized control strategy for a group of robots to herd an
robot for heavy-duty application such as construction. The robot is equipped with a simple docking mechanism called a C-type
plication is labor-intensive and physically strenuous, which motivates the need for automation. We present PLASTR, a receding
obots are able to self-reproduce, such that the assembly system may scale its own parallelization. Robots and structures are m
-level change over time using multi-mission SLAM. Many applications require such a system, including construction, robotic na
gorithm for underwater assembly. Infrastructure is autonomously modified to suit the needs of manipulation tasks based on an u
materials require different loading techniques. It's, therefore, difficult to find a single controller capable of handling them all. On
ce hazard risks and labor-intensive work. To this end, recent studies have investigated using reinforcement learning (RL) to aut
ough visual inspection alone, and must be regularly inspected with other methods to maintain structural integrity. It has been de
 obtain three fundamental capabilities (reasoning, planning, and interaction) for solving long-horizon language-conditioned man
heir universal representation of real-world domains. In this paper, we leverage foundation models to tackle grasp detection, a pe
ften compute and execute actions that accomplish one task at a time. However, efficiency can be improved by anticipating upc
enchmark, which we call ``Language-World,'' which allows a large language model to operate in a simulated robotic environme
in robotics and AI. Recently, Large Language Models (LLMs) have achieved impressive results in creating robotic agents for p
robots with concept grounding capabilities to interpret natural language tasks. Leveraging these capabilities, this work develops
signing intelligent robots. Common approaches that leverage LLMs for planning are unable to recover when actions fail and res
nable subsequent manipulation tasks. To model the complex relationships between objects, tasks, and grasps, existing method
eliably deploying any data driven robots in both production and simulation environments. Compared to existing systems, DOS fe
d aerial vehicle performing a visual target-tracking operation in the presence of turbulent wind. Image information is extracted a
Unmanned Aerial Vehicles (UAVs) in the presence of unknown time-varying disturbances, and model parametric uncertainty.
proposed on the basis of the Active Disturbance Rejection Control (ADRC) paradigm. The controller is synthesized for the sys
s and meeting constraints. However, the substantial computational burden associated with solving the Optimal Control Problem
gorithm for fully-actuated 6-degree-of-freedom (DOF) underwater vehicles which estimates unknown plant and actuator model p
or handling limits in order to achieve minimum lap times. Autonomous race cars require highly accurate perception, state estima
d, controllers seldom consider estimation uncertainty and only rely on the most likely estimated state. Consequently, sudden ch
atellites system(TSS) under outer disturbances, and the effect of PI/PD compensation in restraining chattering on sliding surfac
ct to constraints. Yet, in the case of kinematic-controlled (i.e., high-gains position or velocity) robots, closed-loop QP control sch
allenges, primarily concerning the safety of interactions between robots and humans. To address these safety concerns, we pr
gnificant attention due to their ability to generate real-time motions and reach predefined targets. However, the conventional con
collaborative robots and the safety of human workers. In our previous paper, we introduced a time-optimal path tracking algorith
on visible-hand handovers, thereby heavily relying on hand detection. Large objects in human-robot interactions present a uniqu
navigation tasks despite the complexities and uncertainties inherent in human behavior. Underlying these interactions is a mutu
o interact with an uncertain human-driven vehicle. Predicting human motion is challenging due to inherent uncertainties in diver
er, the intricate relationship between robotic perception metrics and actuation metrics remains unclear, leading to ambiguity in th
evertheless, due to the unconstrained nature of passivity-based impedance control laws, the robot is vulnerable to infeasible an
s, thus demanding stringent safety standards and quick response times. However, adversarial attacks pose a significant threat t
and dangerous tasks has proliferated thanks to advances in sensing, navigation, manipulation, and on-board computing techno
are usually limited to inspection tasks, wheeled robots like LineRanger can ensure a broader range of applications. In this paper
hapes and propelling their bodies, they easily vibrate under sudden or periodic applications of external forces. Suppressing vib
rgency response situations, demand optimal performance and efficiency from every component involved to maximize the succe
representation of the environment for efficient large-scale exploration, while it integrates a novel uncertainty awareness scheme
lems with routing constraints. However, it involves solving two NP-hard problems: maximal coverage and traveling salesman pr
blem is reformulated as submodular maximization subject to intersection system constraints. The objective function is submodu
based on approaches utilized in wireless network sensors. The problem is motivated by search and inspection missions with m
onomous long-term exploration with fewer and cheaper sensors. Requiring just an RGB-D camera and low-power computing h
erception NaivPhys4RP, essentially based on cognitive emulation to understand the past, the present and the future of the stat
explainable way in human environments. In the literature, there has been an extensive focus on object labeling and exhaustive
ulate objects, and answer questions based on their environment. While existing approaches often rely on extensive labeled data
or autonomous driving in off-road environments. However, learning-based ap- proaches often yield unreliable results when conf
nder-explored imbalances on the radial axis: points are unevenly concentrated on the near side, and the distribution of foregrou
 nnotated training data that is both arduous and expensive to obtain posing a significant challenge for their widespread adoption
tation of low-light or nighttime scenes is challenging. This is primarily due to the limited visibility of objects and the reduced text
approach to improve 3D semantic segmentation performance. Recent transformer-based methods have also demonstrated sup
objects being swapped, added, removed, or moved. We introduce Lifelong LERF, a method that allows a mobile robot with min
ge about their surroundings in order to perform their actions competently and human-like. In this work, we investigate the use o
LAM systems, enabling numerous robotic and AR/VR applications. We present a new methodology for real-time semantic mapp
using a semantic topometric map. As it will be presented, the semantic topometric map is generated by segmenting the grid m
often rely on predefined concepts during training or are time-intensive when generating semantic maps. This paper presents O
sses of autonomous agents safely interacting in dynamic environments. With this intention, we propose Mask4Former for the ch
. Panoptic segmentation of 3D LiDAR scans allows us to semantically describe a vehicle’s environment by predicting sema
eviously unseen scene. To tackle this task, the robot must learn a nimble navigation policy by utilizing spatial patterns and sema
extensions of partial 3D scene graphs, that enable efficient high-level task planning with partial information. We propose a grap
opic models. The Gaussian-Dirichlet Random Field (GDRF), a state-of-the-art VTM technique, models these semantic topics in
navigation. State-of-the-art (SoTA) approaches generally extract features from RGB images and an additional source or modal
logy predominantly adopts a fully supervised framework. Such paradigms exhibit an intrinsic dependency on extensive labeled
ance in dynamic environments. LiDARs and cameras tremendously enhance scene interpretation but do not provide direct mot
ween multiple views to build 3D maps annotated with object classes in real-time, is a recent topic of interest. This paper highligh
ds have attempted to construct HD maps in real-time using vehicle onboard sensors. Due to the inherent limitations of onboard
cene understanding in adverse weather and lighting conditions. However, the previous studies mostly focus on designing a mul
here have been many advances on different fronts ranging from SLAM to semantics, building an actionable hierarchical semant
s such as 3D detection, map segmentation, etc. However, the approach struggles with inaccurate camera BEV estimation, and
usage to navigate complex environments (e.g., forests and large buildings). However, such environments often contain occlud
is robust to adverse weather conditions like fog, rainstorms, and blizzards but offers sparse point clouds. Current techniques en
es in 3D object detection. Camera-LiDAR Fusion has been a prevalent solution for robust and accurate driving perception. In co
s is crucial for real-world applications, such as self-driving. Traditional RGB-based detectors often fail under such varying lightin
eption system of autonomous vehicles. Several 2D and 3D fusion methods have been explored for the LIDAR semantic segme
ations is vital, especially for environments with low-light and adverse conditions. Multi-spectral fusion techniques aim to maintai
he sparsity of the LiDAR point clouds, a dense and accurate RGB/3D projective relationship is difficult to establish especially for
of autonomous driving. Existing methods have achieved good performance in this task by fusing different sensor modalities, su
online motion estimation. Quaternion-based cubic cumulative B-splines are exploited for parameterizing motion states continuo
egration of the Global Navigation Satellite System (GNSS), Inertial Navigation System (INS), and camera has garnered conside
ving object are mixed together. Typically, existing methods mainly align the background events to the same spatial coordinate s
detection in autonomous driving systems. Yet, the fusion encounters difficulties with extended distance detection due to the con
gle-view image and a sparse, noisy Radar point cloud. The direct fusion of heterogeneous Radar and image data, or their enco
behavior. Humans are adept at using semantic information for object localization and path planning, a skill that robots need to
D information in scenarios where the RGB modality is affected by noise or captured in dark environments. Existing methods typ
ns in intensities projected from camera images to the LiDAR point cloud. As the input, the proposed method uses a sequence o
h is important for applications, such as robot navigation and self-driving. Traditional frame-based methods suffer from performan
ro and camera measurements. The use of rate gyro measurements facilitates a more reliable estimate of homography in the pr
aneous Localization And Mapping (SLAM) algorithms against gyroscope saturations induced by aggressive motions. Field robo
ns in robotics. Currently, most 3D trackers focus on driving scenarios with neglected jitter and uncomplicated surroundings, whi
 on are needed. Classical user interfaces pose issues for the physically impaired and are partially not practical or convenient. G
t and smoothly transition to conventional contact-based impedance control. These techniques are suggested for application in f
control. In this work, we present a computationally low-cost framework for synthesizing barrier functions over point cloud data fo
sk is designed in proximity sensor space. We utilize a multi-sensor arrangement of proximity sensors that forms a proximity arr
and its consistency in the time domain directly determine feature detection and tracking performance, which further affect the ro
understanding tactile contacts. In this paper, we introduce MagicTac, a novel high-resolution grid-based tactile sensor. This se
g, designed to deliver robust performance in various real-world environments. Existing deep learning (DL)-based approaches fa
esence have become an essential part of our day-to-day activities. In human-computer interaction, speech is the preferred way
lds of robotics and artificial intelligence (AI). The task of reinterpreting calligraphy presents a formidable challenge, as it involve
eal with occlusions, which are naturally present in agricultural scenarios. In this paper, we address the problem of estimating 3D
n to counteract position uncertainty. Freedom and precision in driving allows scientists to reach and study sites of interest. Typic
eyond the reach of traditional platforms thanks to their ability to carefully select footholds and flexibly adapt their body posture w
nments, requires a system with high generalizability and adaptability to tasks and environments. In this paper, we first develope
quilts during sleep, are promptly and accurately re-covered. Our approach is formulated into two subsequent steps: interference
precision positioning. The intermediate representation of image servo policy is important to sensor input abstraction and policy
bots, but it requires accurate camera calibration including camera intrinsic and extrinsic parameters. However, it is labour-inten
vanced treatments, medication management, and robotic assistance while overcoming resource limitations. In the current healt
ivery of food and medicine as well as the movement and transfer of patients. However, the traditional continuum manipulator of
ents, efficient placement of square or circular markers can be challenging due to limited space. In this study, we developed a w
ct detection, pose estimation or scene understanding have been developed. While performing mobile manipulation actions, a ro
te area coverage is crucial for a floor cleaning robot, and its morphology design plays a vital role in realizing complete area cov
erformed by humans, the complexity of open-world environments poses a considerable challenge. Addressing this imperative,
ough frictional contact. The formulation incorporates a new corotational material model with positive semi-definite Hessian, whic
forcement learning has been adopted for its capacity to capture complex interactions, the training process often relies on simul
arning. Recent advancements in Deep Reinforcement Learning (DRL) have demonstrated great potential in this area. Nonethel
etic data generation and skill learning. A relevant aspect of simulation for a variety of robot applications is physics-based simula
he realm of robotic applications. Traditionally, many studies have favoured simplified travelling-wavy undulations derived from o
Along with the complexity associated with these environments, one of the main issues remains the evaluation of said algorithms
ustry is to use manual creation to generate scenes with specific characteristics and automatic generation to generate canonical
ontacts as the Linear Complementarity Problem (LCP). Compared to existing differentiable simulations, Adams offers features in
and testing scenarios for development and maintenance purposes. However, simulations are an imperfect representation of th
ciency and robustness of point cloud registration. Existing learning-based 3D descriptors and keypoint detectors are either inde
ned to support research into autonomous vehicle technology. However, these datasets are rarely suitable for a wide variety of n
sformation between two sonar image frames. Considering the unique challenges, such as low resolution and high noise, of son
oint clouds. They usually take up a huge amount of storage space for large-scale environments, otherwise much structural deta
n a future with increased autonomy. For many autonomous systems, estimation and odometry remains a single point of failure,
estimates, yet state-of-the-art SLAM systems predominantly support monocular or stereo setups. This paper presents a generic
ng (SLAM) system for ground vehicles. Our system features efficient initialization, effective sensor anomaly detection and hand
merous applications such as autonomous driving and virtual reality. Recent progress on neural implicit SLAM has shown encou
the field of SLAM(Simultaneous Localization and Mapping). Previous NeRF-based SLAM methods have difficulties with object-
d long-term SLAM backend. Given a pose sparsification policy, we aim to scale the SLAM problem with space explored instead
 s allow correcting the drifting pose estimates from a sensor odometry pipeline. In this paper, we address the problem of effectiv
d simultaneous localization and mapping (SLAM) because it provides constraints for data association, which further determines
paper, we address the problem of pose estimation and map creation using only radar sensors. We focus on two odometry estim
O) is a technique to refine an initial estimate of a set of poses (positions and orientations) from a set of pairwise relative measur
son to conventional hand-crafted methods and recent data-driven approaches, there is significant interest in plug-and-play algo
d increasing attention due to its ability to provide high-level semantic information while maintaining computational efficiency. Som
) has recently gained considerable attention. Within self-supervised SVO, feature representation inconsistency between seman
behaved data fitting models. However, due to the potential presence of zeros in their denominator, rational functions tend to yiel
ng sub-problems by establishing implicit data association in a way that the generative model of events is exploited. The main bo
single optimizable factor graph (1) a pose graph, as a set of robot keyframes comprising associated measurements and robot p
this area is Visual Place Recognition (VPR), which is the ability to recognize previously seen places in the world based solely o
s, relying on sensors such as cameras, LiDAR, and IMUs. However, in real-world scenarios, the performance of these sensors i
g in visual SLAM pipelines. The proposed method constructs a high-dimensional semantic descriptor for each detected ORB fea
have made significant progress. However, how to make full use of visual information as well as better integrate with inertial me
roach is characterized by being dense, therefore point clouds are not reduced to pre-selected features in advance. Furthermore
motion distortion and ranging error make it a bottleneck. Most existing methods are limited in accuracy and robustness becaus
mplicit scene representation. Existing implicit mapping methods for LiDAR show promising results in large-scale reconstruction,
own environment. In modern mobile robots, 3D LiDAR-inertial systems are often used for this task. By fusing LiDAR scans and
y systems by providing comprehensive environmental measurements. However, current multiple LiDAR-inertial odometry frame
static world, which does not align with real-world scenarios. To enable robots to safely navigate and plan in dynamic environme
fectively improve the accuracy and efficiency of LiDAR(-inertial) based simultaneous localization and mapping (SLAM). This ma
for robot vision that is often achieved by Bundle Adjustment (BA). This paper presents an efficient and consistent bundle adjus
, dynamic objects, such as cars and pedestrians, can introduce ghost artifacts during the map generation process,leading to re
eve the goal of simultaneous localization and meshing in real-time. This proposed framework termed ImMesh comprises four tig
on capabilities. Ensuring that sensors are mounted in a manner that enables accurate detection, localization, and mapping is es
ed reality, where feature-based methods are prevalent due to their accuracy and robustness. However, existing methods empl
cheme for 6-degrees-of-freedom (DOF) object pose compensation. Although the pose estimation of the 3D object by the camer
ber of edges for measurement (edge) sparsification in pose-graph SLAM. The greater the amount of measurement data, the larg
The convergence of a state estimator to the correct belief over the robot state is dependent on the proper tuning of noise models
ng (SLAM). However, these sensors are prone to failure under low visibility (e.g. smoke) or places with reflective surfaces (e.g.
bjects. In this study, we design a physical-enhanced radar-inertial odometry system that capitalizes on the Doppler velocities an
ent the limitations of the short-range and viewing angles of RGB-D cameras. Herein, the four-point LiDAR can detect distances
mation by converting 3D scans to 2D images or construct constrained bag-of-words (BoW) representations reliant on specific fe
rehend and interact with their surroundings. Existing fixed-class models limit the adaptability of indoor mobile and assistive auto
lem in computer vision and robotics, with broad applications spanning autonomous navigation, surveillance, and augmented re
ile robots in the real-world. Besides high-accuracy state estimation, a suitable and globally consistent representation of the 3D
rowing availability of SLAM-enabled devices provides the option of simply localizing in a map of another robot or device. In case
uction. Recent works have extended their applications to autonomous implicit reconstruction through the Next Best View (NBV)
navigation map, it is important to create both a stable pose graph and a high-quality occupancy map that covers all the navigab

akes them more robust than traditional frame-based cameras to highly dynamic motions and poor illumination. It also means tha
 vel filter design we dub the multi state constraint equivariant filter (MSCEqF, in analogy to the well known MSCKF). We define
ulated drift error in odometry, this paper proposes L-VIWO, a Visual-Inertial-Wheel Odometry based on lane lines. This method
mans, which requires robots to navigate efficiently and safely in crowd environments with diverse pedestrians. Pedestrians may
A human may recognize that it is not appropriate to disassemble the sports car and put it away as part of the “tidying.”
ng in shopping interactions. In this role, a robot needs sometimes to reject a customer's choice. Thus, we investigate different r
and their presumptions of robot behavior. Explanations can serve to alleviate detrimental impacts on humans and amplify their
n and challenging task for mobile robots. Previous works have shown the efficacy of using reinforcement learning frameworks to
s using textual descriptions. Our approach involves the integration of GPT-4 with dense captioning, enabling robots to perceive
ate into human life which necessitates the understanding of human-robot interactions (HRI). This study unravels the details of H
eaching language, science, and math, while applications focusing on the enhancement of cultural competency are quite scarce
fore, it is desirable for the robot to adapt its actions to fit users' preferences. Human feedback is essential to facilitating robot ad
imitations of existing reconfigurable electromagnetic devices by seamlessly integrating antennas with robotic structures. Based
s), have gained prominence in applications such as minimally invasive surgery and hazardous environment exploration. Howev
the environment, such as the predation of frogs, leopards, and eagles. Despite its ubiquitousness in nature, capturing fast-mov
cal state due to the accumulation of particles after removing the particle jamming,which means poor restorability, and the comp
ver unknown biodiversity, and provide critical insights into the Earth’s history and the impacts of climate change. Due to the
romised when tip-mount devices are added to enable carrying of payloads. We present a new method for securing a vine robot
assembly, using a soft wrist that can operate more safely and tolerate lower-frequency control signals than a rigid one. Previou
h as interaction with a varying environment. Due to their compliant continuous structure, they are inherently safe and adaptive t
nical deformation to adapt to different shapes without explicit control. These characteristics are particularly appealing for agricul
diverse environments due to their structure and weight. Thus far, soft wall-climbing robots with integrated functions that can loco
orm multiple goal-directed actions including locomotion in confined spaces, environmental mapping, object retrieval and object m
erate motion by deforming seamless muscles rather than having rotational joints or sliding cylinders, as well as having excellent
soft material, or following a continuous movement pattern, has attracted extensive attention due to its unique features, such as
ure of the connector caused by uncertain loads such as high temperature or vibration in the launching process of the spacecraf
adigm shift in surgical practice, significantly minimizing patient morbidity. However, the variability of inner diameter and the inter
obots, because their deformable bodies limit the magnitude of the forces they can exert on the environment. To achieve high m
a traditional diet. However, blackberries have extreme fragility (resulting in up to 85% of harvest batches sustaining damage) an
c robots. However, considering the issues of insulation and waterproofing, as well as the driving module of soft materials, their
ng simple control strategies, but it remains difficult to design mechanisms with a desired embodied intelligence. In this article, w
oto-actuators with practical use cases. One potential application is for the control of intelligent electromagnetic surfaces, or two
o envelop themselves around an object. Few hands focus on the palm even though human hands greatly benefit from their cen

nd comply to a variety of irregular shapes. To date, several soft adaptive grasping strategies have been reported, however, mos
uctures, robotics, and metamaterials. However, the design and fabrication of origami structures rely on human experiences and
ncreasing attention and is demonstrating great potential in small-scale applications, such as drug delivery, minimal invasive surg
deformations and offer speedy responses when exposed to voltage. Owing to their high electromechanical conversion efficienc
structural design. Here, lattice metamaterials are exploited to design pneumatic soft robots with programmable bending, twistin
to insufficient stiffness and single motion mode limitations, existing soft manipulators usually exhibit low load capacity and sma
around humans. However, manufacturing of soft robotic hands with several degrees of freedom requires a complex multi-step m
d and their thumb remains curled within the palm. Due to the underactuated nature of the hand, the dominance of flexor muscle
 -robot interaction, making them suitable for various applications. Herein, a novel soft flat tube twisting actuator (SFTTA) is prop
tion. However, real-world applications of these robots are limited by challenges in achieving reversible and repeatable operatio
htly constrained, unknown environments such as inside a jet engine or the human heart, current robots are less adept. In such
ators for positioning the module in 3D space. This is achieved by utilizing a ball joint as the rotation center and leveraging the s
challenge for soft and continuum style manipulators. The inability to carry sufficient load and effectively reject external disturba
, that can move in three dimensions and need to avoid any contact with their environment. On the contrary, contact should no
h which objects are handled. Vision-based sensing methods often suffer from fabrication complexity and low durability, while ap
cribed. The focus is given on the interest of using a deep learning model to explore the design space with a genetic algorithm. I
ng for patients with impaired movement function of hand, but also assist in completing daily grasping activities. However, most s
omatically sort produce in high operating speeds. Despite their benefits, these multi-component assemblies are prone to faults,
gid layer sliding mechanism. It aims to mitigate ankle sprains and address the progression of chronic ankle instability by provid
al metal tools for their ability to safely interact with delicate tissues. In this paper, we introduce a proof-of-concept soft, capacitiv
shape sensing through functionalized tubing used as soft optical waveguides. The sensor is fabricated by laser patterning an of
. The interaction aspect of prostheses has become a highlight research topic in the field of rehabilitation robotics. However, mo
ckaging constraints, relying only on intrinsic sensors. While various sensing ap- proaches have been studied towards curvature
However, this domain mainly focuses on kinetic teaching, and does not consider imitation of interaction forces which is essenti
ers from rate-dependent viscoelasticity, mechanical vibration and configuration dependency, making the generalized dynamic m
d adaptability, making them ideal for applications such as minimally invasive surgery and inspections in constrained environmen
daptability of robots, expanding the application areas and environments. We propose a tendon-driven soft bending actuator that
ies found in the natural world. However, many of these biological capabilities are yet to be realized. For example, vines in natur
input pressure, resulting from its morphological shape that locally deforms the soft material to different degrees by simultaneou
tly increasing attention has been dedicated to developing fully soft robots to exploit their full potential, with a recognition that ele
nd crucial for the development of safe soft robots. Sensing and understanding the environmental interactions with a minimal foo
rface (FMI) standard. The goal is to control any (industrial) robot with electrical drives using a customizable vendor-agnostic co
mlining the development and deployment process and operations of robotic applications becomes ever more important. Automa
robots using AI planning that fully supports partial observability and noisy sensing. The AOS provides a code-based language f
one by human expertise or trial and error. To evaluate automatic approaches to this combined optimization problem, we introdu
ity, with high-stakes search-and-rescue applications. Existing methods vary widely and each has its strengths and weaknesses
mplexity, temporally-extended interconnected tasks, and an almost infinite space of possible actions. Our key insight is that com
ptive assembly remains labor-intensive, demanding expertise in multiple domains and often resulting in task-specific, inflexible c
es for multi-robot multi-arm systems in a shared workspace in MoveIt2. The proposed method leverages a centralized schedule
oviding the underlying basis to develop a variety of high-level functionalities. However, when used in outdoor environments such
mote sensing data, such as optical satellite imagery acquired by lunar orbiters, is key for the identification of future landing and
c investigations and technological advancements. However, controlling these platforms in zero-gravity environments presents u
solution for the active removal of large space debris. Guaranteeing the successful capture of such space debris is dependent o
sed navigation about a non-cooperative spacecraft. Spaceborne Neural Networks (NN) are susceptible to domain gap as they a
ess the intricate challenge of the robotic pre-grasping phase under microgravity conditions. Leveraging reinforcement learning
ions such as rendezvous, docking, and proximity manoeuvres, leading to increased interest in employing Deep Learning-based
bile robots operating in unknown environments, enhancing their onboard perception systems for collision-free operation with fa
cost-effective, and accurate orbit determination (OD). This paper introduces a solution to this “lost-in-space” problem th
shing effects impact its dynamics and performance. To predict this behavior and control such systems, sloshing and flexible app
nology. However, several of its key components, such as standard interface design, compliant control methods, and path plann

within its shell rather than relying on a chassis with wheels to navigate. In this shell, it is completely shielded from dust and the e
nd heavy robotic systems. The trend is towards lighter modular systems that can be (re)configured in situ according to mission
es in a vacuum using their microwedge structures. This makes them an especially attractive solution for perching on and grasp
ue to their high flexibility and anti-overturning capabilities. Nevertheless, compared with legged and traditional wheeled robots,
d model-based controllers, dependent on the knowledge of SMS parameters, can be employed. However, these parameters ma
ge. We address the problem of estimating the probe pose with respect to the laparoscope without using markers and additiona
al fields. Various methods based on mechanical modelling, data-driven approaches or explicit feature tracking have been propo
h increased dexterity and precision. A major drawback, however, is the loss of sense of touch due to a lack of kinesthetic coupl
ty and autonomy in robot-assisted surgery. This paper presents a multimodal transformer architecture for real-time recognition
anipulation task for surgical applications using Reinforcement Learning without expert demonstrations or explicit curriculum. The
nging in autonomous surgical robotics. One challenge in these tasks is nonprehensile manipulation such as pushing tissue, pres
ency, but it is difficult to achieve due to the need for robust perception algorithms. In particular, 6D pose estimation of surgical in
cient, and minimally invasive procedures, sparking a new era of robotic surgical intervention. This enables doctors, in collaborat
ficiency and quality. Learning-based methods have been recognized as an interesting paradigm and been increasingly investig
mes a pivotal role in the early detection of lung cancer. However, the complex anatomy of the airway and the fixed length of the
sformation between the robot and the camera. Uncertainties in mechanical systems as well as camera calibration create errors
y rely on extracting discriminative features from the template and videos to recover corresponding matches. However, it is diffic
tic surgery, where soft tissues are being displaced, stretched, and dissected. Many DOM methods can be powered by simulatio
human surgeons. There are a variety of surgical subtasks that require geometric information of subsurface anatomy, such as th

ment, autonomy, and error detection in robot-assisted surgery. However, existing recognition models suffer from the limited ava
due to changing demographics, making automation an important part in future surgical treatment. In this work, we focus on a su
tion, and locomotion. Yet, a fast, accurate, and robust surgical simulation environment remains a challenge. In this paper, we p
ithm, which is robust to outliers and can align nonrigid shapes. Few works have managed to achieve both real-time and accura
eir miniature diameter, high dexterity, and compliance with soft tissue. CTRs comprise individual pre-curved tubes usually comp
rically nested, laser-cut tubes with offset stiffness centers. The distal tips of the tubes are attached to each other, and relative d
rate needle placement remains challenging due to limitations in current tracking methods related to interference, reliability, reso
us endovascular intervention procedures to treat peripheral arterial diseases, where magnetic actuation is used to steer the cat
al one. The ValveTech robotic platform comprises a flexible articulated manipulator and surgical interface supporting the effecti
cture operation. However, affected by respiratory motion, intraoperative accurate location of the tumor and its surrounding anat
ti-agent system, such that it achieves a desired formation, in a resource-constrained setting. Our controller is fully distributed an
ed for locating a moving target or source in a GNSS-denied indoor setting. The approach is completely on-board and can be de
ensive in terms of space and resources. The situation is even worse for studies involving multiple, possibly heterogeneous robo
botics. Morphogenetic engineering in robot swarms, however, proposes new requirements for platforms: precise motion control,
at any location at any orientation with simple control and without alignment. The robot improves upon previous FireAnt-series ro
qual and opposite forces on each other. By studying the effects of reciprocal and non-reciprocal interactions we are better able
performed by the swarm, an automatic method produces the required control software for the individual robots. Automatic desig
or drone fleets. One of the challenges in maintaining a large fleet of drones is ensuring that the fleet performs its task as expect
proposing a long-horizon object transport task called Long-HOT and a novel modular framework for temporally extended naviga
 g task planning with motion planning; the task planner proposes a sequence of actions towards a goal state and the motion pla
onformal geometric algebra. The CDTS was first defined using dual quaternion algebra and is a well established framework for
arch and continuous motion planning, collectively known as combined task and motion planning (TAMP). However, prevailing m
plication of autonomous trolley collection. For environment representation, a task-oriented environment partitioning algorithm is
meta-) decision process, where early decisions concern the skeleton (sequence of logic actions) and later decisions concern wh
the use of mobile robots for Gas Source Localization (GSL) tasks presents significant challenges. These aspects increase the d
mobile manipulators. Logic-geometric programming (LGP) has shown promising capabilities for optimally dealing with hybrid TA
eract with multiple movable objects and navigate narrow passages. Such settings are notoriously difficult for Task-and-Motion P
se of the Prolog language. Our framework hinges on a special organisation of knowledge base that enables: 1) its efficient popu
en high-level natural language instructions, such as ''Rinse a slice of lettuce and place on the white table next to the fork''. To su
s for planning and decision making. NARS enables robots to effectively handle uncertainty in real-time with complete sensor an
lose proximity manipulate objects while satisfying constraints and a given goal. In particular, we formulate the plan refinement p
tiple robots while minimizing the overall completion time subject to timing and precedence constraints. Our approach uses the T

searches for a sequence of symbolic actions, i.e. a task plan, and also computes collision-free motion paths. As the task planne
m to visit a set of locations that can provide the most useful data to reconstruct an unknown scalar field. In the budgeted version
w before performing a manipulation task. Named this as a coupled active perception and manipulation (CAPM) problem, we mo
ssembly tasks. RAMP consists of beams that a robot must assemble into specified goal configurations using pegs as fasteners.
s application of physical human-robot interaction (pHRI). One reason for this is that the severity of harm caused by such edged
ameworks such as Integrated Task and Motion Planning. However, task planning is computationally challenging and struggles
OON) were introduced as a knowledge graph representation for robots. A FOON contains symbolic concepts useful to a robotâ
epare a meal successfully. This paper introduces a novel task tree generation pipeline producing correct planning and efficient
ighborhood search to address large-scale problems, thereby reducing computation time. With an increasing number of agents,
complish this problem efficiently has become a focus of research. However, in this kind of task, the formulation of strategies for
ities represents a challenge for human operators, who have to learn how to operate the many degrees of freedom of the robot t
ration, communication impairments are unavoidable, hindering the human operator from delivering high-quality demonstrations
n of dexterous manipulation tasks with aerial robots or in ground service applications, facilitating the transferability of cognitive
d limited degrees of freedom (DoFs) at the operator side. Autonomous teleoperation is proposed to overcome this difficulty by p
tive and ergonomic control interfaces for teleoperation has become crucial. The rapid advancement of Mixed Reality (MR) has
sed model (OBM) to spatially map the environment impedance for the emerging use cases in Tactile Internet. Different from th
ling users to remotely control a robotic arm and create drawings in real-time from a remote location. The proposed control fram
n erupted from blast furnace. Current manual labor to remove lump iron near the outlet, which is essential to prevent lump iron
example for activities of daily living. Using shared control to provide task-specific assistance -- for instance with the Shared Con
g complex control devices or interfaces. To address the limitations and challenges of existing teleoperation frameworks, such a
y between a user and a robot using the minimum-jerk model coupled with a recursive least-square filter, and (ii) synchronizes th
has been mostly focused on the ``pick'' task, leaving the ``place'' task relatively unexplored. In this paper, we address the probl
pose of the end-effector by exploiting the null-space of the Jacobian matrix. Capitalizing on that feature, we developed a control
bot collaboration using a single smartwatch and smartphone. By integrating probabilistic differentiable filters, iRoCo optimizes a
s that may be difficult or dangerous to work in. Current systems are able to give operators direct control of many components o
manipulation to improve the efficiency of manufacturing and assembly processes. Robots are expected to assist in executing th
g manipulation tasks such as moving large boxes within warehouses. We define such tasks as Dynamic Mobile Manipulation (D
 ed levels of experience. This is primarily due to the non-linear properties of such robots in addition to the difficulty of controlling
and autonomous driving. The pre-training fine-tuning paradigm has shown great potential in the point cloud domain. Full fine-tun
practical solution to embed semantic understanding in autonomous systems without expensive point-wise annotations. While p
ohibitively expensive due to sample inefficiency. To address this challenge, transferring a pre-trained policy in the source domai
ge multi-modal model for 3D point cloud understanding. Existing strategies are quite expensive in computation and storage, and
driving. Recent works mainly focus on improving efficiency or accuracy but neglect the challenges when facing environment ch
havior cloning. We train a grasp evaluation model from a small number of demonstrations that outputs higher values for grasp
earning (RL) often necessitates simulation-based environments due to high physical expenses. However, when simulation takes
kle unstructured and complex scenarios that are difficult to model. However, due to constraints related to equipment lifespan an
ce aims to break down physical limitations and barriers, celebrating diversity and self-expression, while fostering social inclusio
anoid robot. To protect the Force/Torque(F/T) sensor and achieve the mechanical filter effect, we propose a robot foot that mim
bik's Snake unit. Each module has standardized mechanical structure and electrical interface, enabling rapid connection with an
heel which is designed for machining purposes on variable curvature freeform surfaces. The climbing robot demonstrates stable
e into the frame. The insertion task is often required in production assembly as a component of cabinets and windows. Our dem
tem is a departure from traditional single-task-focused cooking robots, as it is designed to multitask operations and complete e
or module for smart cities and robots, that fully reconstructs the 3D indoor environment with a single free-moving Wi-Fi device. I
onmental pollution and rising CO2 emissions severely impact ecosystems worldwide. Despite significant efforts, humanity has y
us applications; however, significant challenges remain for human-robot communication. This project focuses on integrating tec
y monitoring, and sports training. Our perception engine only uses one monocular RGB camera, produces accurate 3D human m
esulting in diverse architectures and applications. Robotics stands out as a field with substantial potential for leveraging deep le
dating back as early as 2007. However, with the advent of modern deep learning (DL) and deep reinforcement learning (DRL)
other agents "on the fly". State of the art frameworks for ad hoc teamwork are often data-driven, using a large dataset of prior o
navigation, urgent evacuation, human activity analysis, Autonomous Valet Park- ing (AVP), etc. Existing wireless IPS (e.g., Wi-
e hardware, vision, and control aspects of Easys. Most parts of the Easys body are made up of readily available products or ar
, this procedure is criticized for its operational inconvenience and stress on marine animals. Recently, the methods with Autono
omenon such as granular jamming have been developed for disaster robotics application. However, as their contact interface is
that promote the overall well-being of patients are on the rise. One such form of care is music intervention, which has been pro
l healthcare costs. Approximate 30% of CVD cases are attributed to structural heart diseases (SHDs), which encompass variou
but they still are expensive and lack autonomous capabilities. An open-source, inexpensive platform for research and developm
this is by using the Learning from Demonstrations (LfD) method to transfer human-like skills to robots. However, human motion
intrinsic stability, safety, and affordability. It leverages a buoyant helium-filled upper body paired with two slender carbon fiber le
harvesting missions in outer space. These missions will likely involve multiple robots from various entities operating in the sam
igh-level movement intentions from input electromyography (EMG) signals. However, during grasping tasks, commanding man
y zones that are projected to the operator as augmented reality. The context of the demo is a pick and place task performed by
le robot designed to climb steel structures. EEWOC can be deployed for close-contact inspection, maintenance, or surveillance
e challenges faced by the elderly population. The goal is to empower older individuals to lead independent lives in their own hom
tively collect artifact samples and perform various underwater manipulation tasks are constructive and significant in multiple fiel
x environments, enduring impacts, and adapting to confined spaces. However, controlling their coupled dynamics has posed ch
eration of these actuators is based on the process of plant root growth. Namely, this mechanism allows not to moves the entire
the spermaceti in its head, enabling it to ascend and descend in the deep-sea. Inspired by this biomimetic principle, we have d
mplex procedures remain beyond robots' reach. What restricts the application of robots in intricate surgeries? In this video, we in
 to work with humans. The device is designed to provide the wearer with better auxiliary ability. A Series-Parallel-Reconfigurable
umans and impedance control is one of the most popular control methods to specify the interaction between these devices and
the human with an extra robotic arm to enhance its operation ability. The interaction method of SRL and its enhanced performa
In this work, we propose to leverage the ensemble paradigm to enhance the robustness of latent planning systems. We rely on
usly. It is being used due to lower latency and higher temporal resolution compared to traditional frame-based camera. We prop
ct from 2D point correspondences in monocular images. Despite promising use-cases such as the grasping of deformable objec
ation to ensure robust and safe operations. For ambulating mobile robots like robot snakes, traditional methods for sensing requ
more effective to discriminate between objects by attending to objects’ compositional visual regions and contrasting with th
in a video sequence while keeping a unique identifier for each object. It is a challenging and fundamental problem for robotics.
multiple position-unknown Ultra-wideband (UWB) anchors to construct an indoor localization system with both accuracy and rob
stems. Unlike the current state-of-the-art method, which heavily relies on the accuracy of a pure visual SLAM system to estima
vironment. To achieve low latency, dense maps are typically computed on-board the robot, often on computationally constraine
ve emerged as a de facto framework. However, there is more to compositionality beyond the decomposition of the learning loa
d effective in varying conditions. Teach and Repeat (T&R) navigation has shown high performance in autonomous repeated tas
entation for object navigation task within complex scenes. We propose an object-centric image representation and correspondi
n this work, we propose a mobile assistive furniture suite that is designed to ease the life of people with special needs in indoor
rol (MPC) and model-based reinforcement learning (RL) for navigation planning of an autonomous model car across offroad, un
when a robot is deployed in an unseen environment, building an accurate metric map is time-consuming. Can an autonomous
ject identification and robot control in real-world environments. While typical visual markers use two-dimensional (2D) patterns
tical robot applications. Current optical flow approaches show remarkable accuracy, but top-performing methods have high com
n multi-modal cues from real-world urban environments and natural language instructions. While existing outdoor VLN models p
specific components improves navigation performance by enabling the robot-independent part to be trained with data collected
s based on first-person visual observations. The ability of agent to comprehend its surroundings plays a crucial role in achieving
a framework centered around the Probable Object Location (POLo) score. Utilizing a 3D object probability map, the POLo scor
hasizes visual grounding, commonsense inference and locomotion abilities, where the first two are inherent in foundation mode
ance degradation when there is a shift between the distribution of training and test datasets, often requiring re-training the netw
e to differences in terrain traversability, diverse obstacles, and the necessity for fast decision-making. Established approaches
dynamic changes, and the lack of a cost-efficient, accurate localization system. We introduce UIVNav, a novel end-to-end und
ated preferences over terrains on which the robot should navigate to ensure both robot safety and mission success. However, c
ost existing cascade and joint enhancement methods may provide undesirable results, suffering from severe artifacts, deteriora
ect keypoint (OKP) extraction approach, AnyOKP, which leverages the powerful representation ability of pretrained vision trans
nomous driving, which quantifies 3D scenes into grid cells with semantic labels. Recent works mainly utilize complete occupanc
arning techniques. While earlier methods directly learn a mapping from images to depth maps and camera poses, more recent
stablishing the dense correspondence between a pair of images that captures dynamic scenes undergoing substantial transfor
s object detection, segmentation, and identification. While there have been substantial advances in these individual tasks, integ
pervised cross-modal contrastive learning in the field of point cloud video understanding. Traditional supervised learning method
ns and autonomous driving. This study proposes a comprehensive self-supervised framework for accurate scale-aware depth p
ross various computer vision tasks, offering notable benefits in terms of dynamic range, temporal resolution, and inference spe
mpressive progress enabled by deep learning, single-view RGB pose estimation models perform poorly in cluttered dynamic en
3D reconstruction, robotics, and autonomous driving. In particular, thermal-based depth estimation has unique advantages, inc
ng robotic surgical activities stands paramount. Existing algorithms dedicated to surgical activity recognition predominantly cate
n real-world 3D labeled data. Our goal is to predict the 3D shape, size, and 6D pose of objects within a single RGB-D image, op
quality issues such as motion blur and extreme lighting conditions. Event cameras offer potential solutions to these challenges
ete image sequences. Recently, video generation models have been widely used in robotic research. However, generating con
conditions due to thermal radiation differences. However, they lack texture compared to visible images. Conversely, visible ima
a “tightness” projection (also referred to as cuboid) to facilitate 2D-based training, which fits the object contour well on th
wo-step learning of self-supervised transformer-based models. We first benefit from the generic pretrained models oriented tow
ding multi-view street images. Valued for its unified spatial representation aiding multi-sensor fusion, BEV is pivotal for various a
ontexts. However, the complexity of outdoor environments, combined with the restricted viewpoints in driving scenarios, complic
However, reconstructing neural radiance fields (NeRF) from realistic event data remains a challenge due to the sparsities and
s sensor technologies to achieve precision. Traditional passive sensors, such as RGB cameras, offer high-resolution imagery a
the field of view (FOV) of the camera proves beneficial for enhancing environmental perception. Unlike image outpainting techn
ks, including camera re-localization. While existing learning-based methods estimated from a series of RGB images have signi
benefiting from their exceptional capabilities in modeling complex data distributions. The standard diffusion-based policy iterati
on tasks from a small set of visual observations, thanks to the latest advances in computer vision. However, learning bimanual
d remote control of lunar rovers during exploration missions. Conventional sensing methods often struggle to capture the intrica
has experienced an increasing surge of interest due to recent advancements in image processing and deep learning-based tec
thin industrial environments, encompassing navigation, docking, and object grasping tasks. However, in contrast to controlled la
ative perception, which seeks to improve perception abilities through information exchange among agents. To achieve this spa
ulting images often contain waterdrops adhered on the glass surface, and these waterdrops significantly degrade the image qua

timation method, which requires neither its CAD model nor hours of object-specific training. Recent NeRF-based methods provi
s. Specifically, skeleton-based action recognition has been actively researched to enhance recognition performance at a lower
ed depth estimation methods have gained research interest due to their ability to leverage large-scale, unlabeled real-world dat
to the physical complexity of haze, the variability in haze conditions, the tendency to capture more details in noisy scenes, and
the surrounding environment only from front-view (FV) images, including static elements (e.g., roads) and dynamic elements (e
are of great significance for real-world AI-driven applications, such as autonomous driving and embodied AI.Unlike traditional c
from transportation and robotics to healthcare, has always captivated the research community. In the wake of rapid advancem
es due to the low signal-to-noise ratio in the images. Previous work has demonstrated the use of burst-imaging techniques to pa
romising potential in autonomous driving by mitigating the challenges posed by high-velocity motion blur. However, training the
eneration Network (PillarGen), which facilitates the transformation of point clouds from one domain into another. PillarGen can
such as high-precision assembly, reverse engineering and additive manufacturing. In order to meet the demand of the product
in spine X-ray images. With the emergence of deep learning techniques that employ landmark detection, tilt prediction, and spi
-nature vehicle make and model recognition (VMMR)—to the best of our knowledge the first entirely synthetically generated la
nd assembly. The state-of-the-art deep learning-based approach uses an indirect strategy, i.e., first finding local correspondenc
nipulations. In prevalent design paradigms, a position-fixed piezoelectric transducer (PZT) is utilized to generate ultrasound wa
wth. In this paper, we present an end-to-end method for 6-DoF antipodal grasps from cluttered scenes. Our approach includes t
and object pose estimation based on point clouds are key processes. Recently, learning-based segmentation and pose estimat
(AMRs) since they encompass intention and indicate future movements. We propose a new method that allows us to infer veh
ion. Endoscopic videos therefore often contain large rotational motions, which require keypoint detection and description algorit
d surgical education. In practice, the needs of trainees are constantly evolving, such as learning more surgical types and adapti
due to the thickness of just 1-2 micrometers and the intricacies associated with its varying density and adhesion, the difficulty o
ulators in 3D space using 2D stereo images without needing to perform stereo reconstruction. We use a stereo camera in an ey
cular camera, not using any internal sensors. In food factories, robots are required to handle a wide variety of foods without dam
bilities. In contrast, visual servoing methods allow efficient task transfer in high-precision scenarios while requiring only a few de
t controllers in this domain generate velocity control signals to guide the cameras to desired positions and orientations. Howeve
ned Aerial Vehicles (UAVs). The proposed scheme combines image moments descriptor and event-triggered Nonlinear Model P
ape of elastic deformable objects using robotic arms. Our approach works by forming a lattice around the object, binding the ob
ck” architecture - first using a light enhancer to brighten the nighttime video, then employing a daytime tracker to locate the o
ng perception with action. Yet, challenges persist, including generalization across scenarios, maintaining reliability, and ensurin
ed significant progress, with various paradigms providing attractive performances. However, the effectiveness of traditional MO
ey are no longer perceivable through the senses, is a crucial aspect of human cognitive development. In this work, we seek to in
rse a video stream in terms of moving objects with names. Tracking-by-detection has long been a dominant paradigm for objec
eat advances regarding Siamese-based 3D tracking have been made recently, it remains challenging to learn the correlation be
n methods are trained with synthetic data, and therefore struggle somewhat due to a train/test gap. Self-supervised methods ho
ighly accurate real-time track identification and estimation. For 3D multi-object tracking, recent approaches process a single me
due to strong interest from the research community and steadily improving object detection methods. The majority of tracking m
ropose a combined vision-based tracker-localization interplay integrated as part of a task-space whole-body optimization contro
ss, attributed to their strong modeling capabilities. However, as performance improves, running latency correspondingly increa
a correlation to reduce cumulative error. Our article presents a Stereo Visual-Inertial Odometry (VIO) system based on ORB-S
l actions for navigation planning. In this paper, we present ST-VIO, a novel approach which tightly fuses a single-track dynamic
based on two innovative techniques called adaptive interframe alignment (AIA) and dynamic inverse distance filter (DIDF). Alth
if available, its intermittency and/or inaccuracy calls for fusion with other sensors. In this paper, we develop a novel GNSS-Visu
ms from back-to-back dual fisheye cameras setup, providing 360 degree coverage of the environment. Firstly, we employ a slid
e vision algorithms directly on the image sensor. Unlike conventional cameras, FPSPs perform computation on the image plane
ving onboard pose feedback control for aggressive motion. In this paper, we propose robust and real-time event-based visual-in
cy of loop closure in visual-inertial SLAM systems, particularly when approximating Jacobians using the Finite Difference Metho
artially overlapped camera systems, named MAVIS. Our framework fully exploits the benefits of wide field-of-view from multi-ca
ed mobility inspired us to develop a vision-based wearable steering assistance. To ensure broad applicability, we focused on a
ntaining a compact size, as well as providing body weight support. This paper proposes a new robotic limb, Variable Grounding
ed time delay affects stability. In the design and control of wearable robots to augment human stance, the neural delay dynami
f energy recovery mechanisms remains unexplored. To push the boundaries further, this study introduces a portable pneumati

, helping patients to regain manipulative abilities. However, as the patient is not always capable of following the robot, safety iss
onditions such as herniated discs, spinal stenosis, or lumbar radiculopathy. Improper posture while lifting heavy weights is a co
esistor (FSR) array for accurate three-dimensional ground reaction force (GRF) estimation during human's various motions. Jo
ng individual, comfortable, and effective control strategies and assistance parameters of exoskeletons during locomotion. Resea
egenerative con- ditions. To aid individuals in regaining knee functionality, supportive exoskeletons were designed to be affixed
swing phase. This paper focuses on providing a continuous ankle joint assistance throughout the gait cycle using an actuated a
nkle joint movement during walking with the use of an actuated ankle foot orthosis (AAFO). A projection-based adaptation mec

otion on flat terrain and negotiability on rough terrain. This study presents a highly dynamic maneuver of the robot to leap onto
e to complex mechanisms and other factors, ground contact detection can be challenging to obtain in certain situations. This pa
 ded in models for real-time algorithms. We present a closed-form dynamic model for rolling with two planar smooth curves, and
et, their use is still severely restricted by the lack of efficient control algorithms for overcoming obstacles such as stairs. We take
central pattern generators (CPGs) for the robot to navigate over a range of terrains. A transformable leg-wheel design is consid
te the safety of the car-like four-wheeled mobile robots (4-WMRs) with Ackermann steering. To classify a planned trajectory to
nvironments. The bicycle and self-balancing vehicle are two main modes of the two-wheeled vehicle, and their combination allo
ems of car-like mobile robots. Currently, literature does not suggest globally asymptotically stable controllers solving this proble
he progress in locomotion strategy, safely recovering from unexpected falls or planned drops is still an open problem. It is furth
bilities of these robots through the integration of articulated robotic arms. However, achieving successful loco-manipulation, esp
truly challenging terrain, it is nearly inevitable that, at some point, the robot will fail and subsequently need to recover if it is to c
ehaviors for robots with various degree-of-freedom arrangements and link lengths. Pseudo-generalization is performed by anal
real-world scenarios. This paper proposes a new robust strategy for a humanoid to stand up in challenging scenarios where no
DF) that extends the principle of signed distance fields (SDFs) to articulated kinematic chains. Our method employs a combina
ooth task transitions while maintaining strict priorities. The weighted generalized inverse is adopted to derive a hierarchical cont
d manipulation tasks. Current reinforcement learning approaches to locomotion and manipulation rely implicitly on forceful inter
ving human-level capabilities on humanoids. Specifically, teleoperation with wholebody control is a promising hands-free strate
ods employ pretraining and fine-tuning strategies on natural scene datasets, which do not exploit the feature interaction across
different datasets can cause domain discrepancies, which hinder knowledge transfer and model generalization, resulting in deg
uch as shop floor operations, this assumption may not hold. Thus, it becomes necessary for a robot to recognize a specific targ
andidates and re-order them to increase the likelihood of the top candidate being correct. However, state-of-the-art re-ranking m
ew training paradigm for Action-SGFA is also devised to improve pre-trained VL models using datasets with SG annotation. Wh
xtracting and matching global descriptors in a database, but they often must balance the extraction of comprehensive contextua
e prone to generating low-quality response in challenging UAV scenarios, e.g., fast motion and background clutter. In this pape
ods rely on object-specific training and therefore do not generalize to unseen objects. Recent novel object pose estimation meth
k of transformers is their quadratic increase in computing resources with larger inputs and dependence on considerable data fo
r’s voluntary motion and supports the implementation of multiple assistive paradigms. In this paper, a novel variable transm
to reduce the electrical energy consumption of a robot. This letter presents an optimization of the arrangement of springs for a
incorporate mechanical springs into their joints. The forthcoming generation of soft robots necessitates adaptable joint stiffness
ironment or human. This paper presents a new compact serial variable stiffness actuator ((SVSA-III)) with linear stiffness profile
ophisticated, they do not fully provide control over energy transfer timing. To overcome this problem on the basic system level, t
gns of cycloidal gearing. The open source designs are 3D-printable combined with off-the-shelf components, achieving a high p
origami structure of designed Quadrangular-Expand pattern and the foldable pneumatic bellow. Integrated origami structure reg
ep mode with tooth-based gripping to ensure precision, 2) a continuous mode available locally for fine positioning. The actuator
nical energy. Motor-driven artificial limbs and industrial robots can help humans do these tasks, but motors require energy to pr
n-taut tether with controllable length. revF{To the best of our knowledge, this is the first method that addresses the trajectory pla
ght (LOS) communication with the ground station (GS) due to the obstruction of the mountains and are unable to retransmit info
tion-and-energy-aware motion planning for UAVs in GNSS-denied environments. The proposed planner solves the trajectory pl
le of reproducing zero-G environment with high degree of fidelity, these infrastructures can be complemented with multi-rotors e
rdware constraints (ie sub-SI{100}{milliwatt} processor). When deployed in unknown environments not represented in the traini
fixed-wing trajectory generation, as it considers a realistic six-degree-of-freedom (6-DOF) flight dynamics model, including aer
work, namely EVOLVER, to mimic the bio-behavior for robotics to achieve rapid transient reaction ability and high-precision stea

e through narrow spaces. However, their tiny size and payload restrict the possibilities for on-board computation and sensing, m
gital and robotic sketching, leveraging behavior cloning and reinforcement learning techniques. This paper intro- duces an appro
of the works have demanding hardware requirements and ignore safety measures when generating the swing stoke. To addres
ut fragile instruments. We introduce an approach to characterize geometric models of chordophones and their audio onset resp
borating colleges. The MBot platform was designed to meet the needs of teaching robotics at scale to match the growth of robo
models that can be incorporated into robust control paradigms for deploying dense formations. Thus, learning a model for these
l (offline) design phase, where the vehicle is exposed to different environmental conditions and disturbances (e.g., a drone expo
multi-object detection due to conflicts arising from the need for high-resolution imaging, efficient object searching, and accurate
utomated supplies in industrial and hospital delivery applications. Current solutions still suffer from limitations, such as the robo
new task that bridges these two trends, namely, multimodal planning problem specification. The aim is to generate a problem d
MHE's weightings, has shown promise in further pushing the accuracy and efficiency boundary. Currently, NeuroMHE is trained
e most efficient solution uses the memory bank extracted from the nominal samples to detect outliers. We approach our strateg
ver, due to the prohibitive costs of real-world data collection and precise object simulation, it still remains challenging for these w
LMs for complex, long-horizon sequential task planning challenges in robotics. LLMs are advantageous in offering the potential
 anaging real-world odor interference, ensuring resource efficiency for mobile robotics, and accurately estimating gas features fo
rediction accuracy, few research efforts have been devoted to enhancing prediction models’ interpretability and out-of-distr
nt clouds as input use voxelization to create a pseudo-image for real-time running. However, the voxelization process often res
ccount for the intricate dependencies between joints. We introduce CoordiGraph, a novel architecture that leverages subequivar
vironments. However, in real-world robotics settings, these systems often operate under partial observability because of senso
ts and rewards for specific tasks, resulting in the requirements of multiple policies and limited capabilities for tackling complex a
Recent advancements in learning from demonstrations have shown promising results for robot learning in complex environme
ft are not well-suited to robotics settings, where data often arrives in a streaming fashion and may be very high-dimensional. In
ance of plausible future unravelings, calculating an optimal course of action inflicts an enormous computational burden on the a
ormal quantiles as input, to detect distribution shifts with formal statistical guarantees, which we define as conformal policy lear
ue to their non-parametric nature. Yet, in high-dimensional state spaces, these filters face challenges such as 'particle deprivati
nsively studied to assess and enhance the dependability of deep neural networks. However, current research in this area has re
ransformer-based approach combined with a task-planning system and enhanced with several AI sub-modules to run on low-co
, especially when handling dangerous tools such as kitchen knives. This study focuses on enabling a robot to autonomously an
utilizes a coarse-to-fine scheme with a scene graph as the scene representation. Unlike previous methods that rely on either kn
gnificantly improved via prompting techniques, such as in-context learning or re-prompting with state feedback, placing new imp
cts are and where they are located is crucial. While the former has been extensively discussed in the literature that uses the lar
g simulated counterpart, showcasing a diverse range of challenges through example tasks within the Foosball environment. Ini
d visual reasoning. Inspired by the dual-process theory in cognitive science—which suggests two parallel systems of fast and
which heavily limits their scalability to new scenarios. Additionally, they use examples prompting style where the LLM is provide
aptable RL policies on physical hardware. By decomposing complex robotic tasks into component subtasks and defining mathe
rrent to goal locations based on the pre-built map, the local planner produces a kinodynamic trajectory to follow the reference p
f perceptual confusion in repetitive environments, where scenes in different locations are incorrectly identified as the same. Exi
ding of object stability, perceptibility, and affordance to prepare previously unseen and randomly oriented chairs for a teddy bea
ns or to program simulation environments and to code up reward functions for reinforcement learning. Such human involvemen
and demonstrations) to guarantee that the generated plan can be executed correctly. However, in such conditions, LLMs may
er, existing systems grapple with challenges such as unexpected open set environments and the complexity of black-box mode
Ms) have shown promise for translating natural language into robot action sequences for complex tasks. However, existing appr
ion, and placing it). ASC consists of three components – (1) a library of basic visuomotor skills (navigation, pick, place), (2) a
memory (WM) to long-term memory (LTM) leads to high data volumes and challenges in efficient information processing and acc
eraction process between humans and drones to enable seamless collaboration and efficient co-working. In this paper, we pres
, we present THEMIS, a novel compiler-directed approach for automated and proactive regulation enforcement. By expressing
ation areas such as aerial manipulation, payload delivery, and moving target tracking, it is critical to ensure dynamic consistenc
estimation by Extended Kalman Filter. In this system, a piecewise-polynomial parameterization of trajectories is adopted for the
able even for self-intersecting paths. However, the method requires significant gain tuning for practical use. In particular, for un
meras to obtain location data. However, harsh environments such as denial GNSS or strong winds make it difficult for UAVs to
sors e.g. a large amount of cameras could be prohibitive. In such scenarios, usage of fisheye camera with a wider field of view
bees, and other large groups of animals capable of moving in a decentralized yet coherent manner, our method does not rely
based works have shown success in exploring static environments, their inherent sampling randomness and limited utilization o
tect potential failures at each step. Our system needs no prior knowledge of the type of gauge or the range of the scale and is a
bility to data-driven controllers. Methods that operate only over trajectory rollouts tend to be data-hungry. In prior work, we have
 tical to implement on small, resource-constrained robotic platforms. We present TinyMPC, a high-speed MPC solver with a low
se a single cell with gap effect. The robotic manipulator can control the position of the microfluidic chip. The microfluidic chip wit
be satisfied to ensure parameter convergence in the conventional adaptive robot control. Composite learning robot control (CLR
n a self-driving vehicle operating in the physical world. It creates virtual traffic participants with risky behaviors and seamlessly i
e search space to fixed parameter spaces. To address these challenges, we propose a deep reinforcement learning approach t
apts to follow a target vehicle while maintaining a safety distance. To train our model, we do not rely on steering labels recorde
angerous to other agents or to the robot. However, reducing the risk of accidents requires prior knowledge of where potential co
on is the incorporation of realistic traffic models that align with human knowledge, an aspect that has proven challenging due to
al reasoning abilities of LLMs obtained during pretraining. However, while considerable effort has been made to teach the robo
hods often simplify the handling of perception data from multiple vehicles. In these approaches, the ego-vehicle aggregates obs
prediction framework called Partial Observations Prediction (POP) for congested urban road scenarios. The framework consis
studies have attempted to capture interactions between road entities by using the definite data in history timesteps, as future in
rch to higher-level trajectories. With this, real-world trajectory-based scenarios can be augmented with a diverse set of human
e mission and environment can enhance performance at individual and mission levels. However, this is difficult to achieve, part
s smaller than the number of tasks. The proposed DMCHBA assumes that agents employ an implicit coordination mechanism a
n, into multiple shorter paths corresponding to shorter prediction horizons. This implicitly enforces a Hankel structure in the inpu
bed target object in a previously unseen environment. Our approach makes use of Large Language Models (LLMs) for this task
ously challenged with the task of exploring novel areas, yet so far, humans decide for the robots where to go. Even then, priori
a novel method to achieve robotic adaptive tasks based on LLMs and Behavior Trees (BTs). It utilizes ChatGPT to reason the
er choice for capturing highorder interactions among agents in this field. However, most existing works only consider static hyp
tecture that merges vectorized numeric modalities with a pre-trained LLM to improve context understanding in driving situations
s' dynamic states, interactions with neighboring agents and surrounding traffic infrastructures, and their stochastic and multi-mo
erometer. Two different motion models are combined to estimate motion with three degrees of freedom. For this purpose, we pr
ction and Ranging (RADAR) enable environmental perception regardless of lighting conditions, they face significant challenges
nclear. The newly released nuPlan addresses this issue by offering a large-scale real-world dataset and a standardized closed
(HD map) to cruise across the city. Without the prior map, it's hard for vehicles to find right-turning and left-turning ways in larg
ents, inferring occluded ones, and modeling their interactions based on vectorized scene representations of the partially observ
mmercial applications. Traditional vision-based approaches focus on texture features that are susceptible to changes in lighting
tures. The generated predictions must accurately represent the probabilities of predicted trajectories, while also capturing differ
eir implicit interaction with the ego vehicle. This paper presents a novel planning framework, Multipolicy And Risk-aware Contin
ly formulated as a optimization process of a cost function, which can be challenging to manually tune for different traffic scenar
ntly limits wheeled robots’ mobility in real- world, extremely rugged, off-road environments, where part of the terrain (e.g., ir
ic navigation systems with decades of empirical validation to achieve safety and efficiency. However, the many complex factors
owever, few existing works can simultaneously tackle these considerations of tracking single and multiple targets in complex en
r autonomous driving scenarios have largely focused on a pre-defined (closed) set of object categories and driving scenarios, T
sk imitation learning framework for end-to-end autonomous driving that leverages a dual attention transformer (DualAT) to enh
line dataset to achieve broader generalization in the real-world tasks, which, however, faces the out-of-distribution (OOD) issue
ng scenario elements in a unified model: the position of new agents, their initial state, and their future motion trajectories. By pre
a resolution-free LiDAR. Recently, several research works have been reported regarding LiDAR super-resolution. However, mo
irregular obstacles. To predict CAD, we propose a neural network, namely CADNet, with an attention-based multi-frame point
not follow strict rules of motion and there are often multiple occluded entry points such as corners and doors that create oppor
eneous and deterministic with low variance, natural human driving behaviors are multimodal with intrinsic uncertainties, and hen
se safety-critical tasks, it still struggles to plan over extended periods. In this work, we present a skill-based framework that enh
t terrain traversability often exhibits anisotropy. However, the difficulty in obtaining multi-directional terrain labels hinders the em
capabilities, and energy savings. Many conventional approaches impose limitations on the types of objects due to the need for
cation provides a promising way. Existing methods are mostly based on the modularization framework which investigates the ga
ronment and localize the robot using only relative poses. The error-sharing mechanism under this representation effectively red
base for secure cloud storage and off-board compute, and a robust yet low-cost wheeled robot to act in real-world environments
otating the original scenes with perception systems' predictions. Such intermediate representations tie the quality of the motion
on of machine tools by industrial robots. In this paper, a data-driven method for accuracy enhancement of industrial robots unde
have explored visual pre-training with different self-supervised objectives. Still, the generalization capabilities of the learned po
due to manual efforts, operational costs, and safety challenges. A path toward such a universal agent requires an efficient fram
nomously constructing a 3D representation of the scene, where objects can be rearranged virtually and an image of the resultin
uick and easy learning of new tasks. We aim to enable this paradigm in robotic reinforcement learning, allowing a robot to learn
opose methods of improving motion forecasting systems subject to limited compute budgets by combining model ensemble and
d can lead to dangerous system misbehaviors; manually detecting such mistakes is challenging. Current software analysis tech
he dynamic model and the requirements for self-takeoff. Based on the requirements of flight strategy, we designed a system wi
onmental adaptability and low performance efficiency due to the structural design or inherent characteristics of its main constitu
mplexity, so they rely upon the tail for lateral maneuvering. The prototype E-Flap can deflect the vertical tail to make maneuvers
le mathematical model that could illustrate all these motions. Thus, for scientists and engineers, it is quite cumbersome and com
es affect locomotion performance on mud, we study the terrestrial locomotion of a mudskipper-inspired robot on synthetic mud
g. Although the benefits of hydrodynamic interactions in schooling fish have drawn growing interest in fields such as bio
obotic fish is challenging due to the constant swinging of the head during movement, which complicates precise control. In this
iency can compensate for the energy consumption brought by tuning stiffness itself. To explore this issue, we develop a robotic
e swarm robots with “fish-like” intelligence that will enable safe navigation and efficient cooperation, and successful com
he deep reinforcement learning (DRL) framework. Through both exteroceptive and proprioceptive sensing, the agent learns to c
environment render the system fully actuated. We show that the identification of form closed configurations is numerically feasib
by its geometry alone enables the hexapedal robot MediumANT to passively self-right without the need for additional sensory fe
or is designed for before it is needed), either through model-based trajectory optimization, or through deep learning-based meth
cluding Dima - a wheeled platform that used cheetah-inspired inertial tail swings to better execute rapid acceleration and turning
ures. This paper presents the development of a robot STRAY for quadrupedal locomotion, featuring a four-degree-of-freedom s
and a hybrid wheeled-legged motion, enabling adaptability across various tasks and environmental conditions. The robot draws
l and design complexities. Here, we propose a dynamic soft robot named LeapRun that possesses agile locomotion and the ab
es in granular media, we propose the use of a novel machine-learning control strategy to obtain optimal techniques for vertical
mall and lightweight on-board energy source remains a challenge. In this study, we successfully demonstrated the untethered ta
which has a body length of 10 cm and weight of 60 g. On the basis of the extracted skeletal muscle movement of a locust, we m
a, propelling them in the air. This paper presents an autonomous, springtail-inspired microrobot that can crawl and jump. The m
aterlily beetle, flap their wings to actively move around the two dimensional surface — a locomotion strategy referred to as inte
ingle 6-mg high-work-density (HWD) microactuator and undulates periodically due to wave propagation phenomena generated
ing skill remains a subject of ongoing investigation. Drawing inspiration from the learning flights of honey bees and wasps, we p
hrough ”force vectoring” mechanism recently. In this paper, to explore the agile flight ability of dragonflies on man-made
D) thrust offers advantages by generating propulsion without moving parts, but real-world use is limited by insufficient thrust gen
creasing attention in the fields of neuroscience and robotics. However, current research suffers from unstable sensory-motor ma
se their extension mechanism relies on elastic deformation or folding structures. To address this challenge, we propose a cont
formance of wheeled obstacle-crossing. To tackle this problem, this paper presents a novel multi-modal hybrid robot with the a
mbless robotic systems for effective movement across diverse terrestrial terrains. However, the integration of compliant mechan
h two Degrees of Freedom (DoFs), is normally subject to its limited Range of Motion (RoM). Although the reaction wheel circum
materials and generate composite products such as electric devices. Conversely, there have been attempts and endeavors to m
due to their cable routing method. To address these limitations, several scholars have proposed new mechanisms and control s
vement. This letter presents the first hardware design and implementation of a 2-DOF robotic eye driven by super-coiled polyme
tate the careful design of joint poses and the use of specialized calibration markers, while most recent learning-based approach
ch. However, geometric feature-based methods still have requirements on the scene. Deep learning methods, while achieving
ommonly been formulated as a specialized visual odometry problem, without taking into account the principles of stereo rectifica
CAMO and LTO, that monitor and track the LiDAR-Camera extrinsic calibration parameters online. Calibration monitoring provid
over a broader area. In multi-camera RGB-D setups, the goal is typically to reduce camera overlap, optimizing spatial coverage
ch employ planar boards with distinct geometry and image patterns, have been a popular choice. These methods simplify the ta
en LiDAR and GNSS/INS for autonomous driving. The first stage can quickly calibrate the extrinsic parameters between the sen
ce suboptimal and unrobust results. In this paper, we propose an online two-stage calibration method that estimates robust and
ement of the cameras. The standard approach is to use a printed pattern with known geometry to estimate the intrinsic and extr
ystems. Many existing calibration methods are constrained by predefined mis-calibration ranges in the training data, essentially
d Human-Robot Collaboration technologies. However, predicting and identifying the behavior of collaborative robots remains a
material property parameters of objects based on a set of observations. Our focus is on estimating these properties based on ob
tion is indispensable. Recently, some researchers have proposed deep learning-based methods that utilize convolutional netwo
ons, rely on regression methods that require collecting pose data from different robot configurations, which can produce low acc
ze, and cost. In this paper, we propose a learning-based approach for estimating the mass and center of mass (COM) of unkno
y of sensor calibration, which is challenging due to the difficulty of identifying common features across different data modalities.
otential applications in fields such as mobile robotics and autonomous driving. When combined with LiDAR, they can provide mo
ception in all weathers. However, the extrinsic calibration of multiple long baseline 4D radars is rarely researched in V2X, which
nvironments where robots must anticipate and adapt to local terrain, which requires visual perception. In this paper, we propose
ts lead to degraded sensing, or when the perception algorithm misinterprets the scene due to limited generalization. In this pap
ehaviors. We introduce a zero-shot navigation approach, Vision-Language Frontier Maps (VLFM), which is inspired by human re
on (i.e., searching for a goal in a novel setting). Typically, these roles are handled by separate models, for example by using su
as been widely applied in robotics to address the curse of dimensionality when learning from high-dimensional data, the inheren
Gaussian process-based algorithm for modeling a robot's configuration space and query collision checks. CollisionGP introduces
erred from real-world data via learning-based approximations. These approximations provide useful information on the pedestri
roach integrates multimodal motion predictions of dynamic obstacles with predictive control for obstacle avoidance. Multimodal
lyhedra. A novel collision avoidance constraint is proposed based on scale-based collision detection and the strong duality of co
e states accurately in simulations, it's exceedingly challenging in the real world. Therefore, a viable alternative is to directly map
ance approaches struggle in complex and various scenarios by avoiding collisions based on rules for specific collision scenario
blem in high dimensional spaces. The approach alternates between sampling points from selected regions in the search space
At the local level, we maintain a tree structure that is dynamically extended with the navigation. This structure unifies the plann
enerator for map images and corresponding trajectories, an image-conditioned diffusion planner for mobile robots, and a trainin
y of the feasible space.Our goal is to seek large convex domains in configuration space, which contain no obstacles.In these do
 primarily focus on training a policy that directly commands the robot with low-level controls, like linear and angular velocities, w
motion under uncertainty. However, the numerical methods required on each front have resulted in a mismatch of representation
ed by popular randomized methods including RRTs and PRMs, as well as those generated by many methods in trajectory optim
network (CVCDIC) to improve the compression quality by using decoder side information. The CVCDIC’s decoder utilizes
, lacking the capability to directly incorporate time constraints. In this paper, we propose a method that leverages a generative m
bstacle map and robot footprint are arbitrary. We propose a Neural Potential Field: a neural network model that returns a differe
lity of collision that naturally models the propagation of uncertainty over the planning horizon and is computationally cheap to ev
ived from reflected point cloud intensities corresponding to a certain height interval. The different layers of MIMs can be used to
a Sparse Gaussian Process (SGP) local map with a Rapidly-Exploring Random Tree* (RRT*) planner. Our approach begins w
mprove localization performance. Our research demonstrates the informative nature of collisions in this context.Moreover, we s
epresentations, such as point clouds. In contrast, our previous graph-based model, GraphDistNet, achieves high accuracy using
grid state check. Specifically, LOS check consumes more than half of the computational time in any-angle path planning algorit
s. Running RL-enabled systems among uncertain dynamic agents may result in high counts of collisions and failures to reach t
mera, 3D LiDAR, and the robot's odometry sensor to train a prediction model to estimate candidate trajectories' success probabi
hape rapidly-exporing random tree (E-RRT*) method is proposed for path planning of HRMs in workspace, particularly those wi
enges of implementing CBF in real-world situations arise from the absence of a model or the lack of an exact one for the enviro
well-positioned to reason about the physical world, particularly within domains such as robotic manipulation. However, current V
vious NeRF-based view planning methods exhibit computational inefficiency since they rely on an iterative paradigm, consisting
resources. Previous view planning methods exhibit inefficiency since they rely on an iterative paradigm based on explicit repres
re used for this task, in which so-called object queries act as candidate objects. Initializing these object queries based on curren
pproaches, relying on extensively annotated datasets to model articulated objects within limited categories. However, this appro
he MOS problem. While previous methods capture motion features from the range images directly, we argue that the residual m
as a unified representation for 3D reconstruction. Equipped with data-driven pre-trained geometric cues, these methods have
ping do not provide sufficient information for transparent objects. However, with the recent advance on datasets and algorithms
age the recent progress in the field of single-view human reconstruction to template-based reconstruct crumpled cloths from the
h integrates optical flow with amodal perception. Instead of only representing the visible regions, we define amodal optical flow

a scene flow estimation framework featuring a unified scene flow estimator, a symmetric cost volume approach, and a geometr
imensional scenes has become a significant challenge in the development of this task. In this paper, we present a novel transfo
them unsuitable for robotic applications with limited compute and energy budget. Moreover, correlation volumes scale poorly w
st from all collaborators, leading to a scalability challenge when dealing with a large number of robots and sensors. In this work,
mpletion of LiDAR point clouds. While existing approaches have demonstrated the feasibility of image-based LiDAR data genera
tinuously tasked with interpreting their surroundings, analyzing intricate sensor data, and making decisions within split seconds
s paper, we introduce a new LiDAR multi-task learning paradigm based on the transformer. The proposed LiDARFormer utilizes
ppers are still challenging due to large structural deformations. In this paper, a model-based method for intrinsic contact sensin
ands. An effective strategy is to obtain the contact properties for measuring rotation angle through visuotactile sensing. Howeve
mitations on vision systems such as transparency, reflections, and occlusions. This study achieves blind surface reconstruction
s are required to be known. This information is often extracted from CAD, limiting scalability and the ability to handle tasks with
es the density of sketch lines to draw light and shadow, resulting in the creation of a 3D-view effect. TacShade, building upon th
erent industries. We introduce a method using Physics-Encoded Graph Neural Networks (GNNs) for such predictions. Similar t
. We propose a stable pushing method that maintains stiff contact between the robot and the object to avoid consuming reposit
 cific examples of robotic contact juggling have been studied before, in this article, we provide the first general formulation and s
n, which is typically estimated in robotics applications using Coulomb's law of friction as a constant coefficient of friction from the
autonomous learning, but they struggle to reach long-term goals in large environments. On the other hand, Task and Motion Pla
dapt itself to novel environments and improve its performance during test time (i.e., lifelong adaptation). Several works for TTA
a challenging setting of lifelong learning with the goal of continuous learning in a dynamic non-stationary environment without fo
n of skills in robots, replay-based continual reinforcement learning has emerged as a promising research direction. These techn
reinforcement learning (RL) techniques (like DQN, PPO, SAC, etc.) struggle to adapt in large domains and suffer from catastro
control primitive skills. However, prior arts fix their skills library and steer the LLM with carefully hand-crafted prompt engineering
ers are only able to operate with a fixed set of skills. We overcome this critical limitation and present a method for using LLM-ba
ging of tactile data distribution as robots encounter new tasks, ANN-based robotic tactile learning suffers from catastrophic forg
re idea behind LOTUS is constructing an ever-growing skill library from a sequence of new tasks with a small number of corresp
roll dynamics, which affects their flight stability when exploring environments with unpredictable disturbances. In this paper, we
support, perceive, and manipulate the tabletop objects. Towards generalizable distributed manipulation, we leverage reinforce
n be easily customized to suit diverse needs. The morphology, i.e., the form and structure of a robot, significantly impacts the p
orkers. Furthermore, most glass facades feature window frames to securely fasten the glass panels to the building structure, en
task. This is because of the high number of internal degrees of freedom and the complexity of predicting its motion.In this pape
on and navigation tasks. Practical constraints of wireless channels, such as limited bandwidth, urge robots to carefully select inf
ment, resulting in a blind robot, and becomes more difficult when comes to a group of robots. However, it could be costly to eq
dynamic pattern formation is, therefore, investigated in this paper. A predictive flocking control algorithm is proposed based on a
gment epithelium (RPE)-Bruch’s membrane complex without causing trauma. Even for skilled surgeons, the inherent hand
st panoramic cameras, their usage in robotics remains constrained by the limited availability of datasets featuring annotations in
xisting works on handover failure detection and prevention focus on preventing failures due to object slip or external disturbance
ate the current localization methods with respect to a new difficulty: large variations in observation distance and, consequently, l
h are memory-inefficient for large outdoor spaces. Point clouds provide a lightweight alternative but existing benchmarks lack o
aches using domain randomization re- quire extensive scene and model optimization. To address these issues, we introduce an
lies. In these environments, numerous interactions occur not just between humans and objects but also between objects thems
algorithms improved in recent years. Moreover, current datasets fall short of comprehensive data-collection protocol for reprodu
. While mainstream works tackle this problem in the RGB domain, we focus on depth data synthesis and develop a Range-awa
ataset, denoted as THUD (Tsinghua University Dynamic) robotic dataset, for training and evaluating their dynamic scene under
ns the robot takes, creating a chicken-or-egg problem. Prior methods ignore such inter-dependency and instead train marginal
techniques is slow. We present nuPlan, the world’s first real-world autonomous driving dataset, and benchmark. The benc
bot interactions. For this, large-scale datasets that contain rich information are needed. We describe a portable data collection s
ompared to traditional narrow top-down step-by-step collection. We collect realistic data by performing any user requests within
augmenting the data with synthetic corner cases greatly enhances the safe operation of AVs in unique situations. However, the
els have shown promise in transferring trained policies to new tasks based on demonstrations. This feature is attractive for enab
eractions from past experiences, learning can enable effective social navigation behaviors directly from data. In this paper, our
ss of identifying and analyzing risks stemming from dynamic traffic participants and unexpected events. While significant advan
s robustly. However, supervised deep learning models require large data-sets to adjust their weights and to generalize to unse
is is particularly critical in developing countries where traffic situations are often dense and unstructured with heterogeneous ro
ethods heavily rely on annotated data and often face domain generalization issues. Unlike 2D images whose domains usually p
at enables precise object localization in 3D space with full six degrees of freedom. However, one critical challenge is the lack of
 sed techniques for autonomous systems in unstructured outdoor environments poses challenges due to limited data availability
s tasks in robotic grasping, bin-picking is considered one of the most challenging because of the cluttered bin filled with objects
es, and difficulties of obtaining ground truth metric trajectories for training and evaluation. This paper introduces the NYC-Indoo
ts over these techniques and accelerate agricultural research, beginning with semantic segmentation and diameter estimation o
ironments. However, they often generate virtual points as reflection artifacts associated with reflective surfaces like glass plane
er, differences in object structures and shapes across industrial datasets create domain shift issues, compounded by privacy co
oss-modality point cloud registration framework with Feature Filtering and LOcal-Global Optimization. The cross-modality featur
oud using Transformer. CAPT uses an end-to-end transformer-based architecture for joint parameter and state estimation of art
cing unwanted noise in the measurements. In this work, we tackle this problem by proposing a novel approach for detecting adv
pact descriptions on edge devices, we introduce EdgePoint, a lightweight neural network. We design a new detection loss Unfo
attention mechanism utilized in these methods considers many low-relevance points, and it has difficulty focusing its attention w
s research predominantly focused on semantic action classification or 2D target region prediction, we argue that predicting the
els to continually changing target domains. However, the pursuit of long-term adaptation often introduces catastrophic forgettin
ed slip detection algorithms commonly difficult to concentrate on key features when faced with dense 3D shape data obtained b
and human--object interactions. The two prediction results are combined to enforce commonsense knowledge, such as ``if the
ticular, when the latter is hard to engineer or the robot's kinematics change unexpectedly, human-free self-modeling is a neces
ed pose can facilitate the corresponding affordance task. Previous methods for affodance-pose joint learning are limited to a pre
xisting methods, which focus on large-scale indoor structures, our approach---Multi-Object RANSAC emphasizes cluttered env
nipulators. Such features can be used for vision-based control and pose estimation purposes, when placing artificial markers (e
ious synthetic renderers in conjunction with perception models is prevalent to create synthetic data to augment the learning in t
side of a well-controlled lab environment. Existing scene transfer end-to-end policy learning approaches often suffer from poor
Unfortunately, the heat signature of these flows is often too subtle to be seen by a thermal camera. While convection also leads
w an ever-growing interest due to novel use cases enabled by their reduced form factor. The price for their versatility comes wit
and for large-scale annotated data and the limited availability of such data in the context of 3D instance segmentation, we study
optic scene understanding. Our representation is trained using an image sequence with noisy robot odometry poses and autom
n produce inconsistent depth estimation in textureless areas and unsatisfactory depth discrepancies at object boundaries. To ad
nt angles makes prediction more complex than simple robot pose estimation, due to its higher dimensionality. Previous method
n for effective driving environment perception is hindered by the lack of comprehensive LiDAR datasets. The sparse nature of L
king and path planning more reasonable. Recently, point cloud prediction (PCP) is utilized to predict and describe future enviro
vigation skills as well as reasoning over long-horizons. Existing methods struggle to control the large configuration space and to
o-labeling process that creates SOGMs from noisy real navigation data. We use a 3D-2D feedforward architecture, trained to pr
venue. However, current methods often assume full observability of the scene and struggle with scale invariance. In many task
ined traction in recent years for addressing the multimodal nature of future trajectories. Goal-based prediction models simplify m
s are reliable, they come with the drawbacks of being expensive and vulnerable to impacts. To address these issues, studies ha
ire centralized training objectives, data, and compute. It is of interest if we can more flexibly create generalist policies by mergin
s limited a priori assumptions about human movements, resulting in a lack of explainability and explicit constraints enforced on
s. The objective of cooperative MARL is to maximize the cumulative team reward. Nevertheless, in partially observable settings
ason about the histories of their actions and observations. In this paper, we explore a new dimension in which large language m
per, we propose a Multi-Granular TRansformer (MGTR) framework, an encoder-decoder network that exploits context features
context of tactile fabric recognition and provide an implementation of information-theoretic exploration strategies based on mini
al signals, by enabling fast current flow between neurons. In contrast, bio-chemical synapses transmit neural signals indirectly,
 tegrating traditional methods. However, the improvements are only modest. In this paper, we propose a novel strategy by incor
ficulty of accessing the entire dataset in a single task are inevitable. Therefore, an incremental reconstruction approach is desir
of real-world images randomly collected from unlabeled everyday objects via the DIGIT sensor. Subsequently, we employ a si
shared human-robot representation space, which facilitates us to synthesize robot motions that interact with humans in social
one of the most intuitive representations of three-dimensional data. However, the substantial volume of point cloud data poses s
eeding on more difficult problems where others fail. At the core of the proposed algorithm is the use of a learned, generative Inv
bility and safety guarantees but are limited to specific system classes. In contrast, policy adaptation based on reinforcement lea
ed path planning, the model has to be evaluated frequently, which can lead to substantially increased computation times. We p
owever, when dynamic motions under constraints are needed and computation time is limited, fast kinodynamic planning on th
arch has primarily focused on single-agent solutions, real-world scenarios often require multiple robots to work together on rear
certain constraints. These constraints are crucial for ensuring the feasibility and safety of actions in real-world systems. We eva
nseen or out-of-domain (OOD), camera and bin settings, etc. In this paper, we present an uncertainty-based approach for onlin
scenarios may involve previously unseen objects, variations in camera perspectives, and bin configurations, among other factor
alibration rely on fiducial markers to establish keypoint correspondences. However, these approaches exhibit significant variabi
manipulation tasks for autonomous robot systems. GANDER relies on mapping input images of a trained task to a learned manif
o generalize to novel objects by recognizing affordances. This object generalization comes at the cost of being unable to recogn
o target segmentation, pose regression, or using fixed target blueprints, our work presents a holistic multi-level framework for p
es from the fact that human-designed objects are optimized for human use rather than robot manipulation. To address this gap
th geometrical extrusions, commonly known as a peg-in-hole task. The accuracy required to detect and orient the peg and the h
jects. However, the manipulation under heterogeneous setups that involve both 1D deformable and rigid objects remains an un
arios can be challenging. In this study, we present a novel approach to enhance the generalization performance of vision-based
metric neural models have been used to significantly improve the sample efficiency and performance of policy learning across a
nto a multi-object manipulation reasoning and planning framework. We propose DOOM and LOOM, which leverage transforme
t be able to flexibly perform appropriate actions based on the current observation. It is challenging due to high occlusion in the
us to learn behaviours directly from data. Most prior methods however rely heavily on simulation, which is still limited by the larg
deep reinforcement learning rely on external models for termination evaluation. To achieve more effective grasp for novel obje
y. In this paper, a novel double-critic deep reinforcement learning framework is introduced to optimize the push-grasping synerg
the specification of a goal shape. This goal has been obtained either by a laborious domain knowledge engineering process or
ften be decomposed into enumerable factors of variation, such as the lighting conditions or the placement of the camera. Empi
g problem in robotic systems and modern industry. Current methods requiring depth input fail on transparent objects due to dep
hout task-specific engineering, or else lack the data-efficiency to do so in an amount of time that enables practical use. In this w
Instead, it is necessary to first manipulate the object into a configuration that affords a grasp. We solve this problem by learnin
ions consisting of segmented images, to mitigate the sim-to-real gap. We address a long-horizon manipulation task requiring a
ntic categories, and kinetic constraints. Prior works mostly focused on recognizing and manipulating articulated objects with spe
ps on diverse and unknown objects have shown promising progress. However, existing approaches only consider a single grou
h as Hindsight Experience Replay (HER) expedite learning by taking advantage of failed trajectories and replacing the desired g
curate geometry by depth cameras. For the first time, we propose ASGrasp, a 6-DoF grasp detection network that uses an RG
small dataset. We propose a two-stage agent consisting of a tentative decision stage and a correction stage, where the tentativ
age and point-cloud observations are two commonly used modalities in visual-based robotic manipulation, but each of these mo
her pre-trained visual models to guide manipulation policies, which face challenges for novel instances in real-world scenarios.
relies on a model of the hardware being optimized, usually simulated based on the laws of physics. However, such a model is o
 earning from real data can be surprisingly efficient through re-using previously collected sub-optimal data. In this paper we dem
otated datasets of interactions or demonstrations. In this work, we argue that well-directed interactions with the environment can
c grasping systems. Most of the current literature constructs models around a single gripper, succumbing to a tradeoff between
ation excavated from global semantics, thus limiting high-quality grasp generation and real-time performance. In this work, we
y desired poses. This letter proposes a differentiable architecture for arbitrary pose servoing: a hyper-network based neural con
Our proposed approach uses visual observations from an eye-in-hand RGB-D camera, and we mitigate the sim-to-real gap wit
epresenting their temporal motion. These predicted point clouds help in other subsequent tasks like object trajectory estimation
predictions over perturbations applied to the inputs or model. However, such a learning paradigm suffers from two critical limita
ns, e.g., checking for bridge health. Crack segmentation under such conditions is challenging due to the poor contrast between
By compressing the weights and activations of a 32-bit full-precision network to 1-bit binary values, the proposed binary point cl
robotics. However, the radar point clouds are relatively sparse and contain massive ghost points, which greatly limits the develo
e the image or the language into a subsymbolic space, like the CLIP model, or create a symbolic space of extracted information
tly using self-supervision, and they can encode actuated and even non-rigid objects. However, learning robust, view-invariant k
and there is a demand for improved reconstruction of feature details. We propose an end-to-end 3D reconstruction system whi
e impractically difficult for some tasks. Sim-to-real transfer with domain randomization offers a potential workaround, but often r
ails to provide adequate cues for reasoning about objects' spatial configurations, limiting the ability to correct errors and adapt t
ese fixtures, accurate estimation of the contact state is essential for preventing and rectifying potential anomalies. However, this
elops joint policies for grasping and pushing, enabling effective manipulation of target objects within untrained, densely cluttered
s to the desired ones via a dedicated controller. This paper proposes a Data-Driven Model Predictive Controller (DDMPC) to co
introduce an innovative approach to learning these synergies by leveraging model-free deep reinforcement learning. The robot
ies across different robot hands or even grasp instances. One of the most direct models of in-hand manipulation is the inverse
environments. However, these approaches mainly focus on human vision or natural language, neglecting tactile feedback. In thi
plex-shaped objects on top of each other (see Fig.1). To design such a system, a robot should be able to reason about the stab
e possessing the skills of manipulating hand tools, for their non-trivial geometries and for the intricacies of controlling their conta
space, integrating Bayesian variational inference and neural networks for implementation. Compared to traditional methods, it a
n. The main limitation of this method lies in the robot's inability to adjust its behavior according to changing contact state during
with fixed grasping positions. However, the potential for dexterous manipulation of DLOs using an anthropomorphic hand is un
and finger design, the everyday finger, that comes close to many characteristics of the human fingers. In particular, we focus on
ethod prioritizing the end-effectors relative position in the objective function have been proposed. However, there is no guarante
existing controllers using geometric features, our controller employs physically based deformation features designed by decoup
) the explosion of contact modes, (ii) stiff, non-smooth contact dynamics and the resulting exploding / discontinuous gradients,
dexterous robotic manipulation (HiDex). This framework explores in-hand and extrinsic dexterity by leveraging contacts. It gene
with eight degrees of freedom to overcome these limitations. We employ DenseTact 2.0 for the gripper, enabling precise contro
s system can dynamically disassemble the rectangular structure into multiple individual UAV modules or integrate these UAV m
lures. We introduce FogROS2-Latency-Sensitive(LS), a Fog Robotics framework that offers secure, location-independent conn
mulation that tethers can also be used to do distributed formation control on very simple robots. Specifically, our simulated age
ted nature of the Alternating Direction Method of Multipliers (ADMM). The first one, Nested Distributed DDP (ND-DDP), is a thre
nts. In this work, the problem of territory defense by large-scale heterogeneous robotic teams is considered. The tasks include e
particularly in expansive, unstructured scenarios with dense obstacles. This paper presents an innovative, decentralized Vorono
O problem, we derive the Local Consensus ADMM (LC-ADMM) algorithm. LC-ADMM is fully localized; therefore, the computat
ion factors. The package loss, bit error, and time delay factors are considered in the distributed task allocation process. The pe
er autonomous maneuvering. This paper presents a novel autonomous underwater robot with fully vectored propulsion, combine
ibility of the morphing systems and high-level flight controller design, with limited discussions on low-level control. In this paper
es due to the high computational com- plexity involved. Sampling-based methods have emerged as a promising avenue for add
existing works concentrate on optimizing the spatial robustness of a system, our work takes a step further by also considering te
have proven to be effective, where deep learning (DL) approaches have shown limits due to data inefficiencies. However, thank
ossibly actuated or not. The simulation is based on the implicit time-integration of the Lagrangian model of these systems, whe
g dynamic programming. Our approach builds upon the pioneering (but largely unknown) O(n+m^2d+m^3) solver by Popov and
e delay of information exchange may cause error and instability. Thus, an appropriate way to determine the interface variables
amidst the complexities of real-world settings. They tend to rely on hand-engineered, rule-based robot models or static weightin
ven be negative when dealing with a deformable receptacle. This narrow clearance leads to complex contact dynamics that are
rol with relatively low compliance has been used, we propose a method that searching and aligning motions can be easily realiz
ges onto binocular RGB image pairs. Our proposed method encompasses three phases: detection, pose initialization, and pose
ent ASAP, a physics-based planning approach for automatically generating such a sequence for general-shaped assemblies. A
hicles (AGVs). A simulation-based approach for automatic roadmap design is proposed. An event-based simulator is developed
multi-session mapping in enclosed or partially enclosed ambiguous environments (e.g., long corridors, industrial warehouses). E
ic time and location. Relying upon such attacker models prevents algorithms from synthesizing strategies that can generalize in
unding the time between subsequent visits to a vertex -- a measure known in literature as idleness. In thedomain of multi-robot
nally, each agent in a sorting facility is tuned individually which leads to suboptimal performance if one agent is less efficient tha
A relaxed hover solution is derived for the bi-copter, by solving an optimization problem subject to rotor and servo stuck failures
o as impact-aware manipulation. Within this context, this paper focuses on developing an online collision monitoring framework
from all potentially occurring errors is included as unforeseeable events may happen over time. Therefore, anomaly detection (
the exact prediction of outcomes is not always possible. Even when the robot behaviors are well-designed, the unpredictable na
working cycle of the robots. In this work, we propose a fault-tolerant (FT) control pipeline based on model-free reinforcement le
able printing. This method can be applied by the printer itself for self-repair and enhance the reliability of commercial 3D printers
uch as the inclusion of aerial manipulators, only extends their versatility. As UAVs become more widespread, preventing loss-o
of the mismatch between the semantic space over which typical driving properties are asserted (e.g., vehicles, pedestrians, int
working on assembly (i.e., collaborative assembly) applications. This paper introduces a novel approach to generating continge
constrained spaces lead to back-and-forth maneuvers in large-scale environments, thereby diminishing exploration efficiency. T
ipped with rolling-capable perching mechanisms and a depth-vision system for the purpose of autonomous power line inspectio
tive models. Compactness enables perception in the large by ensuring that the perceptual models can be communicated throu
optimization method aimed at generating efficient, stable, and smooth traversal motions. To achieve this, we develop a planar
nce and push optimization toward divergence along weakly constrained directions. To overcome this issue, this work proposes i
tely Operated Vehicles (ROVs) operations, this can be compensated by the ability of the operator, but they come with high ope
, i.e. in the presence of complex current flows or wind. Traditional methods often fall short due to simplifications or limited onbo
dius is initially obtained based on three consecutive designated points during the turning process, which is regarded as the base
safety, cost, and efficiency, deploying lightweight autonomous robots, such as unmanned aerial vehicles (UAV), becomes more
is simple and inexpensive, such a setup fails to maximise their potential solar energy production. Single- and dual-axis sun tra
gy to inspect ball joints in parabolic trough collectors, which is a relevant problem that is not adequately covered in the literature
gy consumption of manipulators is necessary to make them affordable and spread their application in the future industry. Previo
or missions call for more autonomous solutions. Addressing the specific challenges of precise UAV landings—especially amids
is of utmost importance in situations such as gas leaks and industrial incidents, so that the efficient use of resources in respon
 poral phenomenon in an environment. We specifically focus on large lake monitoring, where remote sensors, such as satellites
ation method to calibrate the laser scanning system of an agricultural field robot consisting of two industrial-grade laser scanner
ic health and the climate. However, there is no currently available technology capable of measuring individual particles with suc
the environment that are not immediately visible but were visible at a different time. To solve this problem, we propose a metho
d a novel sensor called ViTacTip, which seamlessly integrates both tactile and visual perception capabilities into a single, integr
nciples or different sensitive materials have been proposed, the development of flexible large-area tactile sensors for robots is s
neuromorphic hardware in recent years has furthered our ability to create biologically plausible sensing solutions. While these
through its sense of touch. Among the diverse range of information accessible from tactile sensors, torques transmitted from t
strated impressive learning capabilities using visual input, a lack of tactile sensing limits their capability for fine and reliable con
mote environment with haptic feedback. However, in free motion, unlike unilateral teleoperators, bilateral teleoperators reflect un
ms uniformly or reaches deformation saturation. Here we present L3 F-TOUCH sensor that considerably enhances the three-a
challenging. Vision-based tactile sensors, like Gelsight, can mitigate this issue by providing high-resolution tactile sensing and
ifficult to derive illumination systems that work well within curved geometries. To address this issue, we introduce RainbowSigh
telligence to robotic systems would help intelligent agents understand task constraints and accurately interpret the dynamics of
stimation. Nevertheless, they are usually not easy to fabricate and require individual calibration in order to acquire sufficient acc
ated objects are crucial for successful execution. In this work, we present 9DTact, a straightforward yet versatile tactile sensor
der their ability to perceive tactile information effectively. In this paper, we propose a novel visual-tactile sensor named GelFing
manipulation contacts occur can be challenging, particularly in confined spaces, cluttered environments, or without installing m
n and robotic tactile sensing, providing fast detection of contact and onset of slip. Microphones are an attractive option for imple
en poorly investigated and mainly relies on data-driven methods. Discrimination of object compliance is crucial for enabling acc
are along with the complexity of designing effective controllers for tasks with relatively large state-action spaces. Here we introd
ctile sensors is challenging, primarily due to the large-scale unknown environments and limited sensing coverage of these sens
One of the critical problems is how to extract the terrain-related knowledge from sensor data collected as the robot traversed t
obotics rely exclusively on visual signals, which hinder a robot's ability to manipulate objects effectively. To address this limitatio
ack to solve this task. We present UnfoldIR ("unfolder"), a dual-arm robotic system relying on infrared (IR) tactile sensing and cl
g the bubble surface because local contacts change the global surface shape significantly due to membrane mechanics and air
ming is a pivotal response to labor shortages, addressing the need for increased efficiency, particularly in labor-intensive tasks
temperature sensing can help robots better interact with the environment. In this work, we propose a novel multi-modal vision-
ments, there are ongoing challenges in skin signal processing, particularly in achieving both accuracy and speed in dynamic tou
sing skin designed for physical human-robot interaction (pHRI) tasks such as robotic caregiving. Utilizing a combination of stretc
remains large. Models trained on simulated data often fail in zero-shot inference and require fine-tuning with real data. In additi
tion and multi-modal surface information of objects, thereby facilitating robots in achieving more dexterous manipulations. Howe
he lack of methods for simulating the sensors. Relevant research on simulating vision-based tactile sensors generally focus on
or, our framework estimates and tracks object pose while simultaneously generating motion plans in a receding horizon fashion
(FB) in human bronchial tree. The proposed evaluation criteria and optimization methods combine clinical significance related t
of two action primitives---burrowing and excavating---that can fluidize the scene to un-jam obstacles and enable continued prog
rosstalk and non-negligible equivalent resistance. Hence, this study presents an impedance-separating method, which does no
o an unfolded configuration that facilitates robust dressing. Applying quasi-static grasping actions with vision feedback on garm
g visual and tactile sensory data is key to understanding and navigating the complex dynamics of the physical world, enabling a
dangerous stimuli are detected. In this work, we proposed a method based on full-body tactile sensors that operates at three d
f view, robot motion space, etc.) hinders solving the hand-eye relationship accurately. In this paper, we propose a new generic
 nd help develop consistent and robust estimators. Recent studies focused on analyzing the observability of the state of various
dministering a swab test, tactile information sensed beyond the tool-in-hand is correspondingly crucial for contact state estimati
g cost of sensors has hindered the practical application of force measurement and force control. In this study, a flexure element
force sensing (IFS) methods are limited in application to medical instruments with arbitrary shape, due to high computational t
osed. Compared with other disturbance estimators, this estimator possesses notable advantages in terms of filtering foot-groun
namic information of the manipulator, a hybrid model, consisting of the nominal model and the residual dynamics, is establishe
daptability and flexibility, their potential is yet to be fully exploited. To support robotics deployment in manufacturing, we propos
ow a user to change the robots' task in response to potential future violations, changes to the environment, or user-defined task
del the evolution of the hybrid system as a two-player game, where the nondeterminism is an adversarial player whose objectiv
analysis using reachable set computation. We instantiate the procedure for two concrete classes, namely, the classical discret
f robustness is based on evaluating the robustness of underlying predicates. However, the robustness of predicates is usually
ently carry geometric constraints, such as the unit-norm condition of quaternions representing rigid-body orientations or the pos
sibility. The relaxation preferences are represented as a Weighted Finite-state Edit System and are used to compute a relaxed
al Logic (CaTL) mission, a fragment of Signal Temporal Logic. The agents' motion is modeled as a finite weighted graph, where
ud, that runs compute-intensive deep neural network perception or control models. However, this assistance comes at the cost
ors are crowded with people heading the same way. In this work, we address the problem of routing in such environments with
n of states for the considered task, limiting its applicability in deep reinforcement learning algorithms. To address this issue, by i
of its own actions, but also the effects of other agents' actions and the possible interactions between agents. Previous works h
ded temporal logic formula that specifies how robot should react. Little attention has been paid for autonomous generation of re
unrelated tasks, compatible tasks, and exclusive tasks. To specify such inter-task relationships, we extend conventional atomi
mporal logic (LTL) is a widely used task specification language with a compositional grammar that naturally induces commonaliti
n paint deposition due to the change in the painting angle; however, there was a problem with its versatility. We analyzed this p
ibility. To facilitate these efforts, we construct a framework in which learned behaviors, created by a natural language abstracto
world, to provide answers. We describe a method whereby a robot system efficiently answers such questions on the basis of r

d, which synergizes the advantages of vision and touch, and greatly improves the grasping efficiency of transparent objects. Fi
man hand is often the primary source of inspiration for many technologies addressing this challenge and a notable feature of ou
ransferring grasp experience or demonstration to a novel object that shares shape similarities with objects a robot has previous
stributions and friction coefficients). When a robot encounters a new object, these properties are often difficult to observe and m
derstanding and high-dimensional grasp planning. In this work, we propose an anthropomorphic hand grasping framework to le
rs. Unlike previous works that only transfer a set of grasp poses, FuncGrasp aims to transfer infinite configurations parameteriz
The widespread use of pick-and-place robots and Grasp Quality Networks raises the question of whether such systems are vu
xisting systems either operate on a closed set of objects or perform object-agnostic manipulation which lacks the capability for
ms inefficiently combine global and local information, often focusing solely on the local features of objects or the interactions of o
edded in a soft membrane and get modulated due to a deformation of the channel when pressed by an external force and rece
arse reward nature of this task made the learning process challenging to bootstrap. To avoid constraining the operational space
gies provide a simplified, low-dimensional representation of grasp postures and motions, that can be used for the description of
p. At each iteration, our method VFAS-Grasp builds a set of candidate grasps by generating random perturbations of a seed gra
se objects using only one actuator. Designed to be easily integrated with prevailing parallel jaw grippers, it alleviates the compl
standing: First, the robot needs to analyze the geometric properties of individual objects to find feasible grasps. These grasps n
esent the Grasp Reset Mechanism (GRM), a fully automated apparatus for conducting large-scale grasping trials. The GRM au
olve them renders state-of-the-art learning systems not combat-ready for high-stakes tasks. Recent advancements in interactiv
antly concentrate on the significance of the pre-training phase, the equally important task of extracting more effective representa
for environment instrumentation by building an implicit understanding of the world based on raw pixels, but navigating the cont
nian Mixture Models. The aim is to enable robots to grasp moving objects effectively. Our approach comprises three main comp
and cluttered environments. In this work, we propose a representation that can be effectively mapped to the multi-finger grasp
imal grasping strategy for dexterous hands, especially when it comes to delicate manipulation and accurate adjustment the des
other waste materials in part because traditional industrial end-effectors struggle to reliably grip these irregularly shaped objects
at humans exhibit extraordinary dexterity when handling objects. Such dexterity seems to derive from a robust understanding o
essed by simply counting the number of successful attempts in a given task, and this success rate is then compared to those of

grasping, finger-bending, and pull-in modes. Using these three modes, the developed gripper can rotate and translate a graspe
However, multiple sensors are often required on one finger, leading to significant sensor placement and wire routing complexity
self-adaptive gripper capable of conforming to a wide range of objects, this article presents an innovative topology-optimized de
a compliant pneumatic gripper and adapted trajectory plan is presented to initiate contact at a higher speed while remaining wit
atory transport is more difficult than its horizontal counterpart, and provide guidelines for achieving optimal vertical vibratory tran
by soft hands regarding carrying capacity and stability. The BSF design uses a hybrid variable stiffness mechanism combining m
gripper comprises four fingers, wherein each finger has a maximum size not exceeding 3mm, 4mm and 5.5mm in three dimen
lar consideration to potential challenges of articulated robots. The primary design goal is to achieve maximum repeatability duri
y, we are typically only interested in a subset of product parts, which opens the possibility of using destructive processes such a
es. The hand is synergistic in its operation (because its joint movements are highly coupled through the hand's interaction with t
ion of the different dishes are usually stored in small containers that are often not easily accessible. This paper introduces an in
y high-dimensional and intractable without a suitable underlying representation. This paper uses environmental constraints to s
maneuvering. In the present study, we first examine the parametric design of a squirrel-inspired underactuated gripper for pass
tics research challenge is the realization of flexible and reliable manipulation of soft, deformable, slippery, spiky and scaly objec
he hand impedes the rich, low-level feedback that would improve learning of dexterous manipulation skills. This paper introduce
grasped tool, are difficult for humans to program in advance. Recently, AI-based algorithms that can imitate human force skills
hesion for grasping and placing a wide range of objects in packed spaces. Applications include shopping and restocking in reta
oss a wide range of tasks, such as extrapolating to unseen object locations, incorporating via-point modulation, accurately mod
o-temporal information. In this work, we propose an efficient tactile recognition model, X-Tacformer. This model pays attention t
aily life. The highly under-actuated hand exoskeleton, which can track the 6-DoF motions of each fingertip via a highly under-ac
o compact vibrotactile actuators and two silicone chambers in contact with the user's skin. These components provide high-freq
owever, covering the entire contacting surface of the device with a dense arrangement of actuators can affect its normal use. D
ng their clinical performance as a consequence. This work presents the development and validation of a haptic- enhanced Virtu
urved paths. Yet the kinematic dissimilarity between continuum robots and conventional, off-the-shelf input devices, along with
nitude smaller than other ultrasonic haptic devices and is a good candidate for wearable tactile rendering systems. The device
raction---e.g., velocity, force, acceleration---as well as properties of the sensor and object under test. These dependencies mak
h makes them both mechanically complex and unable to form smooth surfaces. In this work, we build a multi- stable curved lin
s suffer from insufficient data collection problems due to the lack of effective guidance for users during the scanning process, w
ems effective is the implementation of an assist-as-needed (AAN) control strategy that intervenes only when the patient deviate
emory (FS-LSTM) architecture to handle the uncertainty and irregularity inherent in real-world sensor data. Recognising the ch
and inaccuracy of labeling procedures, unsupervised regression of sEMG signals has been employed in literature, however sh
 possibly minimising the physical effort needed to grasp the object. However, altruistic robot behaviours may result in protracted
rge-scale datasets is imperative for training models to acquire implicit knowledge related to the handling of various objects. Cre
quest such assistance is critical: too few requests can lead to the robot making mistakes, but too many requests can overload t
irefighting and infrastructure inspection. The limited prior work with UAVs has focused on user training using modalities like sim
set of human joints repeatedly, which leads to localized joint fatigue. In this work, we present a framework to plan a policy of a
netheless, to obtain realistic robot policies, an arbitrarily large number of queries is required to be answered by humans. In this
ance a robot's performance in unseen environments, data collection and model refinement may be impractical in every environ
However, gathering data that include crucial but rare scenarios - such as pedestrians jaywalking into heavy traffic - can be cost
automobiles. However, managing thousands of shared electric micromobility vehicles, including rebalancing and charging to m
participating in it. This paper presents a trajectory-based flight assistive system that enables various operators to fly the drone
ual's forearm weight, an external load can be detected and adaptively compensated via a feed-forward force reference, facilitati
anar compliant contact control method for planar manipulation to cope with unexpected contact. We first use sheet gel as a mul
ucts could break, possibly resulting in dangerous behavior of the robot. To protect the user from such behavior, this work addre
ght not want them to go everywhere or not in a specific way. For example, users might not want a drone to fly over a table whe
n interactions, the environment, crowd density, and social and cultural norms. Accurate prediction of pedestrian paths is vital for
ds to degrade in particularly challenging or unfamiliar situations due to the models' dependency on representative training data
eive uncertainty and risk in a biased way, which can lead to a diversity of actions and expectations when interacting with others
on for indoor robots. Real-Time Presence (RTP) asks the question, "Where in the home is the human now?", which is an import
en. When developing and assessing socially-aware navigation methods, pedestrian motion modeling plays a significant role. H
e, requires reasoning about the impact of the ego’s intended motion plan on nearby agents’ behavior. Deep-learning-ba
and picking the optimal trajectory in complex environments are critical issues. To this end, this paper proposes a Graphic- and
a solution to the failure-explanation problem, which are automatically-generated environment-based explanations. These explan
some restricted area. To this end, we propose a self-supervised 6-DoF grasp pose detection framework via an Augmented Rea
man behaviors and make trust-aware decisions. Consequently, there is an urgent need for a generalized approach to recognize
d entail the time-consuming processes of data and parameter tuning. To overcome these challenges, we propose an easily imp
hich can only be evaluated as a collective (a meta-behavior) and user preferences often differ between individuals, yet success
major antecedent of trust. By making robots aware of their capabilities, we can allow them to choose when to perform low-confid
the large language model (GPT-3.5) and the open-set Vision-language Model (Grounding DINO) to create an action-aware cos
ecoding signals captured non-invasively from the skin's surface. Portable high-density surface electromyography (HD-sEMG) m
on accuracy increases mental and physical demands on the user. In this work, we propose a system for predicting hand motion
y observing their behavior and, in turn, adjust their actions to facilitate better interpersonal communication and team collaborati
ers' demonstrated trajectories and evaluative feedback in a real-time two-player game. Our research shows that Haru is able to
visual impairments, as they are difficult to rely on visual cues to perceive essential information related to sports participation, su
sk of falling over, it makes it difficult for the rider to go in the desired direction due to interference by the front wheel assist. In or
task of peeling has received less attention. Peeling, conventionally a bimanual task, is challenging for care recipients using one
r knowledge about the interaction task, a large simulation dataset was generated and used to train a multi-output variation of th
guage navigation (VLN), a navigation task based on images and textual paths. Although VLN is expected to facilitate user instru
ion for vehicle guidance and operation. These methods do not consider high-level mission context resulting in cumbersome ma
and not repeatable. Although robotic infrastructure inspections address such limitations, they cannot replace the superior ability
ct for experimental control and orchestration. However, there remains an open challenge in deploying robotic solutions at these
collaborative tasks in robot work cells. We used Standard Usability Scale and the Usability metric for user experience question
 as a central factor in the development of passivity-based pHRI controllers, securing haptic transparency while guaranteeing pH
following given language instructions. These tasks appear in various practical robotics applications, such as household service
s. This paper investigates two control approaches for lower limb exoskeletons: a real-time kinematic state-dependent estimation
e live in, directly controlled by the room’s occupants engaging with it. Robot-rooms represent an advance in human-robot i
t's collaboration behavior to the user expectations. In this paper we investigate if trust can be estimated from observable behav
ans, robots need to respect the social space of the target person. In this study, we propose a dual closed-loop human-following
ncludes the design of performance metrics as well as the representation of user inputs for simulation-based optimization. The f
mplete complex tasks specified using temporal logic formulas that involve human trust. Since accurately observing human trust
ot collaboration uses observational models of humans (i.e. models that treat the robot purely as an observer) to choose the rob
th individual and collective beliefs of the team members. Dynamic epistemic logic has proved to be a viable tool for representin
h flexibility and efficiency of tight-clearance assembly of various complex-shaped parts in the presence of uncertainty without re
onent remains limited and serves little more than to satisfy the coupled localization problem. We present a collaborative 3D SL
ble time and effort, yet it might still not capture the full range of potential environmental variations. The generation of new synth
ct variation and dynamic environments. In this paper, we present CoBT, a collaborative programming by demonstration framew
en updated introducing the ISO/TS15066. However, complying with these regulations often leads to inefficient behavior, such a
sks. In particular, shared control between human and robot can reduce the surgeon's workload and improve the overall surgica
e of an underwater robot, essentially turning the robot's outer surface into the vibrating membrane of a speaker. We employ thi
ation and reproduce these skills in a compliant and safe manner. Therefore, two key components are addressed in this work: m
p reinforcement learning (RL) in the healthcare domain. Previous studies showed that assistive tasks can be formulated as mu
he tools. This paper aims to analyse the different methods of human robot shared control in the surgical domain. We propose a
ential but a non-trivial endeavor: not only do robots need to remember human feedback over time to retrieve the right informatio
ve to humans than translation. In fact, imagining the path to the target orientation is more challenging. In the literature related t
arned through Riemannian dynamic movement primitives (DMP). The proposed framework responds to human-robot interaction
ered shared control framework enabling a robot to learn from expert demonstration, predict operators' intent and modulate cont
uilding on the foundation of Fractal Impedance Control (FIC) introduced in our previous work, this paper presents a novel exten
ondition? Classical AI task planning has been explored for this purpose, but it involves a tedious definition of an inflexible planni
good understanding of how a robot will act in the environment. This can be achieved through explanations of the robot's policy
often translated in collision avoidance behaviour or high speed reduction. Adhering safety however is not the only aspect that m
icient HRC. The proposed control framework enables dual-mode HRC, filling the gap between contactless and physical HRCs.
have been updated in order to provide guidelines on how to asses safety in these new scenarios. In particular, Power and Forc
N methods totally rely on one model's own thinking to make predictions within one round. However, existing models, even the m
essful strategy to increase human likeness in a robot, enhancing mutual understanding and building a more natural and intuitiv
ng (RL) setups. One way that experienced recent research interest directly addresses the problem by considering human feedb
exists only in the input stages. To support interactive, human-robot collaborative painting, we introduce the Collaborative FRIDA
e field of human-centered robotics. To fill this gap, in this work, a wearable vision-based robotic system, MateRobot, is establish
onents: a mobile base, a robotic arm, and the human subject who gets guided by the robotic arm via physically coupling their ha
fixed set of innate MPs is insufficient to deploy them in dynamic and unpredictable environments. Instead, the full potential of M
gate the effectiveness of SuperLimbs in assisting astronauts to their feet following a fall. Based on preliminary observations fro
aling with a significant obstacle: the blanket or sheet that will almost always cover the person's body. We propose a method for
ems or interactions: information presented last among a grouping is most salient in memory formation (recency bias), followed b
on planning, tracking, human-robot interaction and safety monitoring. However, accurate prediction of human trajectories is cha
 of semantic reasoning architectures, these methods assume that all the training data is available a priori. However, each user's
y draw strokes on a canvas, to collaboratively create a portrait drawing through frequent interactions. The key is to enable the r
rfaces exist to operate these devices, brain computer interfaces (BCI) offer a control modality for those who have little to no mo
ng model, which can take full advantage of visible light images and infrared images under different conditions. Meanwhile, we d
els and draw them in oriental ink painting styles by using a robotic arm. The framework consists of a simulation stage and a rob
so visualize the map that the robots create of their environment and can easily send robots to new goal positions. The system m
bles on-site walking acquisition with high sample efficiency, real-time planning capability and generalizability across skill conditi
omotion on deformable, yielding substrates, such as granular terrains. We present an enhanced resistive force model for bipeda
ed via backstepping method to mimic reference passive gait, and a slope estimator for the inclination angle of the terrain. In ad
replicate natural whole-body locomotion patterns and execute seamless transitions by mimicking human motions. First, we pre
ers from generating adaptive behaviors online. In this paper, we investigate this problem and decompose it into two sequential
oxious heat. We use a spiking neuron network whose structure, encoding scheme, and form of information transmission mimic
everaging a 1D convolutional neural network (CNN), our method aims to maximize lead time for fall prediction while minimizing
ulders. This paper proposes a rigid-flexible-soft coupling three-layered robotic mannequin shoulder inspired by human body ana
e proposed approach considers multi-modal measurements (e.g. currents, accelerations, etc.) and non-directly measurable effe
alidating. However, the perception performance is degraded when these simulation-trained models are deployed to the real wor
e that previous methods have limitations in generating fusion BEV features free from cross-modal conflicts. These conflicts enc
ons beyond a single vehicle viewpoint. Currently, two major challenges persist in vehicle-infrastructure cooperative 3D (VIC3D)
ndicators, emergency lights, rear lights, and brake lights. We extend the Waymo open perception dataset with ground truth ann
le some previous learning-based methods integrated a physical scattering function for simultaneous stereo-matching and deha
ulti-class defect detection and segmentation is challenging due to the textural simplicity of road pavement image, the diversity o
World for Natural). It emphasizes the tight integration of physical and digital space in LiDAR systems, with networking being on
such as 3D object detection and bird's-eye view (BEV) semantic segmentation. Recent researchers have extensively explored
notations. Furthermore, the training of these networks typically assumes the prediction of a human bounding box and the accur
e to its potential to enhance the 3D perception of autonomous vehicles. However, existing methods rely on a complex cascade
riving, it is preferable to over-detect objects to avoid catastrophic outcomes due to missed detections. As a result, current state
action. However, the data source to train the various agents is independent and private in each company, leading to the Distrib
ating the limitations of single-agent sensing, such as occlusion. However, the precision of GPS signals can be influenced by a r
asting problems, in a multi-step pipeline. In this complex system, advances in conventional forecasting methods have been mad
terpretable (result cooperation) or flexible (feature cooperation). In this paper, we propose the concept of query cooperation to
adation is correlated with visibility. The fog attenuation coefficient is associated with visibility. In light of this background, this pa
ntly studied to reach a better balance by transferring feature maps. Existing works mainly perform spatial-wise fusion and adopt
sor failure. To address this issue, more recent research proposes using vehicle-to-vehicle (V2V) communication to share perce
o vision transformers, from small to gigantic networks, have been extensively tested on offline image classification tasks. In this
learning have shown that extracting skill embedding from raw human videos is a promising paradigm to enable robots to cope w
rn a value function from historical data, then guide the update of policy parameters by maximizing the value function at a single
ned with DINO. When interacting with a novel object, DINOBot first uses these features to retrieve the most visually similar obj
st to this paradigm, it is often significantly easier to provide raw, action-free visual data of tasks being performed. Moreover, this
y synthesizing preference queries to maximize information gain about the reward function parameters improves data efficiency.
on critical significance in the textile industry and household services. Accordingly, enabling robots to possess the above-mentio
alternatives for model optimization to solve the visual lane-follow application in urban scenarios with an imitation learning appro
 d, we propose the notion of soft fixtures extending earlier work on energy-bounded caging to include a broader set of energy fu
al model of the ball-on-cloth system is introduced. The shape of the cloth is simplified, and the rolling constraint of the ball on th
imitate the flexion/extension (f/e) and abduction/adduction (a/a) motions of human hand. The robotic hand has three identical u
Critic approaches are used for Sim2Real to reduce the rich idealized features in simulation to the accessible ones in the real wo
eciphering these complicated interactions over massive amounts of experience and complex reward functions to learn how to a
osed manipulation task leads to complex contact interactions which can be very susceptible to uncertainties in kinematic and p
s from the fact these skills are learned from scratch without the benefit of any domain expertise. In this work, we aim to improve
troduce a system that leverages visual and tactile sensory inputs to enable dexterous in-hand manipulation. Specifically, we pro
es on carefully synchronizing the fingers' dynamics. Indeed, fingers must simultaneously generate motion while maintaining co
r pyramid shape, the robot provides a good stability during the locomotion and task execution. Three suction cups, called the pe
ensive dataset that mimics the intricacies of cable grasping, factoring in potential collisions between cables and robotic grippers
pically available parameters is combined with an online path iterative learning controller (ILC). The ILC law is employed to com
rated system for the intuitive, interactive automation of surface treatment tasks with industrial robots. It combines a sophisticate
gmented. Typically, an Autonomous Multi-Functional Vehicle (AMFV) is not engaged in other scenarios when idle in a specific o
on. These factors can significantly hinder the performance of computer vision tasks such as object detection and image segme
c camera necessitates camera motion to capture a broader workspace environment. In this work, we propose an autonomous
. We exploit machine-learning techniques such as reverse-mode automatic differentiation and constraint parametrization to ach
based on fixed manipulators for free-motion tasks (e.g. welding, pick and place, or painting). In this work, we propose a control
n of the underlying uncertainty distributions. For this reason, they are often unable to capture more complex distributions (e.g.,
t detection assumes a discrete set of objects and loses information about uncertainty, so any errors compound when predicting
anners when the update interval becomes infinitesimal. By requiring the trajectory parameters to evolve continuously in time, th
d on an accurate model of vehicle dynamics and engine, including weight, drag coefficient, and the Brake-specific Fuel Consum
s, even when subjected to disturbances, which are modeled as random process noise and measurement noise. The approach
jectory from the last timestep, often falls short of providing an adequately close initial guess for the current optimal trajectory. T
presents a novel reinforcement learning approach to solving high-dimensional robot navigation tasks with complex logic specif
ect inputs that satisfy output constraints in dynamic environments. The main contribution of this paper is to propose an output-sa
control spaces, thereby enabling a wider suite of dynamic tasks in real-world applications. Our contributions are three-fold: i) a
orcement Learning scheme to yield a novel solution for unknown workspaces that inherits provable safety, convergence and op
Empirical real-world trajectory data is valuable for defining data-driven model preconditions regardless of the model form (analyt
ad to limited information available to inform planning decisions. Furthermore, resolution and latency limitations of real-world perc
rrent, faster transition times call for robust, modular and interpretable solutions that also allow a robotic system to learn how to
on planners have to experience a variety of planning problems throughout their lifetimes, which are yet to be leveraged for futur
ation of vector-based backbones for prediction and planning tasks may lead to the occurrence of causal confusion. Previous st
n high dimensional planning, learning based methods have shown great potential in achieving time efficient and optimal motion
ate, making it necessary to use Model-Predictive Control (MPC) methods to plan motion. However, these methods need a sho
ecially when a steering function is unavailable. This paper presents a novel approach that adaptively biases the sampling distri
ed calibration methods let the user measure points on the fixture's surface with a tool tip attached to the robot's end effector. M
uch as Remaining Useful Lifetime (RUL) and the State of Health (SOH), based on a, typically, restricted set of measurements t
ormance gap is difficult or expensive to measure through in-person competitions. To bridge this gap, we propose a compact, In
he effects of the boundary input force and gravity on the payload, which results in non-homogenous boundary conditions. This
captures an image of the fabric pattern placed between two acrylic plates with a small opening, utilizing a deep learning model t
 s, different areas of the surface cannot be optimally reached by a simple strategy using a tool with a relatively large and flat finis
ed control due to a lack of a comprehensive manufacturing system model. In response to this challenge, this paper presents a m
g factors regarding the deployment of such algorithms particularly on mobile robots. Although offloading a majority of the algorit
sh the lane changing tasks, profiting from their communication ability. However, dispatching CAVs in mixed traffic remains diffic
es usually suffer from the non-independent-and-identically-distributed (non-IID) problem, which may cause negative effects on t
led with semantic Operational Design Domain (ODD) tags (e.g., WeatherRainy, RoadTypeHighway and ActorTypeTruck) to be
on is: In safety-critical settings, can we avoid the need to have one human supervise one machine at all times? The work forma
verages a hierarchical forward-looking, risk-aware estimation framework with learned parameters to generate human-like drivin
antic map information and modeling interactions, it is important to build an effective mechanism capable of reasoning about be
efficient coordination strategy for CAVs. In this paper, we formulate the cooperative autonomous driving task of CAVs as an op
road participants with different social preferences (e.g., selfish or courteous human drivers). To ensure that autonomous vehicle
ognitive state when delivering warnings, which can result in false alarms and impact the driver's trust in the system. To address
quate safety indicator, recent works leverage domain knowledge to identify safety-relevant VRU and back-annotate the criticali
se to adopt Random Sampling strategy to process large-scale point clouds. However, our quantative and qualitative studies hav
mate their states such as velocity and acceleration in the present. Existing works frequently focus on the association task while e
(CAVs). However, time-varying channel variations in vehicular transmission environments demand dynamic allocation of comm
routes for autonomous buses. We first train a graph neural net model as a policy for constructing route networks, and then use
ic cable. To construct the controller, we employ the backstepping technique and propose an Uncertainty and Disturbance Estim
el 3+ systems to improve the safety and comfort of passengers, traffic authorities need to establish new procedures to manage
cle avoidance, and path planning. While most current methods utilize LiDARs or cameras for Multiple Object Tracking (MOT), th
bot vehicles given the increased emergence of vehicles with different levels of autonomy on our roads. This gives rise to mixed
However, a substantial proportion of existing research predominantly focuses on planning models for robotics that transmute the
reat them as separate modules. In this study, we employ a tree-structured policy planner and propose a dif-ferentiable joint trai
to utilize the positions and velocities of these agents and fail to capture semantic information from the scene. Moreover, to miti
we examine the robustness of motion forecasting to non-causal perturbations. We construct a new benchmark for evaluating a
balance: a method that ensures safety in occluded scenarios while preventing overly cautious behavior. We argue that such ap
ude of diverse traffic scenarios that are challenging for these planning algorithms. Model-free Deep Reinforcement Learning (D
estimated by an empirical distribution. Recent theory has shown that a rollout algorithm with a stable base policy produces a ne
tackle this challenge, a surge of research in scenario-based autonomous driving has emerged, with a focus on generating high
d domain to a new domain with limited data. In this work, we focus on end-to-end autonomous driving as the target task, consis
agent behaviors. However, existing methods have limitations in exploiting the map and exhibit a strong dependence on historic
set-based reachability analysis: First, we identify all possible driving corridors by computing the reachable set for the longitudin
ng-based methods have shown excellent and accurate performance, but the black-box nature makes it untrustworthy to apply t
the forward collision warning and unsafe lane change warning, can reduce the risk of collision caused by human errors. Howev
s, each of whom may exhibit diverse driving behaviors. Many existing methods consider other drivers to be dynamic obstacles a
approaches use non-interpretable control commands as the action space and unstructured reward designs, which are unsuitabl
e-aware lane-change strategy with high-level model predictive control (MPC) through curriculum reinforcement learning (CRL).
mputational methods such as time-series analysis. Our research diverges significantly by adopting an interdisciplinary approach
worlds. In this paper, we present ContextVAE, a context-aware approach for multi-modal vehicle trajectory prediction. Built upo
ate friendly interactions between individuals and rehabilitation exoskeletons. The exoskeleton can also meet the requirements o
nipulation tasks, easing the cognitive load of the operator by assisting them in their anticipated direction of motion. Our novel int
vements by presenting: i) an offline data collection and training pipeline, used to train a classifier for left/right motion intention p
d to be recognized correctly. In this paper, we propose a deep learning magnification method to generate macroexpressions fro
y, such as warning surrounding vehicles in the event the driver is attempting a dangerous maneuver. In this work, we propose a
g complex instructions that are more akin to natural conversation than traditional explicit procedural directives typically seen in r

ngful and systematic approach to identifying challenges for prediction models. In this regard, we also propose a new metric for m
recise and robust HMIP with low computational resources remains a challenge due to the stochastic nature of hand motion and
gh probability of impact between the UAV’s foot and the target. For a successful grasp, the foot must smoothly absorb the
igns lack the compliance and compactness inherent in their biological counterparts. A major limitation of these designs is their r
g individuals with diminished ankle function. This study highlights the advantages of variable stiffness actuators over traditional
e costly or hazardous. To address the opportunity to automate the scrubbing of surfaces, this work focuses on the use of series
s studies have overlooked the kinetostatics of these manipulators, which can result in multi-solution problem. This issue is critic
ath seriously affects the transmission accuracy. In the case of unavoidable friction, optimizing the tendon transmission path to r
m also to buckle under tendon actuation. In this work, we perform for the first time an extensive elastic stability analysis of these
ure between the masts reduces the overall stiffness of the CDPR. This paper introduces a CDPR called Rocaspect, that has fou
rofoundly influences system performance, notably in terms of friction. This study investigates the effects of cable pretension and
dealing with highly individualized and partly unknown systems. This paper unveils a cutting-edge kinematic optimizer, underpinn
tructure. The interaction force between the robot and water surface is complicated due to water deformation, and the difficulty o
We derive a range of feasible target accelerations for each intermediate waypoint of a one-dimensional path using a binary sear
up to three segments. For one-segment, a reachability criterion is proposed, which simplifies the calculation of the neighboring
. By modeling the start and goal configurations of the path as the base frame and final frame of an RRPRR manipulator, we trea
uration to task space, but the Capacity margin, a polytope based kinetostatic index, provides additionally, both an accurate eva
k-and-place. On the other hand, designing an optimal robot for one specific task consumes large resources of engineering time
approaches in the literature represent the body’s trajectory as a point trajectory, considering only its translation and neglect
is a very classical area of study that is broadly applicable throughout engineering, and significant effort has been devoted to the
on. Given the task of placing a known object outside of the workspace of the robot, the robot needs to come in contact with it a
idges to entire pallets of goods. Many of these systems are under pressure to maximise the use of space by growing in height
tesian space have some shortcomings due to its high nonlinear kinematics. Owing to its unique four-to-two leg structure, the m
f the robot with the Irvine's model cannot be done by checking the spectrum of a stiffness matrix associated with the platform m
aper, we propose a numerical approach, referred to as SEECR, which is designed to estimate the behavior of a CDPR featuring
h and twist. Traditionally, these two quantities are evaluated independently from the actuator's torque and speed limits, which a
ergy-efficient mode of locomotion observed in primates, has inspired the development of RicMonk to explore versatile locomotio
trajectory tracking of the actuated subsystem and balance of the unactuated subsystem can be achieved simultaneously. Howe
er the feasibility of the control policy, which might cause over-conservativeness and even fail to achieve the control goal. To sol
xtensive attention for safety-critical control. However, due to the myopic one-step nature of CBF and the lack of principled meth
ible failures during skill learning increases the complexity of the process and restrains learning an optimal policy. Nonetheless,
olic reasoning about probable future actions and the ability to tie these actions to specific locations in the physical environment.
ainst disturbances. We used a hierarchical model consisting of low-level recurrent neural networks (RNNs) for processing each
to object grasping, resulting in significant shifts in the tool-tip position and the body's center of gravity. Moreover, this deflection
failures, which can be particularly costly in real robot applications. Due to their interpretable nature, causal Bayesian Networks
s requires expensive network training and extensive tuning of a large number of hyperparameters. On the other hand, Covarian
 of an autonomous robot for laparoscopic surgery. Robotic laparoscopic surgery presents two main challenges: (1) the need to
the ability to provide guarantees regarding their learned behaviors, which is critical for avoiding failures and/or accidents. In thi
ents to new task scenes. A major goal of MP research is to improve the adaptability of MP to various target positions and obsta
demonstrations given through imitation learning. A graph representation of the state space is considered to encode the task-rele
enting the demonstrations into movement primitives. Therefore, we propose an unsupervised motion segmentation method that
arning, which can be time-consuming, or learning from human demonstration, which requires human involvement. In the latter c
n of object properties necessary for generalising learnt manipulation tasks to unseen objects, existing Learning from Demonstra
y involve convexifying obstacles and presume Gaussian distribution, which are not universal. Meanwhile, the fast convergence o
stems is challenging because of safety requirements: besides executing correct actions, an autonomous agent must also avoid
ting action primitives, such as push operations, their pre-conditions and effects would enable strong generalization to unseen w
n, but such methods introduce delays. Alternatively, series clutch actuators are easily backdrivable without delay. However, ver
y demanding fewer expert examples and eliminating the need for information about the actions taken during the demonstration
en devoted to relating these observed forces to surgical techniques. Answering questions like “Can certain changes to a su
tness to perturbations and adaptation to moving targets with non-linear function approximators to fit a wide range of demonstra
mains relatively unexplored. As these manipulations require joint-level control to regulate the complete poses of the robots, we d
of-the-art constraint inference algorithms learn constraints from demonstrations, but tend to be computationally expensive and
ess in LfD often relies on kinesthetic teaching or teleoperation as the medium for users to specify the demonstrations. Kinesthet
ative placement tasks from just a small number of demonstrations, when using relational reasoning networks with geometric ind
hey often exhibit unpredictability in unexplored regions of the state space, thereby raising major safety concerns when facing p
at involve static obstacle avoidance, we propose incorporating barrier certificates into an optimization problem to learn a stable
his is particularly difficult for visuomotor policies that utilize high-dimensional images as input, particularly when those images ar
re reactivity and robustness to disturbances, we propose a novel approach that selects a target point at each time step for the r
of striking a delicate balance between achieving precision in learning and ensuring the overall stability of the system. In respon
ice the demonstration learning accuracy. This article solves the issue by learning a stability certificate, represented by a neural
s from an ill-posed nature, leading to inaccurate inference of constraints from demonstrations. To figure it out, we introduce a tr
amework referred to as experience-driven random tree connect with human demonstration (ERTC-HD), which can facilitate the
gent learns from a fixed dataset. We hypothesize that model-based RL agents struggle in these environments due to a lack of l
e been successful in this task but still struggle in the context of robotic manipulation. As it requires impacting a possibly large se
sses, as they engender considerable variety and uncertainty in the parameters (e.g. hardness) of materials which the robot mus
o requires tedious human annotation of hundreds of thousands of instructions. Thus, we propose SPRINT, a scalable offline po
, in order to learn effective state representations, most current vision-based RL methods based on contrastive unsupervised le
lity. They employ approximate models, which lead to compromises in performance. Data-driven approaches have been shown
d RL solutions for vision-based real-world applications require bridging the sim-to-real gap for any world model learnt. Due to its
nd planning through both physical state and latent intent uncertainty. In this paper, we transform this intractable problem into a
d by the quality of the offline dataset, so offline-to-online RL is a more realistic setting. Online decision transformer (ODT) is an
racteristics add to this adaptability, they also pose great challenges to robot control. Instead of perceiving the hyper-redundanc
n. Reward shaping is a popular solution, but existing methods often rely on value functions, which face scalability issues. This p
ts broader application in robotics. While data augmentation appears to be a straightforward solution for improving sample effici
nd can be time-consuming. We propose HyperPPO, an on-policy reinforcement learning algorithm that utilizes graph hypernetw
, safety, and overall training stability, which limits its practical applicability. We present APRL, a policy regularization framework
shown to be more sample efficient and intuitive to combine with other planning algorithms. However, for position control gain t
 a single neural network with concatenated observation input. However, the corresponding learned policy is highly task-specific
. As a follow up to our original dataset, we collected seven hours of data at speeds of up to 15m/s with the addition of three new
tion (UOD) in Autonomous Underwater Vehicles (AUVs), remains relatively unexplored. It may be attributed to several factors: (
nalysis methods, such as discrete element analysis, poses significant challenges when applied to largescale terrains. To addres
n to the limited sensing and compute capability present on affordable, small-scale mobile robots. To address this issue, we pres
d D-GLSNet that directly matches the LiDAR point clouds and satellite images through end-to-end learning. Without the need fo
eve this, the robot system's traversability is usually estimated to plan traversable routes through an environment. This paper pre
that avoid local optimum by designing heuristic functions or high-quality global paths in wild environments are time-consuming
annotated real-world human datasets has been a laborious task, especially for highly interactive scenarios. On the other hand,
d propose a transferring approach with a parameter-compositional formulation. We investigate ways to improve the training of m
distant goals in temporally extended tasks remains a challenge for GCRL. Current works tackled this problem by leveraging pla
ing box) although such ground-truth information is not readily accessible in reality. To address this, we introduce KINet (Keypoi
we propose the Intrinsically Guided Exploration from Large Language Models (IGE-LLMs) framework. By leveraging LLMs as a
non-contact states, it is important to address the partial observability of tactile sensors and to properly consider the history of o
to specify different tasks, which limits the multi-task learning performance and can not generalize to new tasks. Thus, we adap
cise motor skills, and the complex interplay between the two. In this work, we present Teacher-Augmented Policy Gradient (TAP
shifting gears. CIPs have two primary design goals: to minimize what must be learned by exploiting structure present in the wo
reactive targets alter their behavior in response to the motion of the manipulator. Reactive applications range from gently captu
hallenges via data-driven meth- ods, learning dynamical models in combination with model predictive control. Those methods,
tly exhibit limited generalization and narrow distributions, which restrict their effectiveness in multi-task training. Addressing the
fer (TEST) framework, a pioneering offline reinforcement learning (RL) approach that incorporates tactile feedback in the contro
n manipulate unseen and diverse objects? In this paper, we propose to perform Reinforcement Learning with only visual tactile
earrangement and the high-dimensional nature of dual-arm planning. To address these challenges, we formulated the problem
hieving this level of generalization. Our insight is to add equivariance to both the visual object representation and policy architec
DRL-trained controllers are vulnerable to even the smallest of perturbations on its inputs which can lead to catastrophic failures
e deformation models that describe the relationship between robot motion and DLO deformations. Such models are difficult to
tificial neural networks suffer from catastrophic forgetting and may forget the previously visited areas when trained with new arr
everage knowledge at different abstract levels and is thus preferred in complex mapless navigation tasks. However, it is compu
der to achieve diversity. In many scenarios, task requirements are specified as a multitude of reward terms, each requiring a di
locomotion control framework for snake robots that integrates tactile perception to augment their adaptability to various terrain
ditions. Therefore, radar sensors have recently gained attention due to their intrinsic robustness to such conditions. In this pape
RF often struggles with geometric distortions in indoor reconstructions. To correct geometric distortions, we develop VPE-SLAM
uristic-based methods, based on radio frequency (RF) propagation, are intuitive and generalizable across simple scenarios, yet
ds solely rely on deep neural network feature encoders to extract features from raw LiDAR data, lacking an explicit representat
vations using onboard sensors. Recent works have shown that predicting the structural patterns in the environment through lea
HRC under the constraint of guaranteed safety is yet to be made. We, therefore, present human-robot gym, a benchmark suite
single policy to simulate pedestrian movements and could suffer from the over-fitting issue. Alternatively, framing the collision a
ciency. How can autonomous vehicles balance different driving objectives and human values in a constantly changing environm
s challenge, this work proposes a new Fast and Safe Policy Optimization (FSPO) algorithm, which consists of three steps: the f
n, the learned behaviors fail to achieve the desired transformation invariances and suffer from motion artifacts. For instance, a
rces to targets with the minimum total completion time. To achieve the goal, the core idea is to develop an effective object-to-a
which impedes exploration and restrains the overall performance. In many realistic tasks, e.g. autonomous driving, large-scale e
game using model-based reinforcement learning techniques. The method involves extracting low-dimensional observations from
ch to exploration tasks is to combine active mapping with planning. Metric maps capture the details of the spatial representation
teroceptive depth measurements, e.g., from LiDARs. However, the measured depth fails to align with the true support surface i
manipulator can be an asset for legged locomotion at high speeds or under external perturbations, where the arm serves beyo
ciently. In this study, we tackle these issues by relying on a simple yet effective two-stage learning framework to generate dyna
al controller parameter space to a lower dimensional space using a machine learning-based algorithm. This mapping is then util
e, allowing experts to be composed sequentially. Each expert specializes in a single skill, such as a locomotion gait or a jumpin
ghtweight quadrupedal robot that unlocks the agility of the quadruped in an upright standing pose and is capable of a variety of
exist, this task remains particularly challenging since generative models can produce physically unrealistic motions, which will
an easily lose their locomotion functionality if damaged by external accidents or internal malfunctions. In this paper, we propose
threat. Typically, controllers lack the capability to actively sense the condition of their own joints and take proactive actions. Ma
arn because they involve both high-level planning and low-level motor control. We present a novel framework, Cascaded Comp
recovery strategies, we propose a simple modification to the typical reinforcement learning training process. Our key insight is t
and Finetuning Model Predictive Control (IFM) framework to take the strengths of both approaches. Our framework first develo
foot reflection helps the robot to navigate intricate terrains without foot entrapment. In this paper, we manage to bring both asp
ing similar challenges. Classically, this is done by independently engineering perception, actuation, and control systems to very
plicitly modeled by currently deployed locomotion controllers for legged robots. In this work, we propose a risk sensitive locomot
ious methods try to integrate privileged distillation, scene modeling, and external sensors to improve the generalization and rob
terms, such as an energy minimization term, to guide an algorithm like Proximal Policy Optimization (PPO) to good performanc
, such as rugged rocks, non-rigid ground, and slippery surfaces. However, while these controllers primarily address challenges
obustness quantification. This marks the first-ever study to apply STL-guided trajectory optimization for bipedal locomotion push
the ankle, hip, and footstep adjustment. The Nonlinear Model Predictive Control (NMPC) is formulated to take into account non
e a novel metric for locomotive robustness -- the estimated size of the hybrid forward invariant set associated with the step-to-s
same will likely be the case. In this work, we propose a locomotion controller for efficient walking over near-flat ground using a r
of Motion (3D-DCM) framework, which was successfully used for generating walking trajectories in previous works, retains its v
mework utilizes an autoencoder to learn a low-dimensional latent space that captures the complex dynamics of bipedal locomot
on an iterated extended Kalman filter. Beyond state estimation, the foot contact position is also modeled and estimated. This all
des custom-built high-efficiency actuators and transparent, sensorless transmissions. The robot’s reinforcement learning-b
al vehicles so that they can carry heavy loads: heavy enough that flying is no longer possible, e.g., when the thrust-to-weight ra
und reaction force, which can be interpreted as an observer feedback gain. We interpret the oscillator itself as a latent contact s

quadruped animals can still stand or walk even with three limbs. This paper investigates the optimization of dynamic balance du
in three stages: 1) synthesize a physically plausible reconstructed key-point trajectory from monocular videos; 2) optimize a dy
ngth, the locomotion robustness and safety are reduced since more dynamic motions are required to traverse it. In this paper, w
pace-qualifiable components. Furthermore, SpaceHopper's design and controls make it well-adapted for investigating dynamic
w robot, where hyperparameters and reward function weights must be re-tuned to maximize performance for each new system
estigates the relationship between different gait sequences and the motion stability of quadruped robots, assuming a flat terrain
quipped with a dedicated robot arm, which means additional mass and mechanical complexity compared to standard legged ro
To address this issue, we propose a model-based supervised learning framework that combines a world model with a policy ne
which degrades its performance. To address this issue, we propose the controller of integrating a data-driven error model into
 onnected network of reduced-order models, based on the single rigid body (SRB) dynamics, is developed for trajectory plannin
he RL training process by incorporating on-demand reference motions generated through finite-horizon optimal control, coverin
f a planar 3-DOF X-shaped walker with a reaction wheel is introduced, and the mathematical equations are described. Second
putation cost of using the full-body dynamics of the limbs, a reduced-order model that can simulate the dynamic characteristics
nking) by unifying cutting-edge model-based trajectory optimization and iterative learning control. Using a reduced-order soft a

wnstream tasks such as footstep planning. The maps generated by existing mapping algorithms are either sparse, insufficient
s acceleration and a constrained amount of time before transitioning to a new domain (e.g. foot touchdown). Meanwhile, these t
esent significant complexities in accomplishing skateboarding behavior.This paper focuses on enabling the quadrupedal platform
planner incorporates locomotion-specific modules into the high-level contact sequence planner and solves kinodynamically-aw
er while maintaining its ground locomotion capabilities. We designed the specific reinforcement learning method to train the ne
d by many of these robots is their limited capability to navigate stairs, which hampers their functionality in multi-story environme
es have indicated the legged robots which efficiently utilizing their spine can achieve rapid and stable locomotion. However,with
ts currently cannot provide high-quality training data regarding consistency and accuracy. To this end, we present CAMA: a visi
ry and SLAM, to advance the research in autonomous robotics. This work complements existing datasets by simultaneously ad
existing lightweight methods often encounter the challenge of limited representation capacity and increased computational res
nes of work on VPR have shown, on one side, that general-purpose off-the-shelf feature representations can provide robustnes

o have small distance for similar images and larger distance for dissimilar ones in a latent space. However, this approach strug
ed by visual odometry but also for preventing the creation of ambiguous maps in unfamiliar environments. Despite numerous re
m performance. Current methods mainly utilize single-modality RGB images, which are sensitive to environmental changes. To
s have attracted increasing attention in HD map construction task, which can significantly boost the benchmark. However, exist
er, ultra-efficient visual-inertial estimation at the size, weight, and power (SWAP)-constrained edge devices to provide accurate
ework that capitalizes on recent advances in deep learning to transform 2D camera images into 3D semantic occupancy, thereb
multi-volume neural feature fields, called NF-Atlas, which bridge the neural feature volumes with pose graph optimization. By re
w-visibility and nighttime scenarios with varying illumination conditions. Addressing this challenge, we introduce an algorithm des
istic textures. This system adapts the state-of-the-art neural radiance field (NeRF) representation to also incorporate lidar data
cient map, which lacks robust data association approaches. And it is not suitable for precise localization or long-term map maint
he localization accuracy and navigation performance of the robot. Therefore, dynamic objects removal plays an important role f
h semantic and metric information show attractive advantages. In this paper, we present RO-MAP, a novel multi-object mapping
ent mapping method using volumetric voxel grids that represent the probability of occupancy. The main bottleneck of current CP
e detection of traversable regions and the modeling of physical attributes of staircases by point cloud data. To mitigate the influ
-road navigation, strictly-supervised learning approaches remain limited in their generalization ability. To this end, we introduce
onment comprising a combination of traversable and non-traversable spaces, which presents a challenge for estimating travers
ring these navigational patterns from incompletely observed environments is required for intelligent mobile robots operating in u
mera sensors are sensitive to lighting and weather changes, and even slight image alterations can greatly affect VPR efficiency
ry consumption, we utilize a sparse hash table for storing hierarchical 3D voxel maps. To improve the processing cost of BBS in
nstance-aware static mapping framework called DynaInsRemover, which exploits the geometric discrepancies between instanc
agent must search for and navigate to the sound source with egocentric visual and audio observations. However, the existing
hat efficiently learns a label field that may be rendered from novel viewpoints to produce view-consistent instance segmentation
paper introduces a method leveraging Salient Object Detection (SOD) to extract these unique features, employing them as pivo
 ue to the increase in map size, which will affect downstream tasks such as robot navigation and services. To this end, we propo
res are less sensitive to weather, season and surface scene changes. Here, we propose a novel subsurface feature-based loca
t lack crucial scale information, and over time, accumulation of images leads to inevitable drift issues. In contrast, mapping met
the robustness of LiDAR-inertial odometry in geometrically degenerate scenarios, like tunnels or flat fields. We project LiDAR in
ased Stein variational gradient descent (SVGD) with iterative neighbor particle search. This method uses SVGD to collectively u
U factors on a sliding window factor graph. The tight coupling of the scan-to-scan and scan-to-map registration factors enables
mited field of view cameras under 180 degree rotations, has proven to be challenging to achieve. To address this problem, this
method. It upgrades the 2D features extracted from the source image to 3D features with the help of a reference image and ma
ertain benchmarks and struggle to adapt effectively to changes in the application domain or environmental conditions. This rend
long-term tracking. This challenge presents itself in two main aspects: first, the difficulty for robots in revisiting previous location
tion and depth estimation, incorporating specifically designed scale-consistency constraints that utilize estimated depth as a re
aper presents a novel model, called UMF (standing for Unifying Local and Global Multimodal Features) that 1) leverages multi-m
ric structures in challenging scenarios. This paper presents RELEAD, a LiDAR-centric solution designed to address scan-match
o the limited ability of backbone networks and extreme environmental changes, current methods fail to capture foundational sem
canning pattern makes it possible to achieve a consistent LiDAR-inertial estimator by employing a frame-to-frame data associa
a semantic map of the local scene from the image sequence, then extracts points at a fixed height for registering to the LiDAR
based on the map. However, when it comes to using it for robot localization, large geometrical discrepancies between the wayfin
ses. However, spatial and temporal clues from sequential images still remain underexplored, resulting in inaccurate poses and l
most accurate approach is a hierarchical one that splits the task into two stages: image retrieval and camera pose estimation. E
the mapping from image pixels to 3D scene coordinates, which requires a vast amount of annotated training data. We propose
as access to a continuous stream of frames, this task is naturally cast as a sequence-to-sequence localization problem. Neverth
ng techniques, the performance of LiDAR place recognition (LPR) has been greatly improved. However, current deep learning-b
modalities, we propose a hybrid solution that combines proprioception and exteroceptive information for estimating the state of
s opposed to a common frame-to-frame retrieval problem formulation currently implemented in SOTA place recognition method
de codes a colon location by a set of real images, while edges represent traversability between nodes. For close-in-time images
f existing localization systems compatible with this modality, beyond fluoroscopy and its harmful ionizing radiation, impedes its t
xhaustive searches, which is time-consuming. In this paper, we present a novel coarse-to-fine approach to address these probl
y challenges (e.g. difficulties of map storage, poor localization robustness in large scenes) in accurately and efficiently impleme
oal to obtain a map representation that is well suited for robot localization. To this end, we utilize a neural network to learn a dis
features introduces false candidate matches and brings unique challenges to this task. In this paper, we propose a sequence-
mapping (SLAM) are typically adopted. Due to error accumulation, visual SLAM (vSLAM) usually suffers from long-term drift. Th
e traditional methods like SLAM, recent strides use deep learning for direct end-to-end pose estimation. We propose EffLoc, a
ation. This paper proposes a novel semantic information-assisted ICP method named SAGE-ICP, which leverages semantics in
mation on the estimated pose, alleviating the impact of these unreliable predictions. However, existing uncertainty modelling tec
nted/virtual reality applications. Inspired by latest advances in deep learning, many existing approaches directly learn and regre
ymmetry of the spaces, visual localization methods commonly confront difficulties in estimating precise 3D poses. In this study,
ge efficiency, which increases the difficulty of direct visual place recognition in compressed Lidar maps. This paper proposes V
ties are well suited for robotics tasks, SNNs have seen limited adaptation in this field thus far. This work introduces a SNN for V
ained normal equation leads to sub-optimal results. Both of them influence accuracy of the estimated pose. In this paper, we th
ed environments (predominantly urban driving), their performance degrades severely in unstructured environments, rendering m
GPU acceleration. Such computers are ubiquitous on small indoor robots and thus this work enables high-precision, millimeter-
d field of view (FOV) and/or constrained motion, thus degrading the localization performance. To ad- dress this issue, we design
nocturnal visual localization is still an open problem owing to the lack of stable visual information. An insight from most nocturn
e. This paper presents a novel GPS-VIO system that is able to significantly benefit from the online calibration of the rotational e
ng environment. This is a particularly challenging task due to the limited sensing and compute resources on board. This work pr
corresponding depth maps, inertial measurement unit (IMU) readings, channel response between a 5G massive multiple-input a
tive pose observations between them. Previous works, however, typically assumed known ego motion and ignored biases of th
ever, this combination with their associated costs and infrastructure demands, poses challenges for widespread adoption in ma
obots to combine sensor data. The two critical drawbacks of KFs are the requirement for skilled tuning of several filter paramet
o fixed-lag smoothers (FLS) to decouple accurate state estimation and high-rate pose generation for closed-loop control. The fi
he inertial navigation system (INS). While modern implementations achieve remarkable robustness, it is often due to the carefu
opagation of radio waves. This study tackles a fundamental issue in wireless ranging: the unpredictability of real-time multipath
f floor plans in terms of availability and timeliness of updates challenges their wide applicability. In contrast, the proliferation of s
s, existing methods typically resort to matching all possible combinations of detected objects and landmarks with the same obje
multiple observations. Sharp-peaked likelihood functions commonly arise from millimeter-accurate distance observations of car
e office spaces and corridors using 3D LiDAR point clouds in a given Area Graph, which is a hierarchical, topometric semantic
owerful spins. The Magnus effect induced by spin adds complexity to trajectory prediction during flight and bounce dynamics up
ng the sensor's velocity from a single scan. The vehicle's 3D motion, defined by its linear and angular velocities, is calculated ta
velocity relative to the sensor. 4D radars in particular are interesting for object perception and navigation in low-visibility condit
nt cloud domain, the topic has been extensively studied as a global descriptor retrieval (i.e., loop closure detection) and pose re
real-time is crucial for evaluating system performance and enhancing downstream operations such as data association. Howev
generate online trajectories that are reactive to human motion, modulating the speed of the robot. The PL allows for tuning the
driven paradigm of navigation approach with safety verification is preferred where one can train RL algorithms with large amoun
hases. The initial phase is centered on a novel sparse Gaussian process (GP) framework. We first integrate a forgetting factor
edom. This work proposes DeformNet, which utilizes latent space modeling with a learned 3D representation model to tackle th
d formulations - with the robustness and online replanning capabilities of model predictive control (MPC). This paper provides an
del-based control approaches is still an open problem. Although these models have conventionally been evaluated for open-loo
ations to train a manipulation policy for each object, which hinders their applications in our ever-changing world. To address this
to lift boxes of varying weight, position, and orientation while maintaining balance. To address this challenge, we present a sim
struggle to capture model variations due to a limited set of parameters. Data-driven dynamics learning approaches offer highe
FRL), which is incredibly flexible and general but often results in brittle policies. In contrast, model predictive control (MPC) con
uctured off-road environments can be highly complex and difficult to simulate accurately. Domain randomization addresses this
d, and incorporate auxiliary data, such as demonstrations and prior experience. However, despite these advances, robotic RL r
en present in robot dynamics: 1) sparsity, i.e., some components of the state may not affect the dynamics, and 2) physical limit
ing (RL)? Offline RL methods, which learn from datasets of robot experience, offer one way to leverage prior data into the robo
ing information regarding the planet’s internal structure. Steep slopes consisting of low-cohesion regolith are prone to flow
conservative policy to avoid unsafe situations. However, overly conservative policy severely hinders the exploration, and make
egration of different manipulation actions for optimally solving long-horizon task-and-motion planning problems. In this study, w
placing the objects of interest directly, object reorientation is needed for precise placement in most of the tasks. In such scenari
ForceSight determines a target end-effector pose in the camera frame (kinematic goal) and the associated forces (force goal).
om confined spaces, which is unsuitable for conventional visual perception and gripper-based manipulation, is identified and ad
he task may involve threading through or avoiding becoming tangled with objects like racks or frames. Grasping with multiple g
cts the per-point actionable score and thus proposes the best point to interact with, has demonstrated excellent performance an
f-interest (OoI) which it intends to grasp and extract from the workspace, it must reason about which movable objects to rearran
ask, the robot needs to (i) detect the incoming projectile, (ii) predict the projectile's future motion, (iii) plan a minimum-time rapid
and have worldwide coverage, offering a scalable alternative. In this work, we systematically explore the effect of SD maps for r
resentations. However, state-of-the-art methods still struggle with excessive memory usage and non-smooth surfaces. This is p
mapping with implicit neural representations remains problematic due to low efficiency, limited video memory, and the catastrop
Several works have utilized NeRFs for refining camera poses, but they do not account for lighting changes that can affect scene
hallenging problem of how to achieve a concise and uniform representation robustly. We employ quadrics to represent diverse
map distortion and long traces, thereby posing challenges for accurate mapping and navigation. To address this issue, we prop
n this paper, we propose a NeRF-based mapping method that enables higher-quality reconstruction and real-time capability eve
nstruct high-quality object-level maps for unknown objects. Our approach takes multiple RGB-D images as input and outputs de
mage-as-node based graphs lack explicit object-level reasoning and interconnectivity. In this paper, we propose a novel topolog
ch 3D environments, without requiring prior traversals to map out these scenes in detail. We posit that this is enabled by the ab
grid form and suffer from the grid size problem, wherein a large grid size is unfavorable for motion planning while a small grid s
moving objects in the current observation and at the same time to also update the internal model of the static world to ensure sa
asks in such systems. As most of the today's mechanical LiDAR sensors produce sparse data, estimating normals from a single
monly used in 3D reconstruction tasks relying on cameras and LiDARs. However, fusing data from these two types of sensors i
n of an ego-robot with map information, the quality of the reference maps is often neglected. To avoid potential problems cause
efficiency challenges of occupancy mapping. Firstly, we use a depth image to determine the occupancy state of regions instead
g accurate localization and effective navigation. To tackle the problem, a novel map construction framework based on 3D regio
ustment (BA). Modern 3D LiDARs now retain higher resolutions that enable the creation of point cloud images resembling those
cle (µAUV) design. This robot is distinguished by its utilization of mass-produced drone components, novel jet propulsion mec
they can move with dexterity. Here, inspired by the versatile motion abilities of fish, a novel robotic fish has been developed, fe
ampling in the deep sea and thus facilitate the explorations of rare deep-sea ecosystems, the goal of this research is to design
tonomous or even semi-autonomous systems, most UVMS still require human support teams. Developing new hardware and a
m both the water surface and ground. A foldable and self-deployable arm mechanism linked to and driven by a piston variable b
e abilities in robotic fish is challenging; existing designs focus on either fast swimming or directional control at limited speeds, m
ese aspects are contradictory sometimes. In this paper, we present an omnidirectional swimming robot inspired by the whirligig
es the disadvantages of state-of-the-art methods for sediment sampling in terms of increased station-keeping capabilities, low d
rain interaction forces. However, using exteroceptive sensors such as cameras and sonars to recognize underwater terrain's de
ology and computational capacity, AUVs are expected to operate safely and autonomously in increasingly challenging environm
easurements from triangulated features to improve the depth predictions and solve the problem of scale ambiguity. To allow pri
mong many factors, limited visibility, light scattering, cluttered environments, and constantly varying water conditions. An appro
C (SONar Image Correspondence), a pose-supervised network designed to yield robust feature correspondence capable of wit
wever, a central challenge persists: calibrating the prediction confidence of the model to ensure robust and reliable outcomes, es
ch, we propose an algorithm to restore the true color (albedo) in underwater imagery by jointly learning the effects of the mediu
e annotated training data by preparing a comprehensive dataset for semantic segmentation of underwater cave scenes. It cont
of fish and other organisms with their habitats on and around coral reefs. Our system, designed for the CUREE AUV, uses fou
parseness of point clouds. However, the node context within the tree is not fully discovered since the feature space among nod
ng. To this end, an augmented catenary- based model is introduced to deal with the dynamical effects of surge motion, sway m
models of fully submerged UV behavior are well-established, they are unable to model the ubiquitous condition of a UV operat
 rch for the dynamic window facilitating takeoff in waves and avoiding hazardous waves. For the first time, the cross-domain win
and pose. AUVs utilizing thrusters can be highly maneuverable, making them well-suited to operate in complex environments s
ar choice for trajectory tracking problems as it naturally addresses operational constraints, such as dynamics and control const
s environment. Robots that work in water need to minimize the number of drive components to improve reliability and reduce co
s a hemispheric dome window. We then observe optical issues including the distortion effects of the lens, e.g., wide-angle field-
d data are still performed using classical methods in the iterative closest point (ICP) family. In this work, we curate and release
y for these systems to reduce intervention operating costs. Developing autonomous UVMS capabilities is challenging, however,
multi-objective behavior optimization is used to model individual decisions. Using previously reported theoretical results, we show

order to facilitate AUVs for long mission durations without ship support, they can be equipped with docking capabilities to recha
sed in recent years. However, it may be difficult and unsafe to follow COLREGs in congested waters, where multiple ASVs are
ronments or with limited-range sensors. Pose estimation can be improved by incorporating the uncertainty prediction of future p
ypical AMVs cannot meet the diverse demands of different marine tasks. In this article, we design a new type of remote-contro
. In this work, we address the problem of choosing the most informative views that minimize survey time while maximizing clas
nance tasks. For delicate repair operations such as fixing tiny cracks, most existing underwater manipulators face limitations in
60 Mechanical Scanning Sonar (MSS) and a A50 Doppler Velocity Log (DVL), which are low-cost devices for their respective ty
static targets, the research gap in underwater docking to dynamically active targets has been left relatively untouched. We add
or or navigating in confined waters. We introduce the concepts of area water clearance and line water clearance. Area water cle
ion for the AUV to recharge its batteries and receive updated mission information. Various methods for locating and tracking a
a camera and a light detection and ranging (LiDAR) sensor to detect navigable regions in the waterway. A deep learning-based
respondences or calibration targets. It is separable as a distortion model (dependent on refractive index n and radial pixel coor
g re-projection errors, relative inertial errors, and relative magnetometer orientation errors. The results of IMU orientation propa
ter visual-inertial navigation system being inappreciable by tightly coupling the long baseline (LBL) system into an optimization-
sitive, sonar beamwidth. Our approach models the free-space covered by each emitted acoustic pulse, employing volumetric te
d or unlocked. This joint-locking feature allows effective balancing of the payload capacity and dexterity of the robot but significa
per mechanism affixed to the aerial robot’s top, facilitating overhead drilling. The primary aim is to enhance the safety and e
sing dual quaternions for UAVs with cable-suspended loads, focusing on the sling load lifting and tracking problems. By utilizing
method is also cumbersome when loading cargo, and the size of the cargo to be loaded is limited. To overcome these issues, w
al interactions, which offers a way to assist humans in high-altitude and high-risk operations. Tactile sensors, being both cost-e
hance its performance and extend its working time, which is equipped by almost every kind of bird. As a common method of ta
perching and unperching autonomously on/from a ferromagnetic surface during flight, and a switching controller to avoid rotor s
rescue. These systems, which may consist of single or multiple vehicles, can be equipped with passive connection mechanism
nism on vertical surfaces. The suction cup interaction with different surface types is characterized with experimental tests to ac
e rescue area. However, during the fast descent of the rotorcraft, a chaotic flow field rampages as the rotorcraft falls into its wa
h other. In particular, the work sheds light on the impact of the uncertain knowledge of the model parameters available to an est
d flight endurance hinders their application. Inspired by natural perching behaviors, we propose a multi-modal aerial robot syste
orm and the kinematic error amplification effect from multi-link structure of the manipulator, it is extremely challenging to ensure
mical forces the MAV is subject to, with a minimal set of parameters. We only rely on commonly available sensor information on
or optimization-based control. The phi-theory methodology imposes a specific geometry on aerodynamic coefficients that leads
ng to its silent and solid-state nature. Like rotors, EAD thrusters exhibit changes in performance based on proximity to surfaces
ifferent bounds for the sinistral and dextral curvatures, and penalties μL and μR for the sinistral and dextral turns, respective
 A* heuristic search. The Value of a Prevented Fatality provides a basis to convert fatality risk to monetary value terms as a Hu
large adhesive force. To seal on a surface, suction cups must be aligned with the surface and possess proper relative tangentia
enon, trajectory tracking can now be implemented on Monocopters via feed-forward terms that are computed per the trajectory
between the aerial robotic system and the environment. End-effectors with multiple contact points are often used for placing ND
orates two off-the-shelf quadrotors passively articulated to a rigid frame. This innovation enables free pitch rotation without mec
r with stable embedding that extends the system into Euclidean space while ensuring stability of the manifold. Our particle filter
olicies. However, a critical hurdle is the sim-to-real gap, often addressed by employing a robust inner loop controller —an abstr
d body. In contrast, the proposed robot employs passively reconfigurable structures to enable the second mode of locomotion,
morphing) drones represent an efficient solution. However, morphing drones require the addition of actuated joints that increase
sturbance observation and compensation, full attitude controllability, and instant output power generation of the quadrotor. In th
execution of high-speed yaw maneuvers. In this approach, we minimize the cost of yawing measured with a performance figure

pread availability and relatively low cost. Industrial infrastructures often consist of diverse oriented work surfaces. When interac
e this paradigm by proposing Geranos: a multicopter unmanned aerial vehicle (UAV) adept at precisely placing vertical poles, b
drone. It results in sudden disturbances to the aerial manipulator when switching the robotic arm or interacting with the environm
drones is another important area of recent research. However, efficient coordination algorithms for collaborative transportation
acteristic limits its applications, especially when it comes to generating desired wrench with six degrees of freedom (DOF). The
platform, modeled as a spherical double pendulum. A state estimator is designed in order to obtain the deflection angles of the p
as weather conditions, communication reliability, and battery autonomy. This paper introduces the MOAR path planner which a
ns. The proposed actuation module adopts a non-contact magnetic coupling mechanism to transmit rotational power with a pre
nerating holding torques with low power consumption. A key limitation of electrostatic clutches has been their low specific shea
he bionic foldable wing is established, yield the motion law of the crease angles and the variation relationship between the pane
ed over the past decade, are based on vibration mechanisms built on flexible legs that enable motion on the ground. However,
plex mechanical devices lacking in compactness, ease of manufacturing and accessibility. In addition, the scarcity of commerci
be able to grow by adding at their tip Transformable Modules (TMs). We provide a two-module implementation to evaluate the
of the mechanical design of the PULSE115-60 actuator is presented, with particular focus on the design parameters affecting th
d improve their activities of daily living (ADL). First of all, a reconfigurable exoskeleton for upper limb rehabilitation is designed i
eedom on one surface. This flexible display mechanism is expected to be placed inside a car dashboard as a human interface
environments of logistic order picking. The finger part of the gripper was configured with the parallelogram remote center of mo
to detach and attach the robot body with an arm-type robot, and the system to manage the connection state of modularized bo
n of kirigami-origami. The synchronous motion of each module is improved by using slider-crank mechanisms. Using screw the
sporting objects, a novel transformable multi-mode Mecanum-wheeled mobile robot is proposed in this paper. Owing to its uniqu
mitigation, and enhanced balancing. However, creating such mechanisms for legged robots has been challenging due to increa
gy-saving and convenient scheme for deploying detectors. However, due to the limited pulling force provided by a single autono
e efficient and excellent wiping performance than conventional cleaning methods for LiDAR sensors. An water wiping experimen
propose a dynamical model capturing the essential dynamics of the developed transmission/robot ensemble and implement a
The robot exhibits two distinct configurations: it can fold itself for agile navigation through tight spaces, and unfold to cover larg
r, which comprises a rectangular cross-sectional flexible beam and a rotational-type motor, is introduced. Regarding the structu
heir usefulness in constrained environments for various applications. However, in traditional designs, the tip’s eversion nat
tional appendages. For example, Chukar and Hoatzin birds can repurpose their wings for quadrupedal walking and wing-assist
e focus of researchers due to its high contraction force and good controllability. However, the current performance of eSKT is fa
 ve repetitive StS exercises, but these can be physically demanding for therapists while assistive devices may limit patient partic
One uniqueness lies in that both control and tracking of the arterial robotic ultrasound end-effector have been implemented to im
drug delivery, diagnostics, and therapeutic interventions. Existing techniques commonly impart magnetic properties to the target
zed steerable dissector for this surgery, enabling easier access to curved deep artery branches. The handheld surgical dissecto
achieving precise and repeatable execution of surgical tasks has motivated the development of robotic platforms to overcome th
kinematics define a phase variable impedance controller (PVIC), which is implemented as the non-volitional base controller wit
equired precision, and in combination with intraoperative Optical Coherence Tomography (iOCT) imaging, methods for autonom
icate and bulky mechanical structures, which pose practical challenges when moving through the complex and narrow environm
ive Surgical Inc, Sunnyvale, California) has been the most widely adopted robot of this kind, with more than 6,700 systems in c
graphy, robots have the potential to reproduce the skills required to acquire high-quality images while reducing the sonographe
m section, comprising two 7-DoF surgical instruments and a 5-DoF endoscopic arm designed to enhance hand-eye coordination
ility and inconsistent image acquisition hinder the widespread adoption of broader screening programs. Robotic ultrasound sys
geons struggle to estimate instrument positions from 2D microscope frames, necessitating extensive trial-and-error experience
operator variations. However, it is still challenging to autonomously and accurately transfer a planned scan trajectory on a gene
e ribs and lungs because they reduce the clarity of US image. Meanwhile, the anatomical position and size of the heart, lungs, a
matic segmentation of small tubular structures (e.g., the ulnar artery) from cross-sectional US images is still challenging. To add
peration of occlusion, safeguarding the facial nerves and blood vessels, as well as preserving facial aesthetics. In this study, a n
l burden of doctors, which also improves the quality and safety of surgery. However, it is difficult for current commercial electric
he fundus tissue. Typical targets of this type of surgery are the internal limiting membrane (ILM) and the epiretinal membrane (E
ntrol in multiple directions is challenging for robots due to the uneven skull surface. Here a hybrid admittance control algorithm
t can provide the accuracy required for microsurgery without a high scaling factor by utilizing fingertip and wrist motions. Also, f
gical navigation is the prevailing standard, limitation still exists, requiring personalized planning and recognizing the irreplaceabl
However, most of the existing robots spray disinfectant in all areas, and they are still immature to navigate in densely populated
e not visible during operation since they are hidden by the superficial fat layer; 2) intra-operative bleeding may occur during lym
struggle with efficient, rapid adjustments in the field-of-view (FOV), especially when identifying organs and needles outside the
teraction force between the elastic wires of the parallel robots during motion generation processes, the constant curvature assu
ems such as incomplete collection of operating information and unrealistic tactile feedback. The main objective of this paper is
ody. One of the RCM configurations, one-rotation and one-translation (1R1T) RCM based on a closed-loop design, enables two
lows the target plate to move along one axis with amplified displacement in a parallel manner. The amplifier was assembled aft
owever, the deformation capability of current MCRs is limited by fixed magnetization configurations, preventing them from acces
agnetic field. The robot is composed of a shell and a targeted delivery mechanism. Under the actuation of the external magnetic
dual-segment flexible robotic endoscope MicroNeuro, designed to to perform biopsies with dexterous surgical manipulation de
curvature pose significant challenges for shape-based feedback control. The use of Fiber Bragg Grating (FBG) sensors for sha
ents. It can introduce discontinuities, may result in a complicated decision tree, has a high wall-clock time, and cannot be vector
d colonoscopy. In this letter, we presented a controllable and easy-to-use diagnostic tool, equipping a novel vibration module ins
it. Magnetic rotating swimmers have internal magnets and propeller fins with a helical shape. A rotating magnetic field applies t
ge. Therefore, most surgeries still require manual control of the endoscope and the robotic instruments, resulting in surgeons n
nals from a self-propelled vibrational capsule in contact with in-situ bowel lesions were processed and analysed for features that
is harmful for the patient and the clinician, and may lead to severe post-operative sequlae such as the development of cancer.
ness, high accuracy, and small footprint while overcoming the problems of the high cost of serial robots and the small workspac
increase the precision of such systems beyond surgeon capabilities. However, little attention has been paid to the impact of no
ot-assisted eye surgery technology provides ophthalmologists with advanced capabilities such as hand tremor cancellation, han
e precision of teleoperation but hinder the operator from quickly reaching distant targets. The concept of self-adaptive motion s
sicians frequently resort to traditional endoscopic or open surgical procedures for treatment, resulting in more complex, potentia
d. Existing camera designs mainly rely on rigid structure design, resulting in risks of losing magnetic coupling and impacting tiss
opulation. The efficacy of standard clinical eradication therapies is diminishing due to the rise of antibiotic-resistant strains, nec
is controlling the soft surgical tools and measuring the endovascular stress simultaneously. To solve these problems, a flexible
approaches need precise model parameters of FBGs inside the fiber and their alignments with the flexible robots for accurate s
, and precisely identifying tumor edges and ensuring adequate resection margins are critical for optimizing oncologic outcomes
on and a surgical robot. Particularly, utilizing the da Vinci Research Kit (dVRK) platform, we have designed and successfully de
ecent years for microsurgery are expensive and not flexible, as they cannot use standard instruments, and need the surgeon to
ely integrate robots into the surgical workflow, it is crucial to emphasize that an accessibility analysis framework for vitreoretinal
instrument selection and complex anatomy. This paper presents a novel magnetic navigation system that addresses these cha
wing to aberrations induced by soft tissue heterogeneity. In particular for beam motion control to ensure real-time and precise tr
and accessibility in passing a small port. Bendable and steerable surgical tools provides an alternative for this issue. In this wo
pulating the guidewire for branch selection, while also facing the risk of radiation exposure. In this study, we investigated robotic
livering embolic materials or metal coils to fill these bulges, occluding them from the pressure of blood flow. However, clinical m
metries (i.e., having flexural patterns). Leveraging this framework, we also introduce the concept of design space by which the
dle-based approaches relies heavily on the accuracy of needle placement, which remains a challenge even with robot assistanc
dresses the system identification challenges for getting high-quality training data associated with co-manipulated robotic arms t
magnetic soft microrobots. In this paper, an automated assembly system based on a two-fingered microhand is presented for fa
effectively apply a tensile strain on the NiTi thin-film. This actuator is expected to generate 8.7 mN of blocking force, and a free
fficient functionalities in a limited volume. Microscopic natural phenomenon is a great reference for current microrobot design, w
e create a net magnetic moment for precise position and orientation control through external fields, while harnessing opposing
sion, and steering mechanisms suitable for operations within the human body and microscale fluids. We propose the microswim
o enhance precision and safety. In response to these demands, this paper presents a novel ribbon-shaped, tendon-driven conti
many strategies, such as utilizing locally specified magnetic fields, applying electrostatic anchoring, taking the advantages of geo
, posing limitations on speed, precision, and re- producibility. Here we introduce a fully automated robotic mi- crosurgery platfor
in extracting subtle yet critical visual features. With the same system under the teleoperated mode, however, human operators
main challenges to further advance robotic knot tying technique, particularly in terms of decreasing space consumption with bett
ation setups, traditional depth acquisition methods such as depth from focus or depth from defocus directly infer depth from mic
This technology significantly enhances the efficiency of single-cell sorting and the development of antibody-based drugs. Howev
which can automatically and intelligently cut down single cells intact from tissue sections. The proposed method consists of se
y reducing size, weight, and cost pose significant challenges to the existing micromanipulator design and fabrication methods.
scale robot, the smart composite microstructure method (SCM) is getting more and more attention. This method can construct
e what functionalities these robots can achieve. Although some studies have been dispersedly reported, the forces and torques
s of multiple sperms and selection of an optimal sperm is crucial for in vitro fertilisation treatment such as robotic intracytoplasm
multiple magnifications, thereby facilitating repetitive and accurate histologic biopsies. The framework encompasses an active
ration method to correlate the workspace of the stereomicroscopy with the image and the micromanipulator. Even though the fo
Microscope (AFM) and dealing with drift issues by fusing multiple individually AFM topography patches. The merging of the patc
o be immobilized by glass micropipette tapping for subsequent surgical steps. The process requires accurate tracking of the tar
nd quantitative morphology evaluation. State-of-the-art instance-aware part segmentation networks follow a “detect-then-se
n studied in recent years and with the development of deep learning algorithms, extensive efforts have been devoted to sparse
ening, and cell analysis. However, micromanipulation of biological micro-objects, such as cells and micro tissues, suffers from m
nvasive imaging techniques combined. The need for simultaneous actuation and noninvasive localization imposes a strict const
ents. We address the lack of large-scale, accurately-annotated suction grasp datasets by proposing a benchmark synthetic data
mixed-observability, factored states, or temporally-extended actions. We propose the locally observable Markov decision proces
m associated with mobile manipulators poses challenges in achieving real-time whole-body motion planning. To bridge the gap,
he environment to their needs. In this work, we introduce a novel interactive multi-object search task in which a robot has to ope
esired location while avoiding static and dynamic obstacles. In contrast to existing approaches, our focus is on fast online plann
ontrollers to uncertain contact dynamics is a significant challenge. Adaptive Model Predictive Control (AMPC) is an approach fo
e effective observation of the target while approaching it for grasping. In this work, we propose a graspability-aware mobile man
Compared with wheeled mobile manipulators, legged mobile manipulators have higher requirements for cooperative control of
en equipped with onboard sensors, e.g., an embodied camera, extracting task-relevant visual information in unstructured and c
in the operating conditions, the model should be continuously refined to compensate for dynamics changes. In this paper, we p
at exhibit plastic behavior, like modeling clay or bread dough, is also difficult because they permanently deform under stress an
still arise as significant challenges for ADSs due to the intrinsic nonlinearity and fast-paced behavior of such events. These man
uch environments, it is crucial for the vehicle to predict its motion and adjust its controls proactively in response to environmenta
as micro- and nano-scale robots, soft robots, medical robots, and bio-hybrid robots, it is infeasible to anticipate the operating e
ency of estimated inertial parameters. One of the geometric approaches is to use a regulariser that represents how close the p
ot been considered in prior work. Confined workspaces and limited visibility complicate prior assumptions of multi-robot manipu
a cable-slung load poses significant challenges. On the other hand, purely data-driven learning methods do not comply by desi
d slipping. In this work, using Gaussian Process Regression (GPR) and Sigma-Point Transforms, we estimate the non-linear e
uitive interaction and increased acceptability. Human likeness in robot motions has been conventionally measured and realized
o and limbs, but performing massage motions through suitable actuation on a person’s head has been a challenge. In this
arios are long-tailed and hard to collect, which restricts the system from getting reliable predictions. In this paper, we build a new
arly crucial for human-robot collaboration in household settings, where data collection is often privacy-sensitive. In this paper, w
grasping quality and success rate. First, we designed a wearable tool-like parallel hand exoskeleton equipped with optical tacti
es at work, but hardly scale to a large population of workers because of limited consideration of morphology-specific behaviors
ems. However, current open-source models are restricted to a limited range of body parts and often with a reduced number of
ackbone for exhaustive underlying features and then directly estimate human mesh vertices. These factors lead to redundant pa
ven more comprehensive information than EMG to control these assistive rehabilitation devices. Measuring in vivo muscle force
inuum robots. Despite its merits, the development of a control-oriented dynamical model1 tailored for this specific class of robo
d in MATLAB and is used for understanding the underlying physics, for identifying the regions of operation, and for demonstratin
eir hysteresis characteristics. In response to these challenges, we propose an adaptive control method to compensate hysteres
tics assumes Piecewise Constant Curvature (PCC). However, PCC introduces errors when the robot is subject to external force
int configurations that satisfy desired position or pose queries, is a fundamental challenge in motion planning, control, and calib
cost of considerable control challenges. Data-driven model predictive control (MPC) is an approach that has shown promise in
camera images, and/or use distinguishable markers on a deformable body have limitations in realtime tracking of large continu
der approximate model describing the soft robot's dynamics is first derived with the finite element method and proper orthogon
or reasons including the hard-to-model dynamics, the system’s underactuation, and the need of using a prominent feedforw
ement following. As a potential alternative, fluidic elastomer actuators (FEAs) possess compliance and flexibility that is importa
ation as an essential step to be deployed. Available sensors are usually realized by introducing rigid bodies to the soft robot or
einforcement learning that generates, given the target position, an actuation pattern for the tossing task. To deal with the high n
Robot Inverse Kinematics (VR-Soft IK), overcomes fundamental limitations of general-purpose drag-and-drop soft robot contro
ose new modes of locomotion or grasping. However, a methodology for implementing motion planning of these tasks, as exists
on growing robot with an embedded steerable catheter. The current manuscript first models the robot using kinematics (consta
rees of freedom. Their continuously deforming bodies complicate integrating rigid sensors and reliably estimating its shape. In
metric mechanics (also known as gauge theory) that can be applied to systems with low-bandwidth control of the system's inter
e systems have been prevalent in the literature, there is an increasing trend toward designing soft-rigid hybrids with intrinsically
muscle tendons, enabling precise stimulation of hand specimens. Hand movements are monitored through an optical tracking s
e-bodied individuals. The data was then used to recover optimal cost functions in a predictive simulation environment with mus
uited to ascending rocky cliff faces, but most existing designs are either large and slow or limited to relatively flat surfaces such
ed robots prioritize foot design from a bionic perspective to accomplish various climbing tasks while overlooking payload capaci
xed-base frame during tasking due to forces and torques on the base that arise directly from arm motions (e.g. arm drag in low
a key component of geometric motion planning, it is almost infeasible to apply current geometric motion planning to high-dimens
ncorporate new terrain information perceived while the robot is moving within its environment. While hybrid mobility robots offe
these requirements. However, traditional methods of finding the control policy under those STL requirements are computationa
P) problem. Although there are no known polynomial-time solutions to TDSP in general, our recent work has identified a tractab
ns. Among others, using multiple distinct paths is necessary for optimization-based trajectory planners where found trajectories
handle rapidly shifting multimodal optimal action distributions. While MPPI can find a Gaussian-approximated optimal action dis
ells in a single pass with no preprocessing and independently of the number and shape of obstacles. It has a compute and mem
h planning due to their appealing properties. Nevertheless, their real-time applications face considerable challenges, primarily s
ionally efficient and high-quality solutions are lacking, especially in lifelong settings where robots must continuously take on new
. Ergodic search already derives continuous trajectories in these scenarios so that a robot spends more time in areas with high
n presented to improve the performance of the extension and convergence of the random trees, such as Informed RRT*. Howev
r, the study of path planning for CRs while considering external elastic contact is limited. The challenge lies in the fact that CRs
o the constraint of maintaining connectivity of the structure. We develop two reconfiguration methods, one based on spatio-temp
niformly over the free space, which leaves room for efficiency improvement. To accelerate convergence, rule-based informed a
s paper presents performance improvements of more than 500x over the state-of-the-art, bringing planning times into the range
cations. In the setting of persistent monitoring, we develop a method for representing environmental information and risk in a w
While much existing research predominantly addresses robotic motion planning with predefined objectives, emerging problems
angle constraints. A variant of tangent graphs is used as an underlying graph for searching a path using A* in order to reduce th
ized, discusses the properties of this representation that can be leveraged during the optimization process and shows that the
pulators, coupled with the challenges of obstacle avoidance and constraint satisfaction, often makes motion planning the bottlen
ains a novel RL policy using an actor-critic network and arbitrary data collected in real outdoor vegetation. Our policy uses heig
ptability, as they can be directly used off-the-shelf for any new scene without needing specific training datasets. However, witho
on a convex decomposition of free space to design collision-free trajectories using fast convex optimization. In this work, we pre
ch strategies for the pursuit team have shown to be either computationally intractable or permit poor solution quality. This pape
st. Besides,unnecessary heading changes, or zig-zag patterns, exist even when no obstacle is nearby, which is inconsistent wit
ility proof given long enough time. The algorithm trains a manifold using configuration space samples as data and triangulates
to construct a graph of free-space bubbles, effectively compressing the skeleton information in a sparse data structure but reta
pe of multi-goal mission often relies on humans designing a set of actions for the robot to follow in the form of a path or waypoin
y pipeline based on state of the art object detectors and OctoMap. First, a point cloud of surface obstacles is generated, which
 us studies neglect the configuration of the gates, simply rendering drone racing a waypoint-passing task. This formulation often
metric (or kinematic or input) constraints and specifies the homotopy class of the trajectory, and a back-end that optimizes this
anical simplicity and agility, but pose great challenges in planning and control. To realize fully autonomous aerial transportation
ates the quadrotor aggressively to intercept balls that are far away and hit them to distant positions in ways that are beyond the
namic model of the UAV with the suspended load and, hence, an estimate of the load state is obtained by fusing the measurem
etected object is dynamic, we propose to address the combined surface exploration and monitoring of the detected objects by m
provide an intuitive way of delivering high-level commands and allowing humans to perform collaborative tasks. The presented w
ecise motion planning and task allocation for a UAV, traversing through a series of planned space-filling paths involved in the ae
ce through the dual-motor anti-backlash control technology. Specifically, the first three axes of the D-Arm near the base are imp

. The uniform control implies that the passive FTC is able to cover the condition from quadcopter fault-free to rotor failure witho
n a singularity-prone pendulum model, we present a real-time slosh-free tracking technique. This method solely requires the refe
hicle can be especially costly. Scale vehicles have been developed in the literature to serve as a cost-effective transition from te
obots to avoid randomly moving obstacles when the probability distribution of the obstacle’s movement is partially observab
ing disturbances through high controller gains, it is an important factor when considering collaborative settings where smaller ro
d frame cable-driven parallel robots (CDPRs), accurate modeling of the deformable frame poses challenges due to errors and u
en forces on the manipulator or frequencies close to the natural frequency, and mechanical resonance can occur due to the ine
guarantee the safe coordination of the robots while preserving a desired trajectory via a layered controller. The top-level planner
ction in complex, multi-tiered environments. In this study, we consider the problem of operating a quadrupedal robot in distillatio
e control of RCM may be influenced by perturbations (noises) and that the end-effector is frequently replaced to complete differ
fficient model-free method for controller tuning and adaptation. Standard methods, however, are not enough for high-precision
ging due to the robots' large degrees of freedom. To address this challenge, we propose a phase synthesis (PS) scheme for ret
th controllers incorporate an unknown system dynamics estimator (USDE) to estimate disturbances without requiring accelerati
d positioning stages needs to be controlled and synchronized. Therefore, typically suitable model-based controllers of fine stage
osed so-called nonlinear robust internal-loop compensator (NRIC) framework in our previous work. In this paper, we further ext
ominance may reduce surgical safety. This paper developed a safety-enhanced human-robot shared control method by intellige
ntrol approaches are employed for these kinds of physical interaction scenarios. In this work, an extended Cartesian impedanc
n the face of fluctuating environmental conditions. This adaptation process must be performed in a stability-preserving fashion,
raditional robotic arms have limitations in fully leveraging their advantages in such scenarios. To address this demand, we have
y locomotion tasks. Model-free reinforcement learning, on the other hand, has gained much popularity in recent years due to co
constraint, we propose a derivative-based lifting function construction method, show that the corresponding infinite dimensional
ver, learning such a complex model that depends on process conditions is challenging because it requires a significant amount
ds typically rely on either good prior knowledge of the multi-contact model or require significant offline model tuning effort, thus r
or gathering empirical data required to train models. In this work, we address this issue by proposing Data-driven Robot Input V
k does not fully determine its joint maneuvers. Additional constraints or objectives are required to solve the underdetermined co
ion planning approach based on a new state space that is logically factored, integrating the capabilities of the robot through fac
timality. This paper introduces an RRT* approach (ORRT*) that leverages opportunism to (1) find solutions quickly, (2) reduce w
d to simulate global control and the resulting continuous movement of magnetic modular cube structures. We develop local plan
significant challenges. This paper imitates the winding uncoiling behavior of snakes to achieve controllable constrained path foll
mous driving, which has been validated in high-fidelity simulations and real-road demonstrations with practical constraints. The a
ts or obstacles elsewhere in the environment. In this situation, it becomes necessary to divide the trajectory into shorter segme
 t controller module. CBFs, along with a circulation inequality are used to generate safe paths toward the goal while preventing
enarios ranging from object manipulation to legged-robot locomotion. However, the zero-volume nature of manifolds makes the
osed to significant disturbances in non- convex free space. We propose a motion planning method based on the Explicit Refere
ability of an important component of autonomous robots to adversarial attacks - robot motion planning algorithms. We particula
ng-time horizon persistent coverage problem complex. Ergodic search methods consider the spatial distribution of environment
o facilitate sliding and rolling movements across critical regions, which represent semi-free configurations, where the robot and
ent-free (SEF) commands for a tethered differential-driven robot to track a path. Whilst existing path planner has been capable
se in real time. To address this issue, we propose a new generalizable control barrier function (CBF)- based steering controller
bility of the mobile base. However, the mobile base also brings redundancy to the system, which makes mobile manipulator mo
a recent learning-based model that reduces the search space for unconstraint planning for sampling-based motion planners. W
te an approximate motion which is then fed to numerical optimization for refinement. The sampling-based stage finds a long-ter
of single UGV is low due to its limited sensing range. To solve this problem, this paper proposes a risk-inspired aerial active exp
lect a robotic dataset of high-quality 4-channel audio paired with 360 degree RGB data of people moving in different indoor set
er experience. We solve this perception task with a sequence-to-sequence classifier of a potential user intention to interact, wh
rbal cues, specifically gaze and hand gestures, affect the persuasiveness of a social robot. We have designed a protocol to incl
to the robots’ limited sensing capabilities. Previous approaches in the literature typically simplify the handover scenarios, i
a prominent gap persists: in Human-Robot Interaction (HRI), the procedure of communicating commands to robots demands in
n this paper, we propose a dual-modal electronic skin (e-skin) that integrates magnetic tactile sensing and vibration feedback f

hes that employ probabilistic models and token-pruned techniques have shown decent results in multi-view human body recons
quences of 3D skeleton coordinates as inputs and apply Graph Convolutional Networks (GCNs) for spatiotemporal feature lear
s alone cannot determine the context of the human contact---whether it is a deliberate interaction or an unintended collision that
put by active information gathering. Previous approaches often rely on predefined templates to ask disambiguation questions, r
structions to find generic object classes, neglecting the utilization of natural language interaction and the complexities of identify
specifies the target object's category (e.g., bottle). Inspired by pragmatics, where humans often convey their intentions by relyi
ary measurements remains challenging due to EIT’s inherent location dependencies and image artifacts. This study presen
e this problem, we first present an uncertainty estimation method for LLMs to classify whether the command is certain (i.e., cle
ocial settings. In this paper, we developed a system with desktop robot Haru to assist group discussions. The system consists o
tems, active exoskeletons present a more versatile control system because it is possible to select different assistive strategies
le ring-shaped controller equipped with an inertial measurement unit (IMU) and a light-emitting diode (LED). This novel controll
rcrafts within the IAP are interconnected, presenting promising opportunities for enhancing localization. Incorporating the physi
It also makes traditional Appearance-based localization methods unusable. Recently, researchers have utilized the object's se
rs, underground or underwater. In this paper, we introduce Swarm-SLAM, an open-source C-SLAM system that is designed to
onstraints of a single UAV, the efficiency can be greatly improved through the collaboration of multi-UAVs. Nevertheless, when
lenge. In this work, we propose a centralized system designed to facilitate collaborative LiDAR-ranging-inertial state estimation
ot applications. In order to obtain the relative locations between robots, no matter whether they are in relative motion or stationa
egrating multiple LiDARs can be challenging, attributable to temporal and spatial discrepancies. Common practice is to transfo
hes mainly rely on expensive mapping devices, which are time-consuming for city-scale map construction and vulnerable to err
the cable-towed payload. To this end, we first propose novel dynamic models for both the quadrotor and the quadruped robot,
d by individual robots, it is necessary to form subteams that can collaboratively finish the tasks. It is also advantageous for robo
a collaborative and efficient way is a challenging problem that requires online coordination and information fusion. The majority
 tem size scalability are known, for example, from swarm robotics. Example strategies are self-organizing behavior, a strict dece
e that agents have timely and accurate solutions to their state as well as the states of the other robots in the team. To optimize
the diversity optimization algorithm uses the multi-armed bandit algorithm to select the coefficient that predefined space. Howe
al features, as this can lead to deteriorated localization or complete loss of tracking. To achieve this, we propose a hierarchical
Robots are equipped with LLMs to discuss and collectively reason task strategies. They generate sub-task plans and task spac
sed methods have showcased the ability to deploy learned controllers onto single quadrotors or quadrotor teams maneuvering i
eed to communicate and collaborate with other robots in real time, and reach agreed cognition for task assignment, which puts
arge number of optimizations with respect to the number of agents. One-agent-at-a-time algorithms offer computationally efficie
nder the pressure, it could also be a drive for us to stimulate our potentials to the best in order to keep up with others. Inspired
solution to be used in various real situations where radio communication is not available, such as underwater or in a lot of radio
ploration, real-world implementation is limited due to agent coordination complexities caused by communication and sensor lim
pt accordingly. We propose robot reputation to assess the historical performance of agents and make future adaptations in a p
Robot Web solution is based on Gaussian Belief Propagation on the fundamental non-linear factor graph describing the probab
immediate neighbors is insufficient to imitate a centralized control policy. To address this limitation, prior studies proposed inco
d and rarely addressed, although it significantly influences the accuracy of estimation. In this paper, we introduce time synchron
must be robust to unreliable and intermittent communication links. In this work, we present MOCHA (Multi-robot Opportunistic C
and the UGV have predefined areas that they have to cover. The goal is to perform complete coverage by both robots while mi
clay bricks. But, despite the wide variety of space-filling polyhedra known to mathematics, only the cube has been explored in r
nt learning. A cascade controller is formulated, comprising a position control component to ensure containment achievement an
bots have become a significant research interest in recent years. Making effective use of the robots and their unique skills in a t
l (MPC) for each robot and a social long short-term memory model that forecasts pedestrians' trajectories. We formulate the lo
ific agents. These tasks are then imposed through control barrier function (CBF) constraints to ensure completion. To extend e
an existing methods and successfully plans for problems with up to twenty objects, more than three times as many objects as c
ection operations in which safety including intra-agent or agent-target collisions are a serious concern. Additionally, in an on-or
sensor observations based on information from a lookahead sensor. However, current implementations have failed to account
challenges. The core of RecNet's methodology involves a transformative process: it projects 3D point clouds into range images
arts of the same environment. The navigation task is represented as a sequence of semantically labeled areas in the map, that
ot systems (MRS). Given that most of existing works mainly focus on surface inspection of single and fully known 3D structures
s. The framework utilizes spatiotemporal occupancy grid maps (SOGM), which forecast the occupancy status of neighboring sp
sages” enables improved coordination behaviour and communication efficiency in difficult scenarios, such as those where t
narios such as surveillance, herding, and racing can be modeled using game-theoretic frameworks in the belief space. Stochast
s: the number of targets and/or their initial locations may be assumed known, or agents may be pre-assigned to monitor disjoint
ual landmarks are used to quantify the combined impact of process noise and sensor noise on map uncertainty. Additionally, we
d constraints to comply with, computing optimal plans can become especially hard. Moreover, we increase the problem complex
ion-based methods, and bipartite graph matching methods. These methods often assume a form that lends better explainability
arly in the Pilbara iron-ore region of Western Australia. The goal is to paint the technology landscape and highlight issues relev
use. The proposed methodology is based on a three-layer control architecture, which is described as follows: 1) the Top Layer
en approach based on hedonic game to address the task allocation problem of muti-robot systems in emergency rescue scenar
nge and Field of View (FoV) providing bearing measurements and it implements an Information Consensus Filter (ICF) to estim
e from embedded cameras) are an attractive choice for use in decentralized formation control, however this requires that the fo
. Prevailing MuRP solutions for uniform node coverage either incur high (non-polynomial) computational complexity operations
s while maintaining their constant but different speeds. Instead of controlling the spatial relations between the robots, our format
ce inspection task. We task a group of robots to inspect and collectively classify a 2D surface section based on a binary pattern
emove one block at a time. When moving, robots are also allowed to climb or descend a block. A construction plan may thus bu
guration space challenges the real-time performance of existing grasp planning methods and drastically limits the application of
in robot grasping, dexterous manipulation, and bimanual manipulation. The platform combines adaptive and hybrid actuation sy
uild robotic hands equipped with soft tactile sensors, they would be highly accessible and facilitate many robotics projects. To th
been well studied, the study on natural grasp motion planning with a dexterous hand remains a very challenging problem due to
n learning. The existing approaches for acquiring expert data commonly involve utilizing a data glove to capture hand motion in
t, especially holding hands, on a slender finger. In addition, in order to unitize the hand for easy use with any robot and maintain
e dexterity of human hands when manipulating a wide range of objects. While the control of humanoid hands evolved greatly, h
given two types of datasets for each task: expert and mixed. Each expert dataset is collected by a high-skill policy, whereas th
d positioning of the part. In this research, we propose a novel robotic hand with these functions and corresponding finger mecha
mendous development of modern marine industry. To this end, we propose a coordinated multiple UAV-USV landing control al
proach. The proposed framework is primarily motivated by the distributed implementation of Control Barrier Functions (CBFs), w
-angle leader-follower formation is a problem of maintaining the desired relative distance and orientation of robots in a group. O
oal coordination and collaborative mapping are often solved separately, when clearly each of these should influence the others
rvice (QoS) features. Despite the prevalence of 5G QoS discussions in the related literature, there is a notable absence of real
k cell is fully enclosed and features individual lighting, a low-cost Cartesian robot arm with an attached rotatable parallel jaw gri
en toolkit that takes ROS2 nodes as input and automatically runs relevant benchmarks to quickly return a menu of cloud compu
equirements of a different team of task-oriented robots seeking to coordinate their actions in the absence of existing communica
ability of the networked system is guaranteed by ensuring passivity of each subcomponent in the interconnection, as well as the
ed to operate in real-time and eliminates the need for an initial pose guess that is proximate to the sought solution. IFFNeRF ut
ment of low-cost LiDARs (e.g. Livox LiDAR) enables us to explore such 4D perception systems with a lower budget and higher
propose a novel Multi-Task Radar-based Single Stage Detector, termed MTRadSSD, where we adopt instance-aware sampling
per aims to address such a fundamental problem of camera-based 3D object detection: How to effectively learn depth informatio
ow. For the robust performance of object detectors, unsupervised domain adaptation has been utilized to adapt the detection ne
d suffer from occlusions and ambiguities due to object symmetries. We overcome this issue by presenting a novel symmetry-aw
as seen significant advancements; however, the majority of current works focus only on accuracy and ignore uncertainty quanti
methods provide an alternative to this problem yet suffer from inherent scale ambiguity stemming from monocular observation
nformation introduces ambiguity in the translation and size of objects, leading to suboptimal performance. In this paper, we prop
ned categories. This is because object-detection methods are trained to assign a background label to pixels corresponding to th
duces an information bottleneck that limits 3D object detection accuracy and scalability. To address this limitation, we propose
. However, RGB detectors still outperform event-based detectors due to the sparsity of the event data and missing visual detail
een great success in 3D point cloud object detection. However, the lack of hierarchy in a plain transformer restrains its ability to
paper, we propose a detection enhancement method, namely FrameFusion, which improves 3D object detection results by fus
d into three types based on input representation: point-based, voxel-based, and pillar-based. Among these, pillar-based models
ore further analysis. In this paper, we propose a novel video semantic segmentation paradigm for compressed video. Specifical
es that rely on a progressive downsampling of points. While this method effectively reduces computational demands and increa
optic have rich relations, however, which are rarely explored. In this work, we propose a novel panoptic, instance, and semantic
n LiDAR and 4D radar. We introduce multiple alignments on both spatial and feature levels to achieve simultaneous backbone r
D based on self-training have been effective in overcoming changes in the general appearance of images. However, shifts in a
orized into LSS-based and Transformer-based in terms of their view transformation. The former leverages inaccurate depth pre
of input video streams. Recent work has focused on spatially aligning BEV-based features over timesteps. However, this is oft
attention blocks or changing the whole structure of feature fusion networks. However, there is still space of improvement and th
to match natural human communication and convey rich semantics in manipulation tasks as effectively as natural language ex
scue operations. However, drone-to-drone detection presents a more intricate set of challenges compared to regular object dete
d adoption in affordable vehicles. Camera-based detectors are cheaper alternatives but often suffer inferior performance compa
improve out-of-distribution (OOD) detection for segmentation, we introduce a metacognitive approach in the form of a lightweig
evious works perform unseen object instance segmentation (UOIS) by training deep neural networks on large-scale data to lea
g trajectories with other agents and en- suring collision-free paths in a fully decentralized setting is complicated by dynamic elem
ge of potential affordances on real-world objects, and the lack of open-vocabulary support for affordance understanding. In this
often both burdensome and time-consuming. To tackle this issue, the literature has seen the emergence of several weakly sup
cope of applicability and are limited by their efficiency or accuracy. In this paper, we propose HandyPriors, a unified and genera
d using lidar measurements. However, with improvements in radar sensing, resulting in better accuracy and higher resolution, ra
LiDAR is one of the most important sensors for this task. Unfortunately, the existing approaches perform unsatisfactorily due to
robust moving object detection in complex unstructured environments. The central idea of our approach is to incrementally est
nt cloud data, leaving their performance on commonly available visual-inertial simultaneous localization and mapping (VI-SLAM
oints and suffer from distortion of projected RGB images due to the sparse nature of 3D point clouds. To alleviate these problem
sion avoidance. We investigate this estimation task for vehicles using radar sensing. We address moving instance tracking in s
scene exhibits a part-whole hierarchy between the within objects and the scene itself, simple parts (e.g., roads, lane lines, veh

unlabelled datasets through a Masked Image Modeling (MIM) approach, which is robust to augmentation hyper-parameters and
task has important applications in the robot area but the previous DPS models are too heavy to be applied. Thus, we propose E
ctical applications often encounter information disturbance which includes perception packet loss and time delays, and a compr
sets of points belonging to different convex parts of an underlying shape represented as a point cloud. After analyzing its prope
is scenario necessitates the operation of MOT at a low-frame rate. However, algorithms within the MOT research field have be
between the UDA approaches and the fully-supervised approach is still noticeable, which is impractical for real-world applicatio
ata and accurately calibrated sensors which is often not the case in real-world scenarios. Data from LiDAR and cameras often
nventional approaches struggle to adapt to dynamic environments. This paper presents TerrainSense, an end-to-end framewor
ct detection. However, previous radarcamera fusion models could not fully utilize the potential of radar information. In this paper
However, obtaining dense pixel-level annotations for model training is a laborious and expensive process. Active learning hold
se two modalities, we design an early fusion module for joint voxel feature encoding, and a middle fusion module to adaptively f
n iteratively and intelligently label a dataset while retaining high performance and a low budget. In this work we explore AL for Li
model from these samples. We identify two problems with existing active learning methods for LiDAR semantic segmentation.
opose a trajectory-prediction-based dynamic tracking scheme, which includes target localization, trajectory prediction, and UGV
le steering mechanisms with distinct kinematic properties. For example, parallel positive steering (PPS) with four wheels in para
aper, we present a novel approach to jerk-constrained time-optimal trajectory planning (TOTP), which follows a specified path w
erately-large and sparse linear systems. Iterative methods are particularly well-suited for parallel solves of such systems. Howe
this approach is computationally bounded, implementations of direct trajectory optimization that use iterative methods to solve
ing time-optimal trajectories based on the full quadrotor dynamics (typically on the order of minutes or even hours) can hinder i
escriptor exists for representing force trajectories, which appear in contact tasks. This paper introduces invariant descriptors for
und, then a retiming algorithm computes a speed profile that satisfies kino-dynamic and other constraints. While trajectory retim
 e, a high number of degrees of freedom is necessary, as is the case for humanoids and collaborative mobile manipulators. From
c uncertainty, space and time correlations, nonlinear dynamics, and non-convex constraints. In this work, we first introduce a co
bots are capable of highly dynamic motions even with limited motor power. However, the time-dependency of this energy storag
DP) has emerged as a promising tool for achieving highly dynamic tasks. But most of the literature deals with applying DDP to a
k, we propose a novel model-predictive control approach that incorporates direct feedback from force sensors while circumvent
ng the various frameworks of screw theory, Lie algebra and dual quaternions. A unification and generalization of these popular
uncertain system into a higher dimensional deterministic system that accounts for worst case disturbances. Our main insight is t
at satisfy the CBF constraints for high relative degree systems with input constraints. To address these challenges, recent work
these three properties as constraints in an optimization-based controller using control Lypaunov and (integral) control barrier fu
dynamic controllers with bounded suboptimal performance for various underactuated robotic systems by finding good approxim
e made an explicit contact feedback loop in the MPC framework. An algorithm called Multi-Contact Particle Filter with Exploratio
solid-, fluid-, and gas-like behaviors) and high-dimensional (as they could have many grains/particles with at least 3 DOF in 3D)
motivations behind this algorithm are the two main limitations of TO and RL when applied to continuous nonlinear systems to m
operational domains. We start from existing LMPC formulations and modify the system dynamics learning method. In particula
a stochastic model.We focus on safety constraints that take the form of time-varying convex constraints such as collision-avoida
norm-based stability criterion for a biped walking constructed on its linear inverted pendulum model (LIPM). More precisely, an
nnected dynamic system, unknown bounded disturbances and computational errors are taken into account throughout the lear
s particularly advantageous for underactuated systems, which are systems that have fewer actuators than degrees of freedom
ng the time-averaged statistics of a robot trajectory with the spatial statistics of the target distribution. Most existing ergodic con
ormulating the Alternating Direction Method of Multipliers (ADMM) algorithm for solving QPs as a deep, weight-tied neural netw
different vendors to co-exist in the same space is crucial to prevent vendor lock-in. We present the first decentralized system th
the leader lacks complete knowledge of the follower's decision-making model. There is a need for learning-based methods to e
epair impeding vehicular movement in the road network. A primary vehicle, which we refer to as a convoy, wishes to travel from
decomposition of an information map can be used to guide a set of heterogeneous agents, each with different sensor and moti
ent robot social navigation techniques leverage advances in reinforcement learning and deep learning, they frequently overlook
elligent transportation across decades. Simultaneously, recent parallel works in robotics have devised efficient algorithms for mu
ssist another robot in reaching a target in an unknown environment. We propose a framework that sequentially selects the opti
ng planning. Since the wind field effect is prominent when drones have to fly close to each other, we cannot ignore the effect du
robots to gather the most information about the environment. We propose an efficient sparse Gaussian process-based approa
ious multi-robot planning and coordination algorithms have been proposed, including Multi-Agent Path Finding (MAPF) method
ues with search-based methods, focusing on enhancing the conflict-based search (CBS). The CBS ensures optimality but suffe
omposition of the environment's free space into cuboids, convex polyhedra, or spheres. However, such SFCs can be overly con
a modified high-level conflict tree to efficiently resolve robot-robot conflicts in the continuous space, while reasoning about each
s. We solve this problem by combining Conflict-Based Search (CBS), a multi-agent path finding method, and discontinuity-boun
m found at the core of many logistics, transportation, and general robotics applications. Existing learning-based MAPF approach
, and on-demand. The recently proposed horizon-based synchronous algorithms compute the paths for all the robots in each h
s exacerbated in environments with narrow passages that robots must pass through, like warehouse aisles where coordination
mum travel time of all robots. Specifically, we focus on a reduction from the MCPP problem to the Rooted Min-Max Tree Cover
tory-tracking tasks with UAVs. The network is trained on simulated depth images to output a collision score for queried 3D poin
n of the cable's state during the transportation of the load. Based on the advantageous properties of event cameras, our ultima
tion system that enables it to locate powerlines and approach them for landing. A passively actuated gripping mechanism grasp
 hes for feasible trajectory using a set of motion primitives in state lattice and represents the optimal one as a polynomial by sol
wo main limitations to previous approaches: they may not be able to provide a complete map of the environment and assume th
s mainly caused by a lack of training data for such diverse viewpoints and scales. To address this limitation, we propose the Si
V) LiDAR. The point cloud acquired from the low-cost LiDAR is firstly converted into a 3-channel tensor via motion compensation
embedded computer and above 200FPS on desktop- grade 200W GPUs, even in stereo configuration and in the multiscale set
arious collision avoidance methods have been developed, it is still possible for drones to collide with unobserved obstacles due
his challenge, we present a safety-aware approach for online determination of the optimal sensor-pointing direction, psi_d, whi
truction while simultaneously compressing spatial and intensity point cloud data. The strategy employed in this work utilizes Ga
ce limits on mission time requiring efficiency in exploration. We present a novel exploration framework that uses Deep Learning
ath quality in large and complex scenes. To address these challenges, we propose FC-Planner,a skeleton-guided planning fram
sed exploration approaches, we aim to enable robots to methodically explore and observe an unknown set of rooms in a structu
tions, promising to cover large areas at high speeds. Moreover, with the increasing maturity of onboard UAV perception, resear
The proposed solution builds upon a deep collision encoder that is trained on both simulated and real depth images using supe
accurate knowledge of the environment, including the gap's pose and size. In contrast, we integrate onboard sensing and detec
n with Visual-Inertial Simultaneous Localization and Mapping (VI-SLAM) and dense 3D occupancy mapping to guarantee safe
D flow sensors to measure flow from a single source like wind or the drone’s relative motion. Our current work advances flo
field have investigated the 3D object detection problem, the usage of dense point clouds from a heavy Light Detection and Ra
cing estimation accuracy, has often been overlooked. In this paper, we present APACE, an Agile and Perception-Aware trajeCt
ng a farm around a heterogeneous robotic system dubbed SpECULARIA. Within this multi-robot system, mobile robots are dep
novel end-to-end deep learning approach for the simultaneous detection of multiple thermal updrafts and the estimation of their
acy. This FEM with fewer convolutional kernel channels can reduce the parameters, speed up the inference time, and mitigate t
lti-robot systems, and they struggle when image features are weak, ambiguous, or evolving. In response, we propose Natural Q
ds derived from depth image often suffer from reduced scene understanding due to occlusion, ultimately impeding their graspin
on and pressure across complex, non-developable surfaces are essential for empowering next-generation robots with tactile aw
and mobile robotic systems. Among the most prevalent end-effectors, used for the majority of these applications, are the suctio
ain gaps in RGB and depth data and insufficient multi-modal feature alignment. First, a self-supervised rotation pre-training stra
econstructing complete depth maps for challenging transparent objects, leaving holes in the depth reconstruction. Recent work
work that uses monocular image sequences to reconstruct the wiry structures whose skeletons are mainly straight as commonly
auce plating robot, which requires precise control of the thickness of squeezed liquids on a surface. Different liquids demand dif
hand-object scene reconstruction is challenging due to the depth ambiguity of a single image and occlusions by the hand and o
harvested trusses. We propose a method to grasp trusses that are stacked in a crate with considerable clutter, which is how th

ate drastically under low-light conditions. Although supplementary information can be provided by additional sensors, such as d
ctile information. In this paper, we focus on utilizing low-resolution (LR) tactile sensors to reconstruct the localized, dense, and
eep learning methods can provide powerful vision-based priors on the affordance of articulated objects from previous, possibly
since it requires knowledge of the friction coefficient, which can vary from object to object and even from grasp to grasp. Tactile
s on a set level as opposed to a point level. This contrasts methods that predict cross-pose from per-point correspondences, w
neural networks, which poses challenges due to the difficulty in obtaining sufficient pose annotations. To address this limitation,
aints on the DLO's movement through contacts and fixtures. We propose a strategy to estimate the pose of unknown environme
emporal domains. Few existing methods cover all these properties for robot grasping. In this paper, we propose AnyGrasp for g
ich is capable of autonomous harvesting based on crop phenotyping and pose estimation. Specifically, In phenotyping, the dete
s optimised for a specific end-effector, that generate grasp poses directly from sensor input. In the domain of assistive robotics,
ue to the limitations imposed by opaque containers, transparent liquids, fine-grained splashes, and visual obstructions arising f
rom highly cluttered scenes. Real-world learning is a promising approach, but executing picks in the real world is time-consumi
tion with direct touch sensing is not feasible. Adding tactile sensors can complicate the hardware and introduce extra costs to th
art efficient radiance field re-construction method is suitable for highly accurate and robust pose estimation for objects with know
om the goal scene. The key to an image-goal rearrangement system is estimating the desired placement pose of each object b
s to leverage the low-latency and continuous (i.e. not fixed-rate) data capture for high-frequency tracking. We propose the EDO
underlying environment dynamics makes it an interesting research field for statistical learning approaches and neural policies. I
account the linear velocity of the state estimate, improving over current pose-based trajectory analysis. This has the benefit of p
development research in offline reinforcement learning still relies on game tasks with synthetic datasets. To address such limita
ring robotic grasping and manipulation systems among different research groups, also sharing with the community the best pra
ly used dataset in the robotics community, to ensure that our results are comparable to other studies. Additionally, the benchm
, such as cruising, and distance-based metrics that are computed by averaging over all scenarios. Following such a regiment p
cause otherwise reasonable predictions that suffer from such longitudinal localization errors to be treated as false positives. We
tool, the first programmed under the ROS 2 framework, can be used together with different well-known robotics simulators like
baseline. The suite encompasses ROS 2 packages covering the full robotics pipeline and integrates two distinct benchmarking
significance of these properties is evident in the realm of cloth manipulation, accurately categorising and comprehending them
radigm of passivity-based control, virtual energy tanks allow to perform such adaptations in a robustly stable way, by bounding
and social human robot interaction in particular has received special attention as a feedback mechanism for human performanc
when attempting to modify or adjust the loads within a single training set. This paper introduces an innovative method for achie
odel that uniquely casts the user’s sequential behavior as well as a robot support intervention into a hierarchical multi-obje
such robots in physical human-robot-environment interaction (pHREI) is crucial. Having a wide range of Z-width, which indicate
€™s opening around a part of the human body. The rich contact environment and the deformations of the garment make the ta
compute performance validation and sustain engagement to activate them from sedentary lifestyles. As a result, an at-home-ba
stood. Early research in robot-based gait rehabilitation offered little improvement over current standards of physical care and em
ut under scrutiny due to inherent challenges such as response bias, between-subject variations, and the granularity nature. Indi
yer Perceptron called Transducer and a Transformer. The former estimates, based on the kinematic data from a motion captur
ans in simulated environments, where the policies can later be transferred to the real world. However, a major bottleneck is the
specified, and who should lead the motion. This work is motivated by the ability of humans to communicate the desired destinat
an movement simulation, and the robotics and control theory were used to contextualize human sensorimotor control. There are
rce in a Human-Robot collaborative object transportation task at a middle distance. We outperform our previous predictor by ac
e control performance to comply with human intentions. However, a problem that generally exists with mobile pHRC but often g
uch as human soft tissue or musculoskeletal joint damping, other sources such as soft padding materials and bias strap forces
ort to users' stability or too bulky and poorly maneuverable to be used outside hospital environments. In addition, they all lack ad
b-maps built by the agents. Most of the current exploration strategies directly merge two sub-maps built by different agents whe
PC) that uses i) trajectory predictions of the dynamic environment, and ii) prediction regions quantifying the uncertainty of the pr
The safe corridor is constructed based on the perceived map, however, the non-ideal perception induces uncertainty, which is
ols on-the-fly. In practice, however, MPPI encounters problems limiting its application. For instance, it has been observed that M
n planners and prediction modules to predict multiple possible obstacle trajectories and generate obstacle-aware ego robot pla
. Moreover, existing online Partially Observable Markov Decision Processes (POMDPs) solvers do not support hybrid beliefs d
traditional POMDP models rely on the assumption of complete knowledge of the observation source, known as fully observabl
 hms such as Partially Observable Monte-Carlo Planning (POMCP) can solve very large POMDPs with the goal of maximizing th
ble planners to efficiently find paths with expected low costs through uncertain environments. The roadmap must be sparse so t
ainty does not scale to high-dimensional robots such as manipulators, assumes simplified geometry of the robot or environmen
tate and transition uncertainty. Unfortunately, online CPOMDP planning is extremely difficult in large or continuous problem dom
outside scene content. This limitation hinders their extensive applications in robotics, where the reliability of model predictions h
able to cope with mismatches between the expected and the actual behavior of the system (e.g., the presence of an unexpecte
d. While 3D exploration planning has been extensively studied, the environments are assumed static or only reactive collision a
constraints. Our approach is designed for global planning using limited onboard robot perception in mapless environments whi
into a deterministic motion planning problem and a graph search problem. First, we solve the deterministic planning problem u
jectories in crowded areas, it remains unclear how one can safely incorporate these learned models into a control loop due to th
vel of robust autonomy, these robots should be able to proactively predict and recover from such failures. To this end, we prop
ng noisy local observations. An effective approach for planning under sensing uncertainty is deter- minization, where one conve
ne understanding using measurements from vision-based neural networks. Due to appearance-based reasoning, the measurem
ficantly impact the system's optimal decision making strategy for mission completion. To model such environments, our work co
his specialization increases localization accuracy in domains where the given modality detects many high-quality features, but c
t search and tracking (SAT) in cluttered environments with limited sensing field of view. We develop a novel sigma point-based
an aid in accurately localizing the object of interest. Motivated by this, in this work we model high-precision insertion tasks as pl
ions and Gaussian uncertainty, which can be inaccurate or result in trajectories with excessive cost. While visual perception ca
nodynamic conflict-based search (K-CBS) in conjunction with the efficiency of Gaussian belief trees as used in the Belief-A fram
ersability index with predetermined threshold values, the traversability index cannot be precisely predicted because of the meas
ollision-free trajectories, their solutions often suffer from suboptimal time efficiency and potential unsafety, particularly when acc
ng certain risky edges in a centralized manner. While classical approaches can solve this non-standard multi-agent path plannin
ons, while multiple cranes share a track, necessitating collision avoidance. Our goal is to minimize the average task waiting time
nning and scheduling tasks, and is gaining increasing attention and focus in recent years. There have existed algorithms that ex
. In this study, we formulate and delve into the largest well-connected set (LWCS) problem and explore its applications in layou
coalition over a large domain, to complete spatially distributed tasks as quickly as possible. We focus on scenarios involving ho
ocation due to factors like time-extended tasks, and agent specialization, and 2) uncertainties in task execution. Potential task
ocal congestion and motion conflict may greatly degrade system performance, especially in the presence of uncertainties. Our
mission completion. It also enhances resilience by enabling task transfer in case of a robot failure. This paper presents a formula
Storing and Releasing (ESR) feet have shown deficiency in push-off torque, which results in asymmetric gait pattern, slower wa
into the bones tracked by the optical tracking system. This is normally done by invasive, radiative manners (implantable marke
ol of upper extremity exoskeletons based on user intention and task-specific movements, the prediction of body joint positions b
hance the detection sensitivity of phase transition, this work utilises the LSTM-CRF hybrid model. Four IMU sensors attached to
ature parameters that smoothly vary over gait phase and task according to a data-driven model. However, these recent efforts o
to restore ankle plantarflexion- dorsiflexion. The prosthesis is designed in the form of a spring-loaded three-loop linkage for fun
ed, for example, as promising tools in industrial ergonomics, the effect of using them on human motor control, particularly on int
tion intention and synaptic inputs to motor neurons also remains a significant challenge. The neural decoding methods of surfa
lems with accommodating asynchronous hip movements, achieving precise force control and efficient force transmission, and g
ange of motion (ROM) and load-bearing capacity comparable to the human wrist joint, we designed an elliptic rolling joint as an
e to limit the exoskeleton's weight and mechanical complexity, thereby compromising comfort and kinematic compatibility with th
ward shift of the center of mass (CoM) in the sagittal plane. This paper identifies the limitations of existing design approaches an
s walking, stair ascending and descending, and sitting, is crucial for ensuring adaptive control and enhancing user experience. W
omnidirectional recovery step planning by online optimization of step duration and position in response to external forces. We m
ied torque profiles to the user as well as the locomotion condition; however, it is currently unclear how sensitive participants are
these devices in coordination with their users is still an open research topic. While powered devices can replicate normative join
ng interest in utilizing exosuits in industrial environments to prevent work-related musculoskeletal disorders (WMSDs) by ensuri
heavy considering the limited resources a mobile robot has. We propose a new geometry-informed way of distance candidates
nd, there is a large variety of models for panoramic cameras. Many camera models have been proposed for fisheye cameras, c
ined platforms commonly used in unmanned aerial and ground vehicles (UAVs and UGVs, respectively); such point clouds can
mented scans into semantic objects and representing them only with their respective centroid and semantic class. In this way, e
atures for frame-to-frame matching. The accuracy and sensitivity of existing degeneracy detection methods need to be further im
ever, static tree data structures are inadequate to handle large and dynamically growing maps in real-time. To address this issu
ht detection and ranging sensors (lidars) still suffer from bias related to local surface properties such as measuring beam-to-sur
s from a point cloud usually relies on an accurate removal of ground points. The correct estimation of the surrounding terrain is
3D reconstruction approaches, either trained on stereo setting or structure-from-motion, estimate 2.5D visible models which ge
undreds of hertz with minimal resource usage. Our approach combines wavemap, a hierarchical volumetric map representation
d mines. These environments exhibit inconsistent lighting conditions, locally self-similar textures, and general poor visibility con
(SVGD). MPPI formulates a non-linear optimization problem that determines an optimal control input sequence and solves it u
not account for unwieldy tools or incorporate systematic methods to handle these extra maneuvers needed. Instead, we formali
movements, while a feedback control law is formulated to address any deviations the robot may encounter. In particular, our pr
nge of human-computer interactions, including medical scene monitoring, anatomical target planning, and robot manipulation. H
policies from visual inputs alone Visual imitation learning provides efficient and intuitive solutions for robotic systems to acquire
l pose hard challenges for state-of-the-art motion planners. In this paper, we present a new iteration of a reactive circular fields
on local ranging observation (pointcloud) to learn the terrain elevation profile and extract the feasible navigation subgoals arou
Developing exoskeletons should incorporate specific optimization algorithms to find the best design. This study investigates the
ns hold the potential to facilitate continued rehabilitation at home, but understanding what aspects of the design are most releva
s conventional definition, mechanical transparency was not considered when an interaction force was required during assistanc
compliant behavior is imperative for these tasks. Typical approaches involve force control schemes like admittance controllers t
al functions, each representing one of the four primary phases of gait. These polynomials are computed according to different s
bitrary arm movements. Reinforcement learning (RL) emerges as a promising method for controlling arm movement with succe
ve hybrid controller poses significant challenges. In this over-actuated system, it is extremely difficult to find the right balance be
o self-guided PT is low. Robots have the potential to support physical therapists and increase adherence to self-guided PT, bu
or agents to learn near-optimal policies, these methods encounter limitations when dealing with diverse constraints in complex
to safety constraints. This work introduces a novel approach that combines RL with trajectory optimization to manage this trade
ncing risk-aware decision-making. Previous DRL implementations often inappropriately treat statistical estimations as concrete
current methods make strong assumptions about other agents and often rely on manual tuning to balance safety, feasibility, and
ctive in leveraging past experiences for policy updates, they often overlook the potential of reusing past experiences for data c
d error. In this paper, a novel model-free Safe RL algorithm, formulated based on the multi-objective policy optimization framew
-human level video game agents to dexterous, physically intelligent robots. However, training these perceptive DRL-enabled sy
ng negative task interference still remains a challenge. Addressing this challenge by successfully sharing information across tas
ed semantic information from a single image makes current monocular detection methods unable to deal with complex scenario
ible, and as a result, the total area of the display in a vehicle is increasing. This is leading to increased demand for augmented
ptimize performance. Many robotic control problems deal with complex unstable dynamics, where applying actions that are off t
markable advances in computer vision and large language models. Privacy emerges as a pivotal concern within the realm of em
ng and the lack of diversity in Multi-Agent Reinforcement Learning (MARL) algorithms across different scenarios, restraining the
on the quality of the datasets, which may cause extrapolation error in the learning process. In many robotic applications, an ina
hts. These disturbances, typically due to wind turbulence, electromagnetic interference, temperature extremes and many other
mputing offers an energy-efficient alternative but has been underexplored in robotics. We introduce ReactHD, an HDC-based fr
odulus and bending stiffness. Such diversity poses challenges for developing generalized manipulation policies. However, prev
nt detector can accurately map environmental states to symbolic events; however, uncertainty is inevitable for real-world event
Drawing inspiration from cognitive science studies that emphasize the significance of multi-sensory integration in human percep
nal methods exist to solve object retrieval tasks, but they require overhead cameras for perception and a time-consuming exhau
nate (NDC) parameterization is widely used in NeRF learning for its simple definition, easy calculation, and powerful ability to re
ction and sampling-based motion planning. A recent line of research focuses on utilizing neural signed distance functions of eit
ch as point clouds, neural fields provide a continuous scene representation able to represent 3D geometry and appearance in a
deep neural networks, is a promising approach to address this issue. Existing works usually take the common aid from auto-en
ved robot learning, we propose POCR, a new framework for building pre-trained object-centric representations for robotic contr
s. However, both the objects and supporting items can exhibit substantial variations in their shapes and structures, bringing two
e frame, an ideal task representation would also exhibit this property. Nevertheless, most robotic learning approaches use unp
his paper presents a novel framework with two new architectures: SlotTransport for discovering object representations from RG
PVRs, each trained for five distinct manipulation or indoor navigation tasks. We performed this evaluation using three different
so combine other sensors (inertial sensing, for example) to perform combined state estimation. Our work rethinks this approach
clines in capability even for minor differences. In this work, we examine robustness to a suite of these types of visual changes fo
o be updatable incrementally in an online fashion. We introduce HIO-SDF, a new method that represents the environment as a
r utility has been limited, leading to a recent wave of efforts to pre-train robotics-specific image encoders that are better suited

c perceptive locomotion. While ex- isting approaches have shown promising results in effectively extracting planar regions from
tonomous navigation, and robotics, only a single view of the scene may be available, making the scene reconstruction task cha
ncy to occlusion, this paper proposes a bidirectional correspondence prediction network with point-wise attention aware mecha
aggregation in stereo matching, their effectiveness is limited by the low density and non-uniform distribution. To address this is
e many RGB-D registration methods rely on geometric and feature-based similarity, we take a different approach. We use cycle
nsport harmful payloads or independently cause damage, necessitating comprehensive exploration of countermeasures. MMAU
every possible object. While deep learning approaches may be effective for this problem, manually annotating 3D data for sup
hift clustering is a widely used method for image segmentation tasks. However, the traditional mean shift clustering algorithm is
n a scene from sequences of RGB images and sparse LiDAR measurements. The images are semantically segmented by a pre
er, we propose WeatherDepth, a self-supervised robust depth estimation model with curriculum contrastive learning, to tackle p
mited communication bandwidth, capturing comprehensive situational awareness by integrating connected agents' observation
fferent poses (and possibly including points belonging to the background) and we want to simultaneously reconstruct how all ob
s fail to extract a decent 3D DLO model due to different textures, occlusions, sparse and false depth information. To address th
one dimensional temporal waveform which encodes the arrival time of photons incident on the ToF sensor. Typically, a sensor
ns due to sensor resolution, which leads to a lack of detailed information. Recent advances have shown that Vision-Language M
roaches have attempted to leverage features from large vision-language models to encode semantics in 3D representations. H
s no external state information from physics modeling, simulation, visual markers, or contact as input. The algorithm improves o
 resented in practical settings, leading to poor performance on rarely-occurring categories if the class imbalance is neglected. In
robots to implement complex transformational movements in a lightweight and simple manner. However, utilizing the continuu
s of reorienting solar panels throughout the day in order to face the sun. The energy consumption increase that comes with sol
nforcement learning methods capable of optimizing the 3D morphology of a robot have been restricted to reorienting or resizing
posed a helical bistable soft gripper with high load capacity and energy saving. An off-the-shelf bistable steel shell (BSS) as the
oth architectures, the inverse kinematics can be solved in closed form and the singularities described by trivial geometric and a
ed shape that is capable of an anguilliform swimming gait. A three-module soft robotic fish was designed, fabricated, and evalua
ed wave spring structure directs the water drawn through the center of the robot by an impeller, enabling it to move smoothly in
he same rotation axis, this robot can achieve an axisymmetric workspace, resulting in a large reachable workspace. In addition
sistance, and resiliency. However, implementing tensegrity structures in robotics presents new technical challenges, primarily re
d to a consolidation of pretrained models, with general pretrained backbones serving as a starting point for many applications. C
deployment. SARA-RT relies on the new method of fine-tuning proposed by us, called up-training. It converts pre-trained or alr
ed, smooth gel surface and compact design equipped with a synthetic fingernail. We propose three distinct grasping strategies:
er modalities are trained from scratch. In contrast to the abundance of visual data, it is unclear what relevant internet-scale data
a rich source of data for learning such generalist robots. Unlike typical robot learning approaches which directly learn how a rob
s to complicated nonlinear equality constraints in the configuration space, which are difficult for trajectory optimizers. In addition
fectiveness of these approaches highly relies on the choice of CBFs. Inspired by the universal approximation power of neural n
stems leverage probabilistic state estimators such as Kalman filters to obtain a robot's belief, i.e. a probability distribution over p
s essential for guaranteeing a system's safety. However, control invariance does not imply that a system can move from any st
sk. Recently, collision detection using differentiable optimization has provided a way to compute the minimum uniform scaling f
d flexible manufacturing. Despite the desire for automated cloth manipulation, the thin-shell dynamics and under-actuation natu
at affect their performance, but are unknown a priori. This work develops a safety-verification strategy wherein data is collected
mic but also physical, indicating limited studies and comprehension regarding standard testing infrastructure and equipment. Fo
ite their impressive performance, these vision-based controllers can make erroneous predictions when faced with novel or out-
ance (which results in conservative behavior), or to learn a deterministic dynamics model (which is unable to capture uncertain d
h that can automatically characterize the error of a perception module from data and use this for safe control. The proposed ap
gside the completion of the exploration task. In this paper, we propose a modular exploration framework that (i) identifies safe fr
in the near future. However, ensuring safety with these advanced control methods remains a challenge. A well-known tool to m
robot manipulators. A well-known tool to ensure safety is the Viability kernel, which is the largest set of states from which safet
, they often assume known extrinsic safety constraints, overlooking challenges when deployed in real-world environments wher
ared actuators are intrinsically dangerous due to their high reflected inertia. Recent studies have shown that low inertia and hig
ount these uncertainties, probabilistic state estimators are often deployed to obtain a belief over possible states. Namely, Partic
m robot planning algorithms that rely on pedestrian predictions, to calibrating autonomous manufacturing to be high throughput
sed switching between multiple backup controllers to formally guarantee safety under bounded control inputs while satisfying dr
reedom (DoF). However, certain tasks in robotics, such as category-level pose estimation, involve non-uniformly scaled point c
present a method to accurately estimate pixel-wise uncertainty in semantic segmentation without requiring real or synthetic OO
or efficiency and numerical stability. Although Visual-Inertial Navigation Systems (VINS) have made tremendous progress in rec
hile ignoring the motion patterns of the target over a series of frames, which would cause performance degradation in scenes w
esents NGEL-SLAM, a low-latency global consistent SLAM system that utilizes neural implicit scene representation. To ensure
osing the overarching task into several manageable subtasks to facilitate policy learning and generalization to unseen tasks. P
errors, workspace obstacles, fruit swinging/moving when contacted, and varying stem and branch stiffnesses. The gripper con
 petition between supply chains. Enhancing the digitization of production lines is crucial for gaining a competitive advantage. Str
who are required to visually inspect cereal grains for each individual kernel. Since grain kernels (e.g., wheat, rice) are tiny with h
iculture seeks analytic tools at near single plant level, and this work explores how to improve aerial photogrammetry to enable i
or require specialized know-how and equipment. In this paper we demonstrate an affordable, simple, computer vision-based me
d to produce more food, feed, and fiber. Current breeding practices involve extracting leaf parameters on a small subset of the
p manipulation remains a difficult challenge due to complexities involved in modeling their deformable behavior. In this study, we
ers varying robot priorities, thereby extending the traditional multi-agent path finding (MAPF) approach. The proposed algorithm
onally, this task is done manually, which is time- and labor-intensive. Robots can automate phenotyping providing reproducible
igation schedules for crops. However, such measurements are currently being made via a time- and labor-intensive procedure
s is the precision alignment of the gripper and the target crop. An undesired dislocation can harm both the gripper and the crop
tion and segmentation of fruits from field images plays a crucial role in automating jobs such as harvesting, disease detection, a
complex geometry of plants, it is extremely difficult to identify and track the individual organs in different growth stages of plants
henotype need to be measured a time-consuming process when performed manually. We present a robotic platform equipped w
performance of the scenarios involving abrupt changes in dynamics. This paper proposes a hybrid controller enhancing transien
make small-sized UAVs with limited computing and sensing hardware capable of reliably performing the scanning task in clutter
ces safety constraints for precise landing and visual locking, in the presence of modeling uncertainties and external disturbanc
cal graph where nodes represent different spatial concepts (from low-level geometry to higher-level scene-scale reasoning) and
ta, but the limited resources onboard mobile robots (like processing power, network bandwidth, etc.) pose a challenge for the d
mmercial application of sweet pepper harvesting robots still faces challenges. This paper presents the development of a sweet
nting, harvesting and transporting is required. In this paper, a self-localization method by using LiDAR and a robust control meth
ming task. This paper presents EdgeSoil 2.0, a non-invasive, accurate, and real-time robotic system for soil pH prediction, a ke
ver, pruning also represents a uniquely difficult problem, requiring robust systems for perception, pruning point determination, a
ework to perform real-time and on-the-go detection of trees and key geometric characteristics (namely, width and height) with w
populated fruit. They rely on lower resolution maps and sizing methods that do not generalize to smaller fruit sizes. To overcom
nd efficiently perceive the plant nodes that need to be cut, despite the high levels of occlusion from other plant parts. We formu
ing containers make the automation of different farming operations challenging. In this paper, we explore the use of cobots (i.e
problem, however, existing methods are not optimized for dense prediction tasks in agriculture domains which results in degrad
ems,however, often perform worse in new fields or when the robotic platforms change. While we can alleviate the performance
mation of a critical task that arises in transporting logs from the forest to the sawmill: the log loading operation. This work is mo
ature such as trees or street lights. The approach includes three parts: mapping, bounding, and localization. To map and recon
g forests is a challenging task. Robotics in forestry can have the potential for substantial automation toward efficient and sustai
vironment. Utilization of available a-priori data can be a powerful tool for increasing efficiency. However, the relationships of this
group of robots to herd animals which consists of two phases, a surrounding phase and a driving phase. In the surrounding pha
echanism called a C-type expansion dowel and a novel hydraulic circuit MHSB that matches the modular structure. In this pape
sent PLASTR, a receding horizon optimization-based planning algorithm for robotic plaster trowelling. It samples trowelling seq
bots and structures are made from a set of compatible building blocks, or voxels, which can be assembled and reassembled in
ng construction, robotic navigation, long-term autonomy, and environmental monitoring. We focus on the semi-static scenario w
lation tasks based on an uncertainty model on the infrastructure's positional accuracy. Our uncertainty model can be combined
le of handling them all. One solution is to use a base controller and fine-tune it for different materials. Reinforcement Learning (
ement learning (RL) to automate construction machines. One of the challenges in applying RL to excavation tasks concerns obt
al integrity. It has been demonstrated through previous studies that relying solely on a single non-destructive evaluation (NDE)
 anguage-conditioned manipulation tasks. ERRA is based on tightly-coupled probabilistic inferences at two granularity levels. A
ackle grasp detection, a persistent challenge in robotics with broad industrial applications. Despite numerous grasp datasets, th
proved by anticipating upcoming tasks and computing an action sequence that jointly achieves these tasks. State-of-the-art met
mulated robotic environment using semi-structured natural language queries and scripted skills described using natural languag
eating robotic agents for performing open vocabulary tasks. However, planning for these tasks in the presence of uncertainties i
bilities, this work develops an efficient task planning algorithm for hierarchical metric-semantic models. We consider a scene gr
when actions fail and resort to retrying failed actions without resolving the underlying cause. We propose a novel approach (C
d grasps, existing methods incorporate semantic knowledge as priors into TOG pipelines. However, the existing semantic know
o existing systems, DOS features a unique CI/CD (continuous i ntegration and continuous deployment) architecture which allow
information is extracted and processed to control the positioning and heading of the aerial vehicle. A novel adaptive non-singul
el parametric uncertainty. We employ an adaptive Neural Network (NN), used to approximate the partially unknown system, in t
is synthesized for the systems linear in parameters by combining the classic ADRC algorithm with a recent Parameter Identifyin
e Optimal Control Problem (OCP) at each triggering instant introduces significant delays between state sampling and control ap
plant and actuator model parameters simultaneously. We provide a stability and convergence analysis with proof of asymptotica
e perception, state estimation, planning, and control. Adding to this complexity is the need to accurately identify vehicle model
Consequently, sudden changes in the environment or the robot's dynamics can lead to catastrophic behaviors. Leveraging rec
hattering on sliding surface is appended. By taking advantage of the finite-time convergence of traditional terminal sliding surfa
losed-loop QP control scheme can be prone to instability depending on how the gains related to the tasks or to the constraints
se safety concerns, we propose a novel approach that leverages the capabilities of digital twin technology by tailoring it to incor
ever, the conventional convergence-centric behavior exhibited by DSs may fall short in safety-critical tasks, specifically, those r
imal path tracking algorithm designed to maximize speed while ensuring safety for human workers. This algorithm runs in real-t
teractions present a unique challenge: they inherently block the person's hands and arms from the robot's view. This occlusion
ese interactions is a mutual understanding that (i) individuals are prosocial, that is, there is equitable responsibility in avoiding c
rent uncertainties in diverse human internal states, i.e., driving styles and rationality. To address these complexities, we propos
, leading to ambiguity in the development and fine-tuning of perception algorithms. In this paper, we introduce a methodology fo
vulnerable to infeasible and unsafe configurations upon physical perturbations. In this paper, we propose a novel control archite
pose a significant threat to deep learning models of AI-enabled industrial systems, making it crucial to develop methods to imp
n-board computing technologies. Utilizing AUVs in underwater human-robot interaction (UHRI) has witnessed comparatively sm
applications. In this paper, a suite of mechanical devices is featured, as several new asset management tasks were recently ad
al forces. Suppressing vibrations is difficult in our jet-actuated continuum robot because the movements of its mobile base cann
ved to maximize the success probability of such operations. In these settings, cellular-enabled collaborative robotic systems hav
rtainty awareness scheme to handle dynamic scene changes in previously explored areas. The framework is structured around
and traveling salesman problems. Since the time constraint is critical for search problems, this research proposes a Computatio
ective function is submodular and consists of a coverage function to cover environments and a balancing function to efficiently d
nspection missions with mobile robots, where known sensing ranges may be exploited. We investigate the relay placement, est
nd low-power computing hardware, the system consists of an experimental platform with rocker-bogie suspension. It operates in
t and the future of the state of complex worlds from poor observations. In this paper, as recommended in that previous work, w
ct labeling and exhaustive scene graph generation; less effort has been focused on the task of purely identifying and mapping l
on extensive labeled data or exhibit limitations in handling complex language queries, we propose LLM-Grounder, a novel zero
nreliable results when confronted with unfamiliar contexts, and it is challenging to obtain manual annotations frequently for new
the distribution of foreground object instances is skewed to the near side. This leads the training of the model to favor semantic
 their widespread adoption. Concurrently, recent breakthroughs in visual representation learning have sparked a paradigm shift
ects and the reduced texture and color contrasts among them. To address the issue of limited visibility, we propose a hierarchic
ve also demonstrated superiority in segmentation. However, multimodal solutions incorporating transformers are underexplored
ws a mobile robot with minimal compute to jointly optimize a dense language and geometric representation of its surroundings.
k, we investigate the use of virtual reality technology as an implementation for robot environment modeling, and present a techn
r real-time semantic mapping from RGB-D sequences that combines a 2D neural network and a 3D network based on a SLAM
by segmenting the grid map of the currently explored parts of the environment into regions, such as intersections, pathways, d
ps. This paper presents Open-Fusion, an approach for real-time open-vocabulary 3D mapping and queryable scene representa
se Mask4Former for the challenging task of 4D panoptic segmentation of LiDAR point clouds. Mask4Former is the first transform
nment by predicting semantic classes for each 3D point and to identify individual instances through different instance IDs. To d
spatial patterns and semantic co-occurrence relations among objects in the scene. Prevailing approaches extract scene priors
mation. We propose a graph-based learning methodology for the computation of belief (also referred to as expectation) on any g
s these semantic topics in continuous space as densities. However, ambiguity in learned topics is a disadvantage of such Dirich
dditional source or modality of 3-D information, such as depth or disparity images, using a pair of independent encoders. The e
ency on extensive labeled datasets, presenting challenges in acquisition and exhibiting incapacity to segment novel classes, es
do not provide direct motion information and face limitations under adverse weather. Radar sensors overcome these limitation
nterest. This paper highlights the fusion overconfidence problem, in which conventional mapping methods assign high confidenc
ent limitations of onboard sensors, which include sensitivity to detection range and susceptibility to occlusion by nearby vehicles
focus on designing a multi-modal fusion module without consideration of the nature of multi-modality inputs. Therefore, the net
nable hierarchical semantic representation of urban dynamic scenes and processing information from multiple agents are still c
mera BEV estimation, and a perception of distant areas due to the sparsity of LiDAR points. In this paper, we propose a BEV fus
ments often contain occluded and unknown regions, and without accurate prediction of unobserved obstacles, the movement of
uds. Current techniques enhance the point cloud by the supervision of LiDAR's data. However, high-performance LiDAR is nota
e driving perception. In contrast to the vast majority of existing arts that focus on how to improve the performance of 3D target
under such varying lighting conditions. Therefore, recent works utilize novel event cameras to supplement or guide the RGB m
e LIDAR semantic segmentation task, but they suffer from different problems. 2D-to-3D fusion methods require strictly paired d
techniques aim to maintain the merits of optical (EO) and infrared (IR) modalities, e.g., retaining functional highlights and captu
to establish especially for distant scene points. Recent works try to solve this problem by designing a network that learns missi
erent sensor modalities, such as LiDAR and camera. However, current works are based on raw data or network feature-level fu
ing motion states continuously over time. Systematic derivations of analytic kinematic interpolations and spatial differentiations
mera has garnered considerable attention due to its robustness and high accuracy in diverse environments. However, leveraging
same spatial coordinate system via motion compensation to distinguish the moving object. However, they neglect the potentia
e detection due to the contrast between LiDAR's sparse data and the dense resolution of cameras. Besides, discrepancies in t
image data, or their encodings, tends to yield dense depth maps with significant artifacts, blurred boundaries, and suboptimal
a skill that robots need to emulate for intelligent adaptation in dynamic settings. Training of the spatial anticipation ability, which
ents. Existing methods typically rely solely on RGB images or fuse the modalities at later stages, which can result in lower accu
method uses a sequence of LiDAR data and camera images captured while moving the system. Once the camera motion is calc
ods suffer from performance drops due to the limited dynamic range and motion blur. Therefore, recent works leverage novel e
e of homography in the presence of occlusions, while a Bayesian filtering approach generates both a homography estimate alo
essive motions. Field robotics expose robots to various hazards, including steep terrains, landslides, and staircases, where sub
plicated surroundings, which results in their severe degeneration in complex environments, especially on jolting robot platforms
 practical or convenient. Gesture recognition is an alternative, but often not reactive enough when conventional cameras are use
ggested for application in force control-based robots and position/velocity control-based robots. Strong impact forces have a neg
ns over point cloud data for safe vision-based control. We take advantage of surface geometry to locally define and synthesize
that forms a proximity array to obtain the necessary information in sensor space. For the task considered, we provide closed-fo
which further affect the robustness and accuracy of the entire system. In this paper, an image acquisition scheme with image b
sed tactile sensor. This sensor employs a 3D multi-layer grid-based design, inspired by the Magic Cube structure. This structure
(DL)-based approaches face scalability issues when dealing with various types of microphone arrays. To address these issues
peech is the preferred way of communication, especially in indoor environments. This paper proposes a speech module to rotat
ble challenge, as it involves the decomposition of strokes and dexterous utensil control. Previous efforts have primarily focused
e problem of estimating 3D shapes of fruits when only partial observations are available. Generally speaking, such a shape com
udy sites of interest. Typically, rover global localization has been performed manually by humans, which is accurate but time-co
adapt their body posture while walking. However, robust deployment in real-world applications is still an open challenge. In this
s paper, we first developed a top-level GPSR system for worldwide competition (RoboCup@Home 2023) based on multiple fou
equent steps: interference resolution and quilt spreading. By leveraging the DWPose human skeletal detection and the Segme
put abstraction and policy output guidance. Classical approaches achieve high precision but require clean keypoint corresponde
However, it is labour-intensive to calibrate robots one-by-one in practical use. In this paper, we propose a neural uncalibrated V
ations. In the current healthcare domain, an intermediate robotic communication platform is essential for distributing equal medi
continuum manipulator often has the problems of limited workspace and weak carrying capacity. Compared with traditional ma
s study, we developed a world-first, elongated fiducial marker, capable of high-accuracy 6-DoF measurements, designed to be i
manipulation actions, a robot's perception framework needs to execute a series of these methods in a specific sequence each
alizing complete area coverage. However, floor cleaning robots with fixed morphologies have difficulty in achieving a high area
ddressing this imperative, this study contributes to the field of Large Language Models (LLMs) applied to task and motion plann
emi-definite Hessian, which allows us to extend our previous work on the convex formulation of compliant contact to model larg
cess often relies on simulators to replicate realistic crowd behaviors, ensuring cost-efficiency. Existing crowd simulation method
ntial in this area. Nonetheless, they often face challenges related to low sample efficiency. To overcome this challenge, some re
ns is physics-based simulation of robot-object interactions. This involves the challenge of accurately modeling and implementing
ndulations derived from observed real fish kinematics. This approach often neglects higher-order undulations, thereby missing
aluation of said algorithms. With the regained interest in lunar exploration, there is also a demand for quality simulators that will
ion to generate canonical scenes at scale. However, manual creation is not scalable, and automatic generation typically use ru
s, Adams offers features including intersection-free collision simulation and stable LCP solutions for multiple frictional contacts.
erfect representation of the real world and the so-called sim-to-real gap between simulation and reality hinders the deployment
detectors are either independent or loosely coupled, so they cannot fully adapt to each other. In this work, we propose a tightly
able for a wide variety of navigation tasks. For example, datasets that include multiple cameras often have short trajectories wi
ion and high noise, of sonar imagery for odometry and Simultaneous Localization and Mapping (SLAM), the proposed Direct Im
wise much structural detail may not be kept. In this paper, a novel paradigm of LiDAR mapping and odometry is designed by le
s a single point of failure, from which it can often be difficult, if not impossible, to recover. As such robust odometry solutions ar
paper presents a generic sparse visual SLAM framework capable of running on any number of cameras and in any arrangeme
omaly detection and handling, real-time dense color mapping, and robust localization in diverse environments. We tightly integr
it SLAM has shown encouraging and impressive results. However, the robustness of neural SLAM, particularly in challenging o
ave difficulties with object-level localization and reconstruction and struggle in dynamic and illumination-varied environments. W
ith space explored instead of time in a principled way, so that the number of time-indexed poses can be limited while their influe
ess the problem of effectively detecting loop closures in LiDAR SLAM systems in various environments with longer lengths of se
which further determines the direction of optimization and ultimately affects the accuracy of localization. However, estimating n
us on two odometry estimation approaches followed by a mapping step. The first one is a new point-to-point ICP approach tha
f pairwise relative measurements. The optimization procedure can be negatively affected even by a single outlier measurement
erest in plug-and-play algorithms that make full use of pre-trained backbone networks for multi-scale feature extraction and leve
mputational efficiency. Some researchers have attempted to enhance localization accuracy by integrating the modeled object re
nsistency between semantic/depth and pose tasks presents a significant challenge, as it may disrupt cross-task feature represe
onal functions tend to yield notoriously hard optimization problems. In this work, we present a novel least squares method for 6
is exploited. The main bottlenecks faced by state-of-the-art work in this field include the high computational complexity of map
measurements and robot poses, and (2) a 3D scene graph, as a high-level representation of the environment that encodes its d
n the world based solely on images. This present work is the first tutorial paper on visual place recognition. It unifies the termin
rmance of these sensors is degraded by challenging environments, e.g. adverse weather conditions and low-light scenarios. Th
or each detected ORB feature. When integrated with traditional visual ones, these descriptors aid in establishing accurate tenta
r integrate with inertial measurement unit (IMU) in visual SLAM has potential research value. This paper proposes a novel deep
s in advance. Furthermore, the approach is robust against small overlaps and dynamic objects, since no direct correspondence
cy and robustness because they simply compensate for motion distortion by constant velocity motion assumption without accur
rge-scale reconstruction, but either require groundtruth poses or run slower than real-time. In contrast, LONER uses LiDAR da
y fusing LiDAR scans and IMU measurements, these systems can reduce the accumulated drift caused by sequentially register
AR-inertial odometry frameworks face challenges in real-time processing due to the voluminous data generated. This paper intr
lan in dynamic environments, it is essential to employ representations capable of handling moving objects. Dynamic SLAM is a
mapping (SLAM). This map is a collection of voxels that contains one plane feature with 3DOF representation and correspondin
d consistent bundle adjustment method for lidar sensors. The method employs edge and plane features to represent the scene
ation process,leading to reduced map quality and hindering normal robot navigation. Online dynamic object removal methods ar
ImMesh comprises four tightly-coupled modules: receiver, localization, meshing, and broadcaster. The localization module first
zation, and mapping is essential for the success of downstream control tasks. However, when designing a new robotic platform
er, existing methods employ either hand-crafted or solely learnable point features and are thus limited by the feature attributes.
he 3D object by the camera has been rapidly improved in recent years with the development of deep learning, the estimate still
measurement data, the larger is the computational burden. To reduce computational burden, one can remove a portion of meas
per tuning of noise models. During inference, these models are used to weigh different blocks of the Jacobian and error vector
h reflective surfaces (e.g. mirrors). On the other hand, electromagnetic waves exhibit better penetration properties when the wa
n the Doppler velocities and radar cross-section information. The filter for static radar points, correspondence estimation, and re
DAR can detect distances up to 40 m, and it senses only four distance measurements per scan. In open spaces, RGB-D SLAM
ations reliant on specific feature extraction algorithms. In this paper, we propose a novel method based on incremental and gen
mobile and assistive autonomous systems. In this work, we introduce LEXIS, a real-time indoor Simultaneous Localization and
llance, and augmented reality. In this paper, we introduce a novel methodology that extends state-of-the-art Pose Graph Optim
representation of the 3D environment is indispensable. We present a fully tightly-coupled LiDAR-Visual-Inertial SLAM system a
er robot or device. In cases such as multi-robot or human-robot collaboration, localizing all agents in the same map is even nec
the Next Best View (NBV) based method. However, the NBV method cannot guarantee complete scene coverage and often ne
hat covers all the navigable areas. In this work, we propose a novel probabilistic active loop closure framework which attempts

mination. It also means that every measurement in a scene can occur at a unique time. Handling these different measurement t
 own MSCKF). We define a symmetry group and corresponding group action that allow specifically the design of an equivariant
n lane lines. This method effectively utilizes the lateral constraints provided by lane lines to eliminate and relieve the incrementa
strians. Pedestrians may cooperate with the robot by avoiding it actively or ignoring the robot during their walking, while some p
art of the “tidying.” How can a robot reach that conclusion? Although large language models (LLMs) have recently been
, we investigate different rejection strategies with the goal of changing customer behavior. The implemented strategies have be
humans and amplify their comprehension of robot decisions. We model the process of explanation generation for robot navigati
ent learning frameworks to train policies for efficient navigation. However, their performance deteriorates when crowd configura
nabling robots to perceive and interpret the visual world through detailed text-based descriptions. To assess both user experien
y unravels the details of HRI within interactive storytelling contexts. Through a between-subject experiment with 28 participants
mpetency are quite scarce. In this paper, we present a prototype of a robot-mediation framework for cross-cultural communicati
ntial to facilitating robot adaptation. However, when the task is complex or the robot action space is large, it requires a large am
robotic structures. Based on this concept, we develop a prototype demonstrating enhanced flexibility and reconfigurability in el
nment exploration. However, the intrinsic complexity of continuum robots requires a significant amount of time for their motion p
nature, capturing fast-moving objects is extremely challenging in engineering implementations. In this paper, we report an ultra
estorability, and the compliance of the robots during deformation will be reduced because of the jamming effect.Here,we presen
limate change. Due to the extreme conditions, exploring the deep-sea traditionally requires costly equipment, such as specific d
d for securing a vine robot to objects or its environment that exploits the unique eversion-based growth mechanism and flexibilit
s than a rigid one. Previous studies often use a fully observable formulation, requiring external setups or estimators for the peg-
rently safe and adaptive to environmental conditions. In this paper, the interaction with the environment is performed at the too
larly appealing for agriculture, where items of produce from the same crop can vary significantly in shape and size, and delicate
ted functions that can locomote in complex 3D environments are yet to be developed. This study addresses this challenge by p
bject retrieval and object manipulation. In response to these challenges, we present the Polyflex design concept for fabrication
as well as having excellent passive adaptability. However, limited by insufficient structural stiffness, achieving high payload and
unique features, such as high adaptivity to various unstructured environments and safe interaction with living beings through th
process of the spacecraft. There is currently a lack of suitable unlocking solutions for solar wing re-unfolding. This paper propo
ner diameter and the inter-luminal crossing within the luminal tracts lead to challenge for effective robotic intervention. Inspired
nment. To achieve high maneuverability, this study demonstrates a pneumatic soft crawling robot equipped with wheel-legs cap
hes sustaining damage) and near-ripe and ripe blackberries are difficult to distinguish in normal lighting conditions. These challe
ule of soft materials, their control systems are usually larger and heavier. Therefore, small underwater robots often tethered, i.e
telligence. In this article, we propose an automated design technique that optimizes the desired load-displacement behavior of
magnetic surfaces, or two-dimensional ar- rays that could shape and control an incident electromagnetic field in ideally any man
atly benefit from their central fold and soft surface. As such, we develop a novel structurally compliant soft palm, which enables

en reported, however, most of them focus on adapting to the overall shape of the structure, while the adaptive grasping of small
n human experiences and skills, which are both time and labor-consuming. In this paper, we present a rapid design and fabrica
very, minimal invasive surgery, and manipulation of delicate objects. In this study, we introduce a liquid robot composed of ferro
anical conversion efficiency and great flexibility, they have been extensively used in soft robot applications, such as soft gripper
rammable bending, twisting and elongation deformations. The system comprises an elastomeric tube reinforced by lattice meta
ow load capacity and small working space. To address this problem, we propose a novel soft hybrid-driven manipulator with co
res a complex multi-step manual process, which significantly increases their cost. We present a design of a multi-material 15 D
ominance of flexor muscles over extensors, and the limited surface area available, developing an extension glove with thumb a
 actuator (SFTTA) is proposed. The SFTTA is composed of a folded flat tube sandwiched between two silicone rubber laminate
e and repeatable operations, particularly when faced with buckling during retraction. Although a va- riety of retraction mechanis
ts are less adept. In such cases where a borescope or endoscope can’t reach, disassembly or surgery are costly. One prom
enter and leveraging the spring elements not only as reinforcement structures but also as inductive sensors. Additionally, soft p
ly reject external disturbances are two drawbacks which limit the scale of continuum designs, preventing widespread adoption
ontrary, contact should not be considered as a problem, but as an opportunity to expand the range of UAVs applications. In this
nd low durability, while approaches that rely on direct measurements on the gripper often have limited resolution and are difficu
with a genetic algorithm. In particular, we propose to perform the automatic synthesis of the inner structure of pneumatic actuat
activities. However, most soft robotic gloves are completely composed of flexible structures. Although they have high flexibility a
mblies are prone to faults, difficult to clean, and require frequent maintenance that halts the production lines for considerable am
ankle instability by providing stiffness support. The soft-rigid layer sliding mechanism of the structure is designed to achieve a w
-of-concept soft, capacitive origami sensing module (OSM) that can measure forces during neurosurgical retraction. Using orig
d by laser patterning an off-the-shelf medical tubing, allowing for bidirectional responses to large curvatures in two bending dire
on robotics. However, most of the existing prosthetic interaction methods focus on the use of myoelectricity to classify finite ges
studied towards curvature estimation, we look into being able to sense local deformations. To accomplish this, we use a widely
on forces which is essential for more contact rich tasks. We propose a framework that enables robotic imitation of contact from
he generalized dynamic modeling and precise control elusive. In this work, we present a generalized motion control framework
n constrained environments. With the miniaturization of imaging technologies and the development of novel control algorithms,
soft bending actuator that can change its stiffness using fiber jamming. The actuator is made of an elastomer tube filled with di
or example, vines in nature direct growth via localized responses embedded in the cells of vine body, allowing an organism with
nt degrees by simultaneously tailoring the structural characteristics and orienting the input pressure. To date, design of the morp
with a recognition that electronic powering may limit this achievement. Alternative powering sources compatible with soft robot
actions with a minimal footprint is especially important in the medical context, where portability and unhindered patient/user mo
zable vendor-agnostic control cabinet and an innovative, self-developed software architecture based on exchangeable multi-ra
r more important. Automated DevOps processes and microservice architectures have already proven successful in other doma
s a code-based language for specifying a generative model of the system, making model specification easier and model samplin
zation problem, we introduce a benchmark suite encompassing a unified format for robots, environments, and task descriptions
trengths and weaknesses. Comparisons of different methods are limited due to the lack of a broadly adopted and standardized
Our key insight is that combining a source of cooking domain knowledge with a formalism that captures the temporal richness of
n task-specific, inflexible code. This work explores the potential of Large Language Models (LLMs), like ChatGPT, to automate
ges a centralized scheduler in a distributed set up to prevent collisions while the robots move independently. We argue that this
utdoor environments such as orchards and vineyards, its functionality is notably limited by the presence of obstacles and/or situ
ation of future landing and mission sites. Here, robot- and astronaut-scale obstacles are the most relevant to resolve, however,
y environments presents unique challenges, including uncertainties and disturbances. This paper introduces an innovative app
ace debris is dependent on the ability to reliably maneuver the tether-net system -- a flexible, many-DoF (thus complex) system
e to domain gap as they are primarily trained with synthetic images due to the inaccessibility of space. OST aims to close this g
ng reinforcement learning eliminates the necessity for manual feature design, therefore simplifying the problem and empowering
ying Deep Learning-based Spacecraft Pose Estimation techniques. However, due to limited access to real target datasets, algo
sion-free operation with far look-ahead distances. Utilizing a novel converging covariance cell representation, our approach effe
st-in-space” problem that we call Visual-Inertial Navigation for Satellites (VINSat). VINSat performs OD using data from an in
, sloshing and flexible appendages are modeled. A novel system identification scheme is developed, which identifies all parame
 methods, and path planning for multi-branch robots, still need improvement before practical application. This paper presents a

ielded from dust and the environment. This benefit of geometric simplicity has led to the spherical robot becoming an advantag
situ according to mission specific requirements. To facilitate flexible configurability, a multifunctional interconnect is used to me
or perching on and grasping flat objects in space for free-flying robots. To grasp and anchor onto these flat surfaces, the grippe
aditional wheeled robots, spherical robots face certain limitations. The spherical robot is composed of a closed spherical shell s
ever, these parameters may change on orbit for several reasons. Also, during an SMS task, excitation of flexible appendages, s
ng markers and additional sensors. We propose the first method using a single standard laparoscopic monocular RGB image. I
tracking have been proposed. A unifying disadvantage of these methods is the high computational cost of simultaneous imagin
a lack of kinesthetic coupling between the surgical field and the surgeon. This paper presents a framework for sensorless trans
e for real-time recognition and prediction of surgical gestures and trajectories based on short segments of kinematic and video d
or explicit curriculum. The proposed method is based on a recursive decomposition of the original task into a sequence of sub-
ch as pushing tissue, pressing cloth, and needle threading. In this work, we isolate the problem of nonprehensile manipulation
se estimation of surgical instruments is critical to enable the automatic execution of surgical maneuvers based on visual feedba
bles doctors, in collaborative interaction with robots, to perform traditional or minimally invasive surgeries with improved outcom
been increasingly investigated. However, existing approaches encounter difficulties in long-horizon goal-conditioned tasks due
and the fixed length of the bronchoscope’s bending segment, along with its external propulsion property, pose challenges,
a calibration create errors in this coordinate frame transformation. These errors result in poor localization of robotic manipulator
atches. However, it is difficult to adopt these techniques in surgical scenes, where tissues are changing in shape and appearanc
n be powered by simulation, which ensures realistic deformation by adhering to the governing physical constraints and allowing
rface anatomy, such as the location of tumors, which necessitates accurate and efficient surgical sensing. In this work, we prop

suffer from the limited availability of annotated datasets with both kinematic and video data and an inability to generalize to unse
his work, we focus on a substep of cataract surgery, the Continuous Curvilinear Capsulorhexis (CCC). With a high complexity, t
lenge. In this paper, we present ORBIT-Surgical, a physics-based surgical robot simulation framework with photorealistic rende
both real-time and accurate performance for vascular intervention robots. This work bridges high-accuracy 3D-2D registration t
curved tubes usually composed of NiTi and are arranged concentrically. As each tube is relatively rotated and translated, the ba
each other, and relative displacement of the tube bases generates bending in the CPPR. Previous CPPR kinematic models ass
terference, reliability, resolution or image contrast. This often leads to frequent needle adjustments and reinsertions. To addres
on is used to steer the catheter tip during navigation. However, today's MSRC systems require fluoroscopic imaging for cathete
ace supporting the effective delivery of an artificial valve by tele-operation and endoscopic vision. This manuscript presents our
r and its surrounding anatomical structures remains a difficult problem in existing robot-assisted puncture operations. In this pa
roller is fully distributed and only requires local state-estimation and scalar measurements of inter-agent distances. It does not
ly on-board and can be deployed on nano-drones such as the Crazyflies. The swarm acts on a simple set of rules to identify an
ssibly heterogeneous robot swarms. Augmented reality can provide an interesting solution to these problems, as demonstrated
ms: precise motion control, physical interactions with the environment or neighbor robots, and functionalized shells. Few current
previous FireAnt-series robots by redesigning the locomotion drive system to improve mechanical and attachment reliability wh
actions we are better able to design emergent behaviors in robot collectives composed of agents that exert attractive and repuls
al robots. Automatic design has concentrated on missions that a swarm can execute independently, interacting only with a stati
erforms its task as expected without collision, and this becomes more challenging as the number of drones scales, possibly into
mporally extended navigation. Agents in Long-HOT need to efficiently find and pick up target objects that are scattered in the e
 l state and the motion planner verifies whether this action sequence is geometrically feasible for the robot. However, state-of-th
established framework for the simplified definition of tasks using two manipulators. By integrating conformal geometric algebra,
MP). However, prevailing methods often struggle with the computational burden and intricate combinatorial challenges, limiting t
nt partitioning algorithm is presented to extract diverse information for each sub-task. First, LiDAR data is classified as potentia
ater decisions concern what to compute for such skeletons (e.g., action parameters, bounds, RRT paths, or full optimal manipu
ese aspects increase the difficulties in obtaining gas measurements, encompassing both qualitative and quantitative aspects. M
ally dealing with hybrid TAMP problems that involve abstract and geometric constraints. However, LGP does not scale well to h
cult for Task-and-Motion Planners, as they require interdependent regrasps and solving hard motion planning problems. In this
nables: 1) its efficient population from natural language texts using semi-automated procedures based on Large Language Mod
ble next to the fork''. To successfully execute these long-term horizon tasks, we argue that an agent must consider its past, i.e.,
e with complete sensor and actuator integration, thereby ensuring adaptability to evolving scenarios. We discuss essential parts
ulate the plan refinement problem—which, given a task plan, finds valid assignments of variables corresponding to solution tra
. Our approach uses the Timed Partial Orders (TPO) model to specify these constraints. We translate this problem into a Trave

paths. As the task planner and the motion planner are closely interconnected TAMP is considered a challenging problem. In th
d. In the budgeted version, each robot is subject to a travel budget limiting the distance it can travel. Our interest in this problem
n (CAPM) problem, we model the uncertainty in the perception process and devise a key state/task planning approach that con
s using pegs as fasteners. As such it assesses planning and execution capabilities, and poses challenges in perception, reason
rm caused by such edged or pointed impactors is less well studied than for blunt impactors. Consequently, the standards spec
hallenging and struggles to scale to realistic problem settings. We propose to accelerate task planning over an agent's lifetime
oncepts useful to a robot’s understanding of tasks and its environment for object-level planning. Prior to this work, little has
ect planning and efficient execution for cooking tasks. Our method first uses a large language model (LLM) to retrieve recipe in
easing number of agents, the search space for task planning expands exponentially. Hence, conventional methods aiming to fin
rmulation of strategies for frontier point detection and task allocation largely determines the overall efficiency of the system. In t
s of freedom of the robot to accomplish a desired task. In this context, developing an easy-to-learn and easy-to-use human-rob
gh-quality demonstrations. Poor-quality demonstrations can negatively impact the robot's ability to learn and generalize. In this
ransferability of cognitive capabilities of human workers to robots operating remotely or in high altitude workspaces. The system
vercome this difficulty by predicting user intentions and performing some parts of the task autonomously to decrease the dema
f Mixed Reality (MR) has yielded tangible benefits in human-robot interaction. MR provides an immersive environment for intera
Internet. Different from the existing just-noticeable-difference (JND) based MMT, our method avoids continuous transmission o
The proposed control framework aims to improve bimanual robot telepresence quality by reducing the user workload and requir
ential to prevent lump iron from scattering or blocking of molten iron, is performed by equipped human workers using a long stic
tance with the Shared Control Templates (SCT) framework -- facilitates user control, even with low-dimensional input signals. H
ration frameworks, such as cognitive strain, control complexity, robot compatibility, and user evaluation, we propose TELESIM,
er, and (ii) synchronizes the walking pace by predicting user's stepping frequency using motion capture data and a deep learnin
per, we address the problem of placing objects in the context of a teleoperation framework. Particularly, we focus on two aspec
e, we developed a control technique for minimizing the robot visual appearance when observed from an eye-to-hand camera. S
e filters, iRoCo optimizes a combination of precise robot control and unrestricted user movement from ubiquitous devices. We d
rol of many components of the robot to immerse them in the remote environment, but operators still struggle to complete tasks
d to assist in executing the user's intentions. To this end, robust and prompt intention estimation is needed, relying on behavior
mic Mobile Manipulation (DMM). This paper presents a framework for DMM via whole-body teleoperation, built upon three key c
 the difficulty of controlling various degrees of freedom at the same time. 3D Autocomplete helps mitigate such limitations by as
cloud domain. Full fine-tuning is generally effective but leads to a heavy storage and computational burden, which becomes ine
wise annotations. While previous MM-UDA methods can achieve overall improvement, they suffer from significant class-imbala
policy in the source domain to the target domain becomes an attractive solution. Previous research has typically focused on dom
mputation and storage, and depend on time-consuming prompt engineering. We address the problems from three aspects. First
hen facing environment changing, resulting in severe degradation of transfer performance. For BEV perception, we figure out th
s higher values for grasp candidates that are more likely to succeed in grasping. This evaluation model serves as an objective f
ver, when simulation takes a long time, e.g., in an unsteady 3D fluid simulation with interactions to the controllable solids, existin
d to equipment lifespan and safety requirements, acquiring a number of samples directly from the physical environment is often
le fostering social inclusion. Robot technologies are employed to encourage user participation and self-expression, enhance so
pose a robot foot that mimics the longitudinal arch and transverse arch of a human foot to absorb Ground Reaction Force(GRF
g rapid connection with any other module. This module possesses a rotational degree of freedom, which allows a chain-like rob
robot demonstrates stable adsorptive property and mobility on the surfaces of large, complex, and variable curved components
ts and windows. Our demonstration shows that conventional motion planners can lead such a dexterous task insertion with a ro
perations and complete entire cooking procedures autonomously. One of the standout features is the use of proprioceptive act
ree-moving Wi-Fi device. It is accurate for parsing the human and scene interaction in 3D space. Most previous Wi-Fi sensing t
ant efforts, humanity has yet to overcome these challenges. We need support. Imagine a future where machines and robots are
focuses on integrating technologies to facilitate human-machine interaction for collaborative tasks, particularly in the industrial c
uces accurate 3D human meshes in physical sizes and 3D translations, and enables real-time deployment in both moving and s
tial for leveraging deep learning algorithms, particularly for collaborative robots (cobots). Cobots are especially suited for Huma
orcement learning (DRL) methods we developed, it now has reached a new level of autonomous dexterity never before shown
g a large dataset of prior observations to model the behaviour of other agents and to determine the ad hoc agent's behaviour. I
ng wireless IPS (e.g., Wi-Fi, blue- tooth, UWB) is limited by the pre-deployment of anchors (or beacons) and signal calibration,
ly available products or are 3D printed. This makes it easy to build and modify as an open-source project. Communication with
, the methods with Autonomous Underwater Vehicles (AUVs) have been gaining attention for marine animal tracking, particular
as their contact interface is mainly composed of unrigid and burnable silicone rubber, conventional soft grippers are not applicab
ntion, which has been proven to promote the cognitive, physical, and emotional well-being of dementia patients. However, the n
, which encompass various heart abnormalities. Patients with SHDs typically require open-chest surgery. Nowadays percutane
or research and development can provide better access to hardware, and thus accelerate advancements in the personal roboti
s. However, human motion isn't naturally suited for robots due to differences between human biomechanics and robot dynamics
two slender carbon fiber legs bearing the majority of weight in the feet. Despite each leg possessing only one active degree of f
ities operating in the same celestial vicinity. To enhance operational efficiency, these robotic entities must collaborate, share da
g tasks, commanding manual adjustments of grip force to prevent dropping or damaging held objects can be a significant physic
place task performed by the UR10e robot. The digital twin of the robot cell is produced in Solidworks but operated in Unity. Th
aintenance, or surveillance tasks in hard-to-reach or hazardous areas in towers, factories, bridges, power stations, and ships. E
dent lives in their own homes, achieved through the application of intelligent robotics. To realize this objective, we envision esta
d significant in multiple fields ranging from biology and archaeology research to engineering applications. In this video, we prese
d dynamics has posed challenges in achieving autonomous control, thus limiting their capabilities compared to their traditional
s not to moves the entire structure but to extend from the tip. As a result, even in a narrow gap, it is expected that movement w
metic principle, we have designed a buoyancy regulator for small deep-sea robots, the regulator consists of a silicone rubber sh
geries? In this video, we introduce the work of ILSR Lab in the domains of continuum configurations, surgical robot force sensin
 es-Parallel-Reconfigurable Tendon-driven Supernumerary Robotic Arms requiring clever end-conversion tools was proposed to
etween these devices and its user. However, many challenges remain on how to achieve effective collaboration between the ro
nd its enhanced performance on the human body are research issues of great concern. In situations where human limbs are h
nning systems. We rely on our Latent Space Roadmap (LSR) framework, which builds a graph in a learned structured latent spa
e-based camera. We propose a method of 3D model-based object tracking directly from events captured by event camera. To
asping of deformable objects and visual navigation in a non-rigid environment, NRSfM has had limited applications in robotics d
methods for sensing require multiple embedded sensors or markers, leading to increased complexity, cost, and increased poin
ns and contrasting with the background contextual information instead of sticking to only the semantic visual cue such as boun
ntal problem for robotics. In precision agriculture the challenge of achieving a satisfactory solution is amplified by extreme cam
with both accuracy and robustness. Prior to this, there have been several works that have achieved satisfactory results by fusin
al SLAM system to estimate inertial variables without updating camera poses, potentially compromising accuracy and robustnes
omputationally constrained hardware. Previous works leave a gap between CPU-based systems for robotic mapping which, due
osition of the learning load into modular components. In this work, we investigate a principled way to syntactically combine thes
autonomous repeated tasks under challenging circumstances, but research within T&R has predominantly focused on motion p
sentation and corresponding losses for visual-language model (VLM) fine-tuning, which can handle complex object-level querie
h special needs in indoor movement. To enable intelligent coordination of this system, a key component is the localization of ea
odel car across offroad, unstructured terrains without relying on predefined maps. Our innovative approach takes inspiration fro
ming. Can an autonomous robot directly navigate in previously unseen environments using coarse maps? In this work, we propo
dimensional (2D) patterns and RGB cameras for recognizing object IDs and poses, achieving long-distance recognition necessi
g methods have high computational costs and are unsuitable for embedded devices. Although some previous works have focu
ting outdoor VLN models predict actions using a combination of panorama and instruction features, this approach ignores objec
trained with data collected by robots of different types. However, the navigation methods' performance is still limited by the scar
a crucial role in achieving successful object finding. However, existing knowledge-graph-based navigators often rely on discret
ability map, the POLo score allows the agent to make data-driven decisions for efficient object search. We further enhance the f
herent in foundation models. But for the locomotion part, most works still depend on map-based planning approaches. The gap
uiring re-training the networks to achieve peak performance. However, obtaining accurate ground truth data for new training da
Established approaches have primarily focused on geometric navigation solutions, which work well for structured geometric ob
v, a novel end-to-end underwater navigation solution designed to navigate robots over Objects of Interest (OOI) while avoiding
ssion success. However, coping with out-of-distribution data from novel terrains or appearance changes due to lighting variation
severe artifacts, deteriorating blur, and unclear details. In this paper, we propose a novel ambiguity-aware network (ASP-LED)
of pretrained vision transformer (ViT), and can obtain keypoints on multiple object instances of arbitrary category after learning
utilize complete occupancy labels in 3D voxel space for supervision. However, the expensive annotation process and sometime
mera poses, more recent works enforce multi-view geometry constraints through optimization embedded in the learning framew
going substantial transformations. We introduce Doduo to learn general dense visual correspondence from in-the-wild images a
ese individual tasks, integrating specialized models into a unified vision pipeline presents significant engineering challenges an
upervised learning methods encounter limitations due to data scarcity and challenges in label acquisition. To address these issu
urate scale-aware depth prediction on autonomous driving scenes utilizing inter-frame poses obtained from inertial measureme
olution, and inference speed. However, as a relatively nascent vision sensor compared to Active Pixel Sensor (APS) devices su
rly in cluttered dynamic environments. Imbuing the rich temporal information contained in the video of scenes has the potential
as unique advantages, including night-time vision. However, the existing depth estimation method remains challenging in robus
gnition predominantly cater to pre-defined closed-set paradigms, ignoring the challenges of real-world open-set scenarios. Such
a single RGB-D image, operating at the category level and eliminating the need for CAD models during inference. While existin
utions to these challenges by virtue of their high temporal resolution and high dynamic range. However, they have limited perfo
However, generating controllable videos from image-text pairs is an important yet underexplored research topic in both robotic
es. Conversely, visible images retain more texture information at higher resolutions, particularly during the daytime, but perform
e object contour well on the image while still remaining reasonable on the 3D space. These requirements bring significant challe
ained models oriented towards understanding 3D geometry, followed by self-supervised finetuning on non-annotated videos. W
BEV is pivotal for various autonomous driving applications. Creating accurate street-view images from BEV maps is essential fo
driving scenarios, complicates the task of precisely reconstructing scene geometry. Such challenges often lead to diminished q
due to the sparsities and the lack of information when only event streams are available. In this paper, we utilize motion, geome
high-resolution imagery at a lower cost for SLAM applications, yet they fall short in accurately estimating 3D positions and cam
e image outpainting techniques aimed solely at generating aesthetically pleasing visuals, these applications demand an extend
of RGB images have significantly improved the accuracy of pose, the majority of models still face one or two limitations. First, fe
fusion-based policy iteratively denoises action sequences from random noise conditioned on the input states and the model is t
wever, learning bimanual coordination strategies and complex object relations from bimanual visual demonstrations, as well as
ggle to capture the intricate details of the lunar landscape. In response, we propose an innovative approach that leverages NeR
d deep learning-based techniques. View planning is an essential part of this task that dictates the information capture strategy
in contrast to controlled laboratory conditions, markers installed in factory grounds or equipment surfaces, often face challenge
gents. To achieve this spatial-temporal alignment, traditional methods depend on external devices to provide localization and clo
tly degrade the image quality and perfor- mance of many computer vision algorithms. To tackle these limitations, we propose a

eRF-based methods provide a promising solution by directly optimizing the pose from pixel loss between rendered and target im
n performance at a lower cost. However, action recognition in occlusion situations, where body parts are not visible, is still chall
e, unlabeled real-world data. However, the self-supervised methods often rely on the assumption of a static scene and their perf
tails in noisy scenes, and the risk of overexposure. Many existing single RGB image dehazing methods tend to perform well in
) and dynamic elements (e.g., vehicles, pedestrians). However, the high cost of annotation procedures of full-supervised metho
died AI.Unlike traditional closed-set annotation, open-vocabulary annotation is essential to achieve human-level cognition capab
e wake of rapid advancements in large model research, the foundation model for semantic segmentation tasks, termed the Seg
-imaging techniques to partially overcome this issue. This study proposes a novel feature finder that enhances vision-based re
lur. However, training the spike depth estimation network holds significant challenges in two aspects: sparse spatial information
to another. PillarGen can produce synthetic point clouds with enhanced density and quality based on the provided input point c
he demand of the product variety and short product lifecycle, the factory facilities including photogrammetry devices, should be
ion, tilt prediction, and spine segmentation, automated Cobb angle measurement has become increasingly popular. However, t
synthetically generated large-scale VMMR image dataset. Through the simulation of image data rather than the manual annota
nding local correspondence between the 2D image and 3D model, and then using the PnP and RANSAC methods to calculate
o generate ultrasound waves. However, the immobility of the PZT restricts the modulation of the acoustic field's position and or
. Our approach includes two main steps: finding Potential Grasp Areas (PGAs) from depth image of the bin and detecting suita
entation and pose estimation methods for 3D point clouds have been proposed. However, many of them require supervised lea
that allows us to infer vehicle actions using a shallow image-based classification model. The actions are classified via bird's-ey
ion and description algorithms to be robust to these conditions. While most classical methods achieve rotation-equivariant dete
surgical types and adapting to new surgical instruments/techniques. Therefore, continually updating the VQA system by a seq
d adhesion, the difficulty of manipulation exceeds the physiological limits of human perception and operation. Surgical robot is c
 a stereo camera in an eye-to-hand configuration for controlling the robot to reach target positions by directly mapping image s
ariety of foods without damaging them. To accomplish this, the use of underactuated hands is effective because they can adap
hile requiring only a few demonstrations. In this work, we present a framework that formulates the visual servoing task as graph
and orientations. However, the dynamic characteristics of conventional visual servoing controllers may be unsatisfactory, and
iggered Nonlinear Model Predictive Control (NMPC) for efficient and accurate tracking. The deformable contour model allows a
the object, binding the object to the lattice, and tracking and servoing the lattice instead of the object. This makes our approach
ime tracker to locate the object. This separate enhancement and tracking fails to build an end-to-end trainable vision system. T
ing reliability, and ensuring real-time responsiveness. This paper introduces a perception framework grounded in foundational m
ctiveness of traditional MOT methods is significantly reduced when it comes to dynamic platforms like drones. This decrease is
In this work, we seek to incorporate this understanding into interactive robots by proposing a set of assumptions and rules to re
minant paradigm for object tracking of specific object categories. Recently, large-scale pre-trained models have shown promisin
to learn the correlation between the template and search branches effectively with the sparse LIDAR point cloud data. Instead
elf-supervised methods hold the promise of training directly on real video, but typically perform worse. These include methods t
aches process a single measurement frame recursively with greedy association and are prone to errors in ambiguous associati
The majority of tracking methods, which follow the tracking-by-detection (TBD) paradigm, blindly trust the incoming detections
e-body optimization control. To achieve good perception complementarity between manipulation and localization, a new fast de
cy correspondingly increases, presenting a challenge for real-time robotics applications, especially on edge devices with compu
system based on ORB-SLAM3 to address the malfunction of the Loop Closure Detection (LCD) methods in this environment. I
es a single-track dynamics model for wheeled ground vehicles with visual inertial odometry (VIO). Our method calibrates and a
distance filter (DIDF). Although the sparse image alignment algorithm appears efficient for calculating frame-to-frame motion, i
evelop a novel GNSS-Visual-Inertial Navigation System (GVINS) that fuses visual, inertial, and raw GNSS measurements within
t. Firstly, we employ a sliding-window-based front-end to estimate real-time poses from the binocular fisheye images and IMU d
utation on the image plane – at individual pixels – enabling high frame rate image processing while consuming low power, m
time event-based visual-inertial odometry (VIO) that incorporates event, image, and inertial measurements. Our approach utiliz
he Finite Difference Method (FDM). Traditional FDM techniques often face computational overhead due to repeated perturbatio
ield-of-view from multi-camera systems, and the metric scale measurements provided by an inertial measurement unit (IMU). W
licability, we focused on a representative sports environment, the athletics track. Our efforts centered on improving the speed a
limb, Variable Grounding Flexible Limb (VGFL). The VGFL achieved a novel strategy in which its grounding point tracks the for
e, the neural delay dynamics are often overly simplified or ignored leading to over specified systems. In this letter, the neural de
uces a portable pneumatic knee exoskeleton that operates in both quasi-passive and active modes, where active mode is utiliz

owing the robot, safety issues may arise during the training. Due to the bias in different patients, an individualized scheme is al
ting heavy weights is a common cause of back pain, especially among laborers. However, current back exoskeletons are often
man's various motions. Joint torque analysis, essential for numerous applications in biomechanics and wearable robotics, nece
during locomotion. Research indicates that compared with naive wearers, knowledgeable wearers with abundant exoskeleton
ere designed to be affixed to both the shin and thigh. However, a common issue encountered in knee exoskeletons involves th
cycle using an actuated ankle foot orthosis (AAFO). The control strategy is based on an adaptive active disturbance rejection
on-based adaptation mechanism employing a fuzzy system is used to estimate the unknown time-varying parameters of the L1

r of the robot to leap onto a step using its legged form from its original form of wheeled driving, taking full advantage of the rapid
certain situations. This paper presents a velocity estimation method that combines inertia measurement unit (IMU) and encode
planar smooth curves, and apply it to sagittal-plane locomotion problems. We assumed that the body rolls without slip and main
es such as stairs. We take a considerable step towards closing this gap by presenting a fast trajectory optimizer for generating
eg-wheel design is considered and characterized in terms of key locomotion characteristics as a function of the design. Kinema
fy a planned trajectory to be safe or unsafe before the 4-WMR actually follows it, the conditions of the wheel forces (WFs: longi
and their combination allows for good balance-control stability at both high and low speeds. Four control inputs by two steerabl
trollers solving this problem. Unlike the prototypical method of transforming the model into a nonholonomic chained-form system
n open problem. It is further made more difficult when high horizontal velocities are involved. In this work, we propose an optim
sful loco-manipulation, especially involving interaction with heavy objects, is far from straightforward, as object manipulation can
need to recover if it is to continue its task. One critical subset of recovery is standing back up after falling down (aka “self-rig
ation is performed by analytically calculating joint angles that satisfy the desired support conditions, focusing on link lengths and
nging scenarios where no completely preplanned motion can accomplish the same task. This strategy exploits the concept of th
ethod employs a combination of Bernstein polynomials to encode the signed distance for each robot link with high accuracy and
derive a hierarchical control law compatible with singular and redundant tasks. Moreover, a smooth activation matrix is propos
y implicitly on forceful interaction to accomplish tasks but do not explicitly regulate it. This paper proposes a reinforcement learn
omising hands-free strategy to command humanoids but requires more physical and mental demand. To mitigate this limitation
 eature interaction across other complex domains. In this work, we explore and investigate the problem of domain-agnostic scen
ralization, resulting in degraded performance in target domain. In this paper, we present a novel approach to improve point clou
o recognize a specific target in a crowded environment. To address this problem, we propose a person re-identification module
ate-of-the-art re-ranking methods are inefficient when re-ranking many potential candidates due to their need for resource inten
ts with SG annotation. When learning from image-SG pairs, it captures structure-associated action knowledge for visual and tex
comprehensive contextual information and large model sizes. To overcome this challenge, we propose a lightweight paramete
round clutter. In this paper, in order to model the tracker more robustly, we first conduct an effective regularization analysis from
bject pose estimation methods are solving this issue using task-specific fine-tuned CNNs for deep template matching. This adap
e on considerable data for training. Most action recognition models using the transformer structure rely on a few frames from th
er, a novel variable transmission series elastic actuator (VTSEA) is developed to meet torque-speed requirements in different e
angement of springs for a 3 degrees of freedom robotic arm, with the aim of decreasing the electrical energy consumption for a
s adaptable joint stiffness to accommodate various tasks. Consequently, the development of variable stiffness joints (VSA) has
with linear stiffness profile based on symmetrical variable lever arm mechanism. The stiffness motor is used to regulate the po
n the basic system level, the Bi-Stiffness Actuation (BSA) concept was recently proposed. Theoretically, it allows for full link dec
onents, achieving a high performance-to-cost ratio. Extensive experimental data is presented, that compares two prototypes on
rated origami structure regulates the motion of actuator with high contraction ratio and enables accurate modeling. The origami
positioning. The actuator can also be disengaged to enable direct handling by an operator, for example for comanipulation. Th
otors require energy to provide force even if they supply no net mechanical energy. Springs are energetically conservative mec
ddresses the trajectory planning of a marsupial UGV-UAV with a non-taut tether.} The objective is to determine a synchronized
e unable to retransmit information such as video, which will lead to mission failure or affect the flight safety of UAVs. To addres
ner solves the trajectory planning problem by optimizing a cost function consisting of two indices: the total energy consumption
emented with multi-rotors emulating free flying or free floating conditions, exploiting the similarities and analogies between both
ot represented in the training data, these models often underperform due to domain shift. To cope with this fundamental problem
mics model, including aerodynamic equations. Using a global dynamics model enables the generation of aerobatics trajectories
ty and high-precision steady state performance simultaneously. In particular, the Koopman operator is leveraged to explore the

omputation and sensing, making fully autonomous flight extremely challenging. The first step towards full autonomy is reliable o
aper intro- duces an approach aimed at bridging the gap between simulated and real-world robotic sketching through the integr
he swing stoke. To address these issues, we propose a safe motion planning framework that fully pushes the robotic hardware
and their audio onset responses directly through audio-tactile exploration with a physical robot arm. Initially, the system refines
match the growth of robotics as a field and an academic discipline, spanning all levels of undergraduate and graduate experien
earning a model for these downwash patterns presents an attractive solution. In this paper, we present a novel learning-based
bances (e.g., a drone exposed to different winds) to collect training data. Our work is motivated by the observation that real-wor
t searching, and accurate localization. To address these challenges, this paper presents a hybrid system that incorporates a w
mitations, such as the robot cannot recognize unknown objects in real time and cannot navigate freely in a dynamic, narrow, and
is to generate a problem description (PD), a machine-readable file used by the planners to find a plan. By generating PDs from
ntly, NeuroMHE is trained through gradient descent, with its gradient computed recursively using a Kalman filter. This paper pr
. We approach our strategy, the multi-category decomposition editing network (MCDEN), from a similar viewpoint. Notably, we
ns challenging for these works to achieve broad adaptability across diverse articulated objects. Recently, many works have trie
us in offering the potential to enhance the generalizability as task-agnostic planners and facilitate flexible interaction between hu
 estimating gas features for critical tasks such as odor mapping, localization, and alarm generation. This paper introduces a hyb
pretability and out-of-distribution (OOD) generalizability. This work addresses these two challenging aspects by designing a var
elization process often results in the loss of point-specific features. This gives rise to a challenge in recovering those features fo
that leverages subequivariant principles from physics to enhance coordination of motion control with reinforcement learning. Th
vability because of sensor limitations, environmental conditions, or limited inference due to bounded computational resources.
ties for tackling complex and unknown tasks. To overcome these issues, we present a novel approach that combines adversar
ing in complex environments. While imitation learning of expert policies has been well-explored, the study of learning expert rew
very high-dimensional. In this work, we present an online method for detecting distribution shift with guarantees on the false po
putational burden on the agent. Moreover, in many scenarios, e.g., Information gathering, it is required to introduce a belief-dep
e as conformal policy learning. We show how to design such policies by using conformal quantiles to switch between base poli
such as 'particle deprivation' which hinders accurate representation of the true posterior distribution. This paper introduces a no
esearch in this area has reached a state of stagnation. Adversarial attacks focus on exploiting vulnerabilities in models, while co
-modules to run on low-cost telepresence robots in order to support more advanced and autonomous assistance services. The
robot to autonomously and safely learn food-cutting tasks. More specifically, our goal is to enable a collaborative robot or indus
thods that rely on either known goal priors or zero-shot large models, SG-Bot exemplifies lightweight, real-time, and user-contro
feedback, placing new importance on the token budget for the context window. An under-explored but natural next direction is t
literature that uses the large language model to enrich the text descriptions, the latter remains underexplored. In this work, we
Foosball environment. Initial findings are shared using a simple baseline approach. Foosball constitutes a versatile learning en
arallel systems of fast and slow thinking in human decision-making—we introduce textit{Robotics with Fast and Slow Thinking (
e where the LLM is provided few-shot examples of robot code. This presents a challenge for LLMs to implicitly infer task informa
btasks and defining mathematical interfaces between them, the framework allows for the independent training and testing of the
y to follow the reference path while avoiding perceived obstacles. To account for unforeseen or dynamic obstacles not present
dentified as the same. Existing work has attempted to introduce object-level features or artificial landmarks. The former approac
ted chairs for a teddy bear to sit on. The teddy bear is a proxy for an elderly person, hospital patient, or child. By autonomously
. Such human involvement is an important bottleneck in scaling up robot learning across diverse tasks and environments. We p
ch conditions, LLMs may overlook(unfaithful) the rules in the given context, resulting in the generated plans being invalid or eve
plexity of black-box models. At the same time, the evolution of deep learning introduces larger, multimodal foundational models
ks. However, existing approaches either translate the natural language directly into robot trajectories or factor the inference proc
igation, pick, place), (2) a skill coordination policy that chooses which skill to use when, and (3) a corrective policy that adapts p
mation processing and access. Deciding what information to retain or discard within a robot’s LTM is particularly challenging
ng. In this paper, we present a novel tele-immersive framework that facilitates cognitive and physical collaboration between hum
orcement. By expressing regulations through a specification language and integrating their enforcement into the compilation pr
nsure dynamic consistency in the generated trajectory. The dynamics of these platforms introduce a high nonlinearity, leading t
ectories is adopted for the optimizer, and the underlying state estimator is updated with appropriate models of sensors and qua
al use. In particular, for unbounded paths, the gains will inevitably become ill-suited for efficient path following. We propose a m
ake it difficult for UAVs to approach the landing area, and common solutions cannot be used for automatic landing. A tethered
with a wider field of view is very beneficial. Despite the usefulness of fisheye camera for UAV applications, not much work has
our method does not rely on detecting individual neighbors by each agent and estimating their relative position, but rather we p
ss and limited utilization of previous samples often result in sub-optimal exploration efficiency. Additionally, most of these metho
range of the scale and is able to extract the units used. We show that our gauge reading algorithm is able to extract readings w
gry. In prior work, we have demonstrated that topological tools based on Morse Graphs (directed acyclic graphs that combinato
eed MPC solver with a low memory footprint targeting the microcontrollers common on small robots. Our approach is based on
p. The microfluidic chip with a pipette tip is directly fabricated using 3D printing. Xenopus oocyte was used in the experiment. W
earning robot control (CLRC) is an innovative methodology that guarantees parameter convergence under a condition of interva
ehaviors and seamlessly inserts the virtual events into images perceived from the physical world in real-time. The manipulated i
ement learning approach that generates scenarios by sequential editing, such as adding new agents or modifying the trajectorie
on steering labels recorded by an expert driver but effectively leverage a classical controller as an offline label generation tool. I
edge of where potential collisions may occur and how. Therefore, we propose to gain this information by analyzing locations an
proven challenging due to the need to balance realism and diversity. Towards this end, in this work we develop a framework tha
en made to teach the robot the "dos", the "don'ts" received relatively less attention. We argue that, for any practical usage, it is a
go-vehicle aggregates observations from all neighboring connected cooperative vehicles (CCV), without considering the interac
os. The framework consists of two key stages: self-supervised learning (SSL) and feature distillation. POP first employs SLL to
ory timesteps, as future information is not available and involves high uncertainty. However, without sufficient guidance for capt
h a diverse set of human crossing maneuvers, producing a wealth of new scenarios and addressing the scarcity of rare case d
is difficult to achieve, partly due to the lack of a generic framework for transferring part of the known knowledge (behaviors) be
coordination mechanism and consists of two iterative phases, i.e., the communication phase and the assignment phase. In the
ankel structure in the input and consequently also on the output. Taking advantage of this, a convolutional vision transformer m
odels (LLMs) for this task by leveraging the LLM’s commonsense-reasoning capabilities for making sequential navigationa
re to go. Even then, prioritizing the next target based on previous knowledge is complex. In our proposed work, the robot utilize
es ChatGPT to reason the descriptive steps of tasks. In order to enable ChatGPT to understand the environment, semantic map
s only consider static hypergraphs, and ignore that in a hypergraph, the power of influence varies between vertices (or hypered
anding in driving situations. We also present a new dataset of 160k QA pairs derived from 10k driving scenarios, paired with hig
ir stochastic and multi-modal natural behavior tendencies. However, existing approaches have limitations as they either focus s
m. For this purpose, we propose a novel three-part radar filtering method for outlier detection: By analyzing uncertainties and sy
ace significant challenges in environments with a high density of agents due to their dependence on active emissions. Cameras
and a standardized closed-loop benchmark for equitable comparisons. Utilizing this platform, we conduct a comprehensive stud
nd left-turning ways in large intersections. However, due to the complexity of intersections, producing such a map by human res
ons of the partially observable environment. However, prior work on occlusion inference and trajectory prediction have develop
ible to changes in lighting, season, perspective, and appearance. Additionally, the large storage size of maps with descriptors a
while also capturing different modes of behaviour (such as turning left vs continuing straight at a junction). To this end, we pres
cy And Risk-aware Contingency planning (MARC), that systematically addresses these challenges by enhancing the multipolicy
for different traffic scenarios. This paper presents a graph-based scenario-adaptive lane-changing trajectory planning approach
part of the terrain (e.g., irregular boulders and fallen trees) will be treated as non-traversable obstacles. To improve wheeled m
the many complex factors of social compliance make geometric navigation systems hard to adapt to social situations, where n
tiple targets in complex environments. In this study, the interaction mechanism of target tracking between the robot, the environ
es and driving scenarios, Talk2BEV eliminates the need for BEV-specific training, relying instead on well-performing pre-trained
nsformer (DualAT) to enhance the multimodal fusion and waypoint prediction processes. A self-attention mechanism captures g
of-distribution (OOD) issue and potential robot localization failures in a given map for unseen observation. This significantly drop
motion trajectories. By predicting the distributions of all these variables from a shared global scenario embedding, we ensure th
er-resolution. However, most of the works on LiDAR super-resolution have the drawback that they first transform 3D LiDAR poin
n-based multi-frame point cloud fusion module, Stability-Attention Module (SAM), to encode the spatial features from point cloud
d doors that create opportunities for sudden encounters. In this work, we present a Transformer based architecture to predict h
nsic uncertainties, and hence, accommodating diverse human capabilities is significant for its practical applications. This work p
based framework that enhances offline RL to overcome the long-horizon vehicle planning challenge. Specifically, we design a v
rrain labels hinders the emergence of end-to-end multi-directional traversability network. This paper introduces a framework for
bjects due to the need for specific object-end modifications, such as applying perceptual markers like color-labeled plates and r
k which investigates the gas-source search and robot navigation tasks independently, leading to a decoupled approach that res
presentation effectively reduces the impact of accumulated drifts commonly encountered in navigation tasks. A Reinforcement L
n real-world environments. The robots collect task data and upload it to the cloud where navigation policies can be learned eith
e the quality of the motion forecasting models to the performance of computer vision models. Moreover, the human-designed e
nt of industrial robots under consideration of kinematic, elastic, and thermal effects is presented. A continual learning algorithm
pabilities of the learned policies and the advantages over well-tuned baselines remain unclear from prior studies. In this work, w
t requires an efficient framework capable of generalization but within a reasonable data budget. In this paper, we develop an ef
nd an image of the resulting arrangement rendered. These renders are evaluated by a VLM, so that the arrangement which bes
, allowing a robot to learn a new task with little human effort by leveraging data and models from the Internet. However, reinforc
ining model ensemble and distillation techniques. The use of ensembles of deep neural networks has been shown to improve g
ent software analysis techniques identify such mismatches using dimensional analysis rules and ROS-specific assumptions to a
we designed a system with two parts, including a flapping-wing aircraft with a wingspan of 1.8m and a take-off weight of 870g,
eristics of its main constituent materials. The design decision process and locomotion principle of the flipper are introduced in d
cal tail to make maneuvers out of the longitudinal plane. This work defines simplified equations for the steady turning maneuver
quite cumbersome and complicated to model and control different motions with multiple principles. In this paper, we proposed a
d robot on synthetic mud with precisely-controlled ratios of sand, clay, and water. We observed a non-monotonic dependence
est in fields such as biology, physics, and engineering, and multiple hypotheses for how such benefits may arise have bee
es precise control. In this study, we installed two low-cost inertial measurement unit (IMU) sensors separately on the head and
ssue, we develop a robotic fish with a tunable stiffness caudal peduncle (TSCP), simultaneously allowing for untethered swimm
ation, and successful completion of escape tasks in changing environments. In this paper, a novel fish-inspired self-adaptive ap
sing, the agent learns to coordinate rhythmic behavior among different oscillators to track velocity commands, while at the same
ations is numerically feasible, and introduce the relaxed condition of form boundedness to achieve robustness in the presence o
ed for additional sensory feedback. We designed the shell by specifying the cross-sectional geometry in the yz and xy planes su
deep learning-based methods involving millions of timesteps of simulation interactions. Notably, such offline-designed locomotio
id acceleration and turning motions. Subsequent research suggests that the effectiveness of the cheetah’s tail might be en
four-degree-of-freedom spine design. Using trajectory based reinforcement learning techniques, STRAY is able to trot and bou
onditions. The robot draws inspiration from the joint structures of quadruped animals and incorporates biomimetic features. At t
gile locomotion and the ability to perform continuous jumping. To achieve this, a prototype soft robot (weight of 300 mg, size of
mal techniques for vertical self-burrowing. In this paper, we realize a snake-like bio-inspired robot that is equipped with an IMU a
nstrated the untethered takeoff of a flapping-wing MAV powered by a commercially available LiPo battery in short duration with
movement of a locust, we make full use of the Stephenson six-bar mechanism in designing a jumping leg to achieve power amp
can crawl and jump. The microrobot has a mass of 980 mg and stands 13 mm tall, and has on-board sensing, computation, and
strategy referred to as interfacial flight. Here, we present an insect-scale robot, the gamma-bot, inspired by these interfacial flie
on phenomena generated by fluid-structure interaction (FSI) during swimming. The microactuator is composed of a carbon-fiber
ney bees and wasps, we propose a robot navigation method that directly learns the home vector direction from visual percepts
dragonflies on man-made flapping wing systems, we designed, optimized and fabricated a dragonfly-inspired flapping wing robo
d by insufficient thrust generation, manufacturing challenges, fragility, and cost. This work presents the design and developmen
 nstable sensory-motor mapping due to the random wiring of neurons seeded on multi-electrode arrays (MEAs). Therefore, her
lenge, we propose a continuum robot with a novel extension mechanism inspired by the impressive ability of woodpeckers to e
dal hybrid robot with the ability to actively adjust the wheel's size, which is inspired by the behavior of the turtle's legs when it en
ation of compliant mechanisms into sidewinding limbless robots remains less explored, posing challenges for navigation in com
the reaction wheel circumvents this limitation, its efficiency has been shown lower than the tail in terms of inducing Moment of
tempts and endeavors to make things through self-assembly of parts, analogous to the autonomous and decentralized develop
mechanisms and control strategies. Nevertheless, the cable routing method has not changed, and CDRMs continue to suffer fr
en by super-coiled polymer (SCP) artificial muscles. While our previous work designed and simulated a deep deterministic polic
t learning-based approaches using solely pose regression are limited in their abilities to diagnose inaccuracies. In this work, we
methods, while achieving high accuracy and good adaptability, rely on large annotated dataset and need additional training. We
principles of stereo rectification. In this paper, we first delve deeply into the concept of rectifying homography, which serves as t
alibration monitoring provides a certificate for reference-calibration parameters validity. Tracking follows the calibration paramet
ptimizing spatial coverage with as few cameras as possible. The extrinsic calibration of these systems introduces additional com
se methods simplify the task of extracting corresponding features between the image and lidar point cloud. But many of these a
rameters between the sensors through point cloud surface features so that the extrinsic can be narrowed from a large initial er
that estimates robust and accurate extrinsic parameters between camera and LiDAR. This is a novel work to use semantic info
mate the intrinsic and extrinsic parameters of the cameras. The same idea can be applied to event-based cameras, though it re
e training data, essentially tying the network to a specific data distribution. However, if the range of evaluation data differs from
borative robots remains a challenging problem due to the significant non-linear properties of their unique gearbox, the harmonic
se properties based on observations of scenes with different sets of interacting objects. We propose an approach that exploits
utilize convolutional networks to automatically extract features from LiDAR depth images and RGB images for calibration. Howe
which can produce low accuracy and repeatability. However, the derived transformation should solely depend on the geometry o
r of mass (COM) of unknown objects without using force/torque sensors at the end effector or on the joints. In our method, a ro
s different data modalities. Previously, many calibration methods involved specific targets and/or manual intervention, which has
DAR, they can provide more comprehensive information about the surrounding environment. Nonetheless, due to the distinctiv
researched in V2X, which is the key to fuse multi-radars. The main reasons are three-folds: (1) New sensor. Thus, it is not surp
In this paper, we propose a fully-learned system that allows bipedal robots to react to local terrain while maintaining command
generalization. In this paper, we model perception failures as invisible obstacles and pits, and train a reinforcement learning (R
ch is inspired by human reasoning and designed to navigate towards unseen semantic objects in novel environments. VLFM bu
, for example by using subgoal proposals, planning, or separate navigation strategies. In this paper, we describe how we can tr
ensional data, the inherent reflectional and rotational symmetry of robot structures has not been adequately explored. Drawing
ks. CollisionGP introduces a Pòlya-Gamma auxiliary variable for each data point in the training set to allow classification infere
nformation on the pedestrian’s behavior, but may deviate from the data, which can lead to collisions during planning. In this
cle avoidance. Multimodal Motion Prediction (MMP) is achieved by a deep-learning method that predicts multiple plausible futur
and the strong duality of convex optimization. Under this constraint, the high-dimensional non-convex optimization problems of
ernative is to directly map raw sensor observations into robot actions. However, acquiring skills from high-dimensional raw obs
specific collision scenarios. This has led to learning-based methods using neural networks for adaptive collision avoidance. Ho
gions in the search space and de-emphasizing regions that may not lead to good solutions for MGPF. Our approach provides a
structure unifies the planning with the terrain identification. Besides, it contributes to explicitly identifying the hazardous areas o
mobile robots, and a training/inference pipeline employing CNNs. We validate our approach in several mazes, as well as in real-
n no obstacles.In these domains, simple linear trajectories are guaranteed to be collision free, and can be leveraged for further
 r and angular velocities, which leads to unstable speeds and unsmooth trajectories of the robot during the long-term execution.
mismatch of representation for prediction and planning. In prediction, numerical tractability is usually achieved by coarsely discre
methods in trajectory optimization. Based on Sums-of-Squares optimization, our method provides exact, rigorous certificates of
DIC’s decoder utilizes feature extraction networks to extract features from both the primary image and the side information.
t leverages a generative model to decompose a complex planning problem into small manageable ones by incrementally outpu
model that returns a differentiable collision cost based on robot pose, obstacle map, and robot footprint. The differentiability of o
omputationally cheap to evaluate, and (2) a collision-cost formulation which provably preserves forward invariance (i.e., keeps t
ers of MIMs can be used to simultaneously estimate obstacles' height, solidity/density, and opacity. We demonstrate that MIMs'
er. Our approach begins with the generation of a high-resolution SGP local map, providing an interpolated representation of the
s context.Moreover, we show that a robot is able to navigate in a known environment with only proprioceptive sensors by active
hieves high accuracy using edge information but incurs higher message-passing costs with growing graph size, limiting its appli
ngle path planning algorithms, such as Theta*, Visibility Graph, and RRT. To address this issue, we propose an efficient LOS c
ons and failures to reach the goal. The system could be safer if the pre-trained RL policy was uncertainty-informed. For that rea
ectories' success probabilities based on partially reliable multi-modal sensor observations. We encode high-dimensional multi-m
pace, particularly those with angle limits. This method replaces line segments with ellipsoids to connect adjacent nodes. Firstly,
n exact one for the environment. Additionally, online learning is needed for the robot to maneuver in an unknown environment w
lation. However, current VLMs are limited in their understanding of the physical concepts (e.g., material, fragility) of common ob
ative paradigm, consisting of (1) retraining the NeRF when new images arrive; and (2) planning a path to the next best view on
m based on explicit representations, consisting of (1) planning a path to the next-best view only; and (2) requiring a considerabl
ct queries based on current sensor inputs is a common practice. For this, existing methods strongly rely on LiDAR data howeve
ories. However, this approach falls short of effectively addressing the diversity present in the real world. To tackle this issue, we
e argue that the residual maps provide greater potential for motion information, and on the other hand, range images contain ric
ues, these methods have demonstrated promising performance. However, inaccurate prior estimation, which is usually inevitab
n datasets and algorithms for transparent objects, we can at least obtain noisy estimates of normals and masks for various rea
ct crumpled cloths from their top-view depth observations only, with our proposed sim-real registration protocols. In contrast to p
efine amodal optical flow as a multi-layered pixel-level motion field that encompasses both visible and occluded regions of the

approach, and a geometric/semantic feature based upsampling strategy. The novelty of this work is threefold: (1) developing a
we present a novel transformer-based framework named CVFormer, which leverages two-dimensional circum-views from the e
on volumes scale poorly with resolution, prohibiting them from estimating high-resolution flow. We observe that the spatiotempo
and sensors. In this work, we aim to address scalable camera-based collaborative perception with a Transformer-based archite
-based LiDAR data generation using deep generative models, they still struggle with fidelity and training stability. In this work, w
sions within split seconds to ensure safety through numerous computer vision tasks. In this paper, we present a new real-time
osed LiDARFormer utilizes cross-space global contextual feature information and exploits cross-task synergy to boost the perfo
or intrinsic contact sensing, object perception, and interactive manipulation, is proposed for these adaptive grippers. The contri
uotactile sensing. However, existing methods for rotation estimation did not consider the impact of the incipient slip during the p
nd surface reconstruction by harnessing the robotic manipulator's kinematic data and a compliant tactile sensing module, which
bility to handle tasks with varying geometry. To reduce the need for a priori models, we propose a framework for estimating con
acShade, building upon the strengths of the TacTip, a single-camera tactile sensor of large in-depth deformation and being sen
such predictions. Similar to robotic grasping and manipulation scenarios, we focus on modeling the dynamics between a rigid m
o avoid consuming repositioning actions. We prove that a line contact, rather than a single point contact, is necessary for nonho
 general formulation and solution method for the case of an arbitrary smooth object in a single-point rolling contact on an arbitra
efficient of friction from the literature, even though actual friction behavior is variable and depends on many factors. In this work
hand, Task and Motion Planning (TAMP) approaches excel at solving and generalizing across long-horizon tasks, thanks to the
n). Several works for TTA have shown promising adaptation performances in continuously changing environments. However, o
ary environment without forgetting. We introduce a novel approach to lifelong learning, which is streaming (observes each traini
arch direction. These techniques replay data from previous tasks to mitigate forgetting when learning new skills. However, existi
ns and suffer from catastrophic forgetting. Hybrid planning and RL approaches have shown some promise in handling environm
crafted prompt engineering, limiting the agent to a stationary range of addressable tasks. In this work, we introduce LRLL, an LL
a method for using LLM-based planners to query new skills and teach robots these skills in a data and time-efficient manner for
ers from catastrophic forgetting. To solve this problem, we introduce a novel continual learning (CL) framework called the Prob
a small number of corresponding task demonstrations. LOTUS starts with a continual skill discovery process using an open-vo
bances. In this paper, we propose and design a re-configurable aerial platform consisting of two modular bi-copters with an und
on, we leverage reinforcement learning (RL) algorithms for the automatic discovery of control policies. In the face of the massiv
significantly impacts the primary performance metrics acquisition cost, cycle time, and energy efficiency. However, identifying a
o the building structure, ensuring durability and elegance. In this context, existing robotic cleaning methods are limited by their c
ting its motion.In this paper, we apply the computationally efficient position-based dynamics method to predict object motion an
obots to carefully select information to be transmitted. In this paper, we consider the case where environmental information is m
er, it could be costly to equip all robots with expensive perception or SLAM systems. In this paper, we propose a novel system
m is proposed based on a Gibbs random field (GRF), where bio-inspired potential energies are used to charaterize ``robot-robo
rgeons, the inherent hand tremors during manual surgery can jeopardize the safety of these critical interventions. This has fost
ets featuring annotations in the robot space, including people's 2D or 3D positions. To tackle this issue, we introduce PanNote,
lip or external disturbances. However, there is a lack of datasets and evaluation methods that consider unpreventable failures c
ance and, consequently, large variations in image appearance. Compared to the other publicly available datasets, our dataset
xisting benchmarks lack outdoor point cloud scenes with semantic labels. To address this, we introduce PointSSC, the first coo
se issues, we introduce an innovative physically-based structured light simulation system, generating both RGB and physically r
so between objects themselves. However, current scene-graph datasets predominantly focus on human-object interactions (HO
ection protocol for reproducibility and the evaluation of the precision or accuracy of the recorded trajectories. With this objective
and develop a Range-aware RGB-D data Simulation pipeline (RaSim). In particular, high-fidelity depth data is generated by imi
heir dynamic scene understanding algorithms. Specifically, the THUD dataset construction is first detailed, including organizatio
and instead train marginal intent prediction models independent of robot actions. This is because training conditional models is h
and benchmark. The benchmark is designed to test the ability of ML-based planners to handle diverse driving situations and to
a portable data collection system, coupled with a semi-autonomous labeling pipeline. As part of the pipeline, we designed a labe
g any user requests within the entirety of 3 office buildings and using multiple embodiments (robot, human, human with graspin
e situations. However, the generation of synthetic, yet realistic, corner cases poses a significant challenge. In this work, we intro
ature is attractive for enabling robots to acquire new skills and improve their manipulative ability. However, due to limitations in
m data. In this paper, our goal is to develop methods for training policies for socially unobtrusive behavior, such that robots can
s. While significant advances have been made in the community, the current evaluation of different risk identification algorithms
and to generalize to unseen data. Therefore, we equipped our research vessel with a laser scanner and a stereo camera to rec
ed with heterogeneous road occupants. Existing datasets, predominantly geared towards structured and sparse traffic scenario
whose domains usually pertain to the texture information present in them, the features derived from a 3D point cloud are affect
al challenge is the lack of available large-scale datasets due to the unbearable cost of labelled data collection. To overcome thi
e to limited data availability for training and testing. To address this gap, we present the German Outdoor and Offroad Dataset (
ered bin filled with objects. Also, for the next-level automation, they need to handle unseen objects and discriminate target obje
introduces the NYC-Indoor-VPR dataset, a unique and rich collection of over 36,000 images compiled from 13 distinct crowded
and diameter estimation of trees in forests and orchards. We present TreeScope v1.0, the first robotics dataset for precision ag
e surfaces like glass planes, which may degrade the performance of various robot applications. In this letter, we propose a nove
compounded by privacy concerns preventing dataset sharing. To address these challenges, we propose a novel source-free do
The cross-modality feature correlation filtering module extracts geometric transformation-invariant features from cross-modality
and state estimation of articulated objects from a single point cloud. The proposed CAPT methods accurately estimate joint par
approach for detecting adverse weather effects in LiDAR data. We reformulate this problem as an outlier detection task and use
a new detection loss UnfoldSoftmax to improve inference speed. Futhermore, we propose Ortho-Alignment loss combined with
ulty focusing its attention weights on sparse and meaningful points, leading to limited local structure modeling capabilities and q
argue that predicting the action target's 3D coordinate could pave the way for more versatile downstream robotics tasks, espec
ces catastrophic forgetting and error accumulation problems, which impede the practical implementation of CTTA in the real wo
3D shape data obtained by visuo-tactile sensors. Data from noncontact locations can interfere with slip judgements and the ign
nowledge, such as ``if the human right hand is predicted to be in contact with an object after 1 second, the distance between the
e self-modeling is a necessary feature of truly autonomous agents. In this work, we leverage neural fields to allow a robot to sel
earning are limited to a predefined set of affordances, thus limiting the adaptability of robots in real-world environments. In this p
emphasizes cluttered environments that contain a wide range of objects with different scales. It enhances plane segmentation
lacing artificial markers (e.g. ArUco) on the robot’s body is not possible or practical in runtime. Prior methods require accur
augment the learning in the ground-based imaging domain. However, severe challenges in the austere UAV-based domain re
hes often suffer from poor sample efficiency or limited generalization capabilities, making them unsuitable for mobile robotics ap
hile convection also leads to fluctuations in air density and hence causes so-called schlieren – intensity and color variations in
their versatility comes with limited onboard resources, i.e., sensors, processing units, and memory, which limits the complexity
ce segmentation, we study semi-supervised 3D instance segmentation problem and propose a novel end-to-end framework bas
dometry poses and automatic panoptic predictions with inconsistent IDs between frames. Despite this noisy input, our system is
t object boundaries. To address these issues, in this work, we propose GAM-Depth, developed upon two novel components: gr
sionality. Previous methods either regress 3D keypoints directly or utilise a render&compare strategy. These approaches often
ets. The sparse nature of LiDAR point cloud data prevents an efficient manual annotation process. To solve this problem, we pr
and describe future environmental structures by the point cloud form. In this letter, we propose a novel efficient Transformer-ba
configuration space and to navigate dynamic and unknown environments. In the previous work, we proposed to decompose mo
architecture, trained to predict the future time steps of SOGMs, given 3D lidar frames as input. Our pipeline is entirely self-sup
e invariance. In many tasks and settings, this assumption does not hold as objects in the scene are often occluded or lie outside
rediction models simplify multimodal prediction by first predicting 2D goal locations of agents and then predicting trajectories co
ss these issues, studies have been conducted to estimate external torques using only internal signals, such as joint states and c
neralist policies by merging together multiple, task-specific, individually trained policies. In this work, we take a preliminary step
t constraints enforced on predicted trajectories. We present a dynamics-based deep learning framework with a novel asymptot
rtially observable settings, the team reward is misleading due to unpredictable factors from the behavior and state of unobserve
in which large language models may benefit robotics planning. In particular, we propose Statler, a framework in which large lan
exploits context features in different granularities for different kinds of traffic agents. To further enhance MGTR’s capabiliti
n strategies based on minimizing predictive entropy and variance of probabilistic models. Through ablation studies and human e
neural signals indirectly, through neurotransmitters. Prior work showed that interpretable dynamics for complex robotic control
e a novel strategy by incorporating knowledge distillation and model pruning to overcome the inherent trade-off between speed
struction approach is desired. On the one hand, traditional explicit 3D reconstruction methods such as SLAM and SFM require
sequently, we employ a simulator to generate images by uniformly sampling the surface of objects from the YCB Model Set. Th
act with humans in social scenarios despite not observing any robot in the motion training. We develop a transformer-based arc
of point cloud data poses significant challenges for storage and transmission. Moreover, a considerable portion of the data lose
f a learned, generative Inverse Kinematics solver, which is able to efficiently produce promising entire candidate solution traject
ased on reinforcement learning (RL) offers versatility and generalizability but presents safety and robustness challenges. We pr
d computation times. We present a new approach to compute the entire shape of a tendon-driven continuum robot by a physics
nodynamic planning on the constraint manifold is indispensable. In recent years, learning-based solutions have become alterna
s to work together on rearrangement tasks. This paper proposes a comprehensive learning-based framework for multi-agent ob
al-world systems. We evaluate existing algorithms and their novel variants across multiple robotics control environments, encom
-based approach for online learning of grasp predictions for robotic bin picking. Specifically, the online learning algorithm with a
ations, among other factors. In this paper, we introduce introduce a novel approach, SSL-ConvSAC, that combines semi-superv
s exhibit significant variability in accuracy and robustness, particularly in 2D keypoint detection. In this work, we present an end
ed task to a learned manifold that contains only positive task executions and outcomes. When reconstructed through this manif
of being unable to recognize the object category being grasped, which could lead to unpredictable or risky behaviors. To overc
multi-level framework for part assembly planning consisting of part assembly sequence inference, part motion planning, and rob
ation. To address this gap, we propose a framework to automatically design 3D printable adaptations that can be attached to h
nd orient the peg and the hole geometry in SE(3) space for successful assembly poses significant challenges. Addressing this,
gid objects remains an unexplored area of research. Such setups are common in various scenarios that involve the transportat
rformance of vision-based Reinforcement Learning (RL) algorithms for robotic manipulation tasks. Our proposed method involv
of policy learning across a variety of robotic manipulation domains. This paper explores an application of symmetric policy learn
which leverage transformer relational dynamics to encode the history of trajectories given partial-view point clouds and an objec
e to high occlusion in the clutter, elusive entanglement phenomena, and the need for skilled manipulation planning. In this pape
h is still limited by the large sim-to-real gap of deformable objects or rely on large datasets. A promising alternative is to learn fa
ective grasp for novel objects, we propose a new autonomous grasp control scheme termed HAGrasp that considers the compl
the push-grasping synergy for robotic manipulation in such environments, aiming to significantly reduce pre-grasping redundan
ge engineering process or by manually manipulating the object into the desired shape and capturing the goal shape at that spec
ment of the camera. Empirically, generalization to some of these factors have presented a greater obstacle than others, but exi
sparent objects due to depth cameras’ deficiency in sensing their geometry, while we proposed a novel framework to recon
les practical use. In this work we explore dense tracking as a representational vehicle to allow faster and more general learning
ve this problem by learning a sequence of actions that utilize the environment to change the object's pose. Concretely, we emp
nipulation task requiring a sequence of accurate actions to achieve the target shapes using a robot arm with an RGB-D camera
articulated objects with specific joint types. They can either estimate the joint parameters or distinguish suitable grasp poses to
only consider a single ground-truth grasp orientation at a grasp location during training and therefore can only predict limited gra
nd replacing the desired goal with one of the achieved states so that any failed trajectory can be utilized as a contribution to lea
network that uses an RGB-D active stereo camera. ASGrasp utilizes a two-layer learning-based stereo network for the purpos
n stage, where the tentative decision stage determines a tentative action from the original camera image,and the correction stag
ation, but each of these modalities have their own limitations. Commercial point-cloud observations often suffer from issues like
s in real-world scenarios. In this paper, we propose a novel part-guided 3D RL framework, which can learn to manipulate articu
owever, such a model is often difficult to integrate into an effective optimization routine. To address this, we introduce a proxy h
data. In this paper we demonstrate how the increased understanding of off-policy learning methods and their embedding in an i
s with the environment can mitigate this problem and propose an information-based measure to augment the agent's objective a
bing to a tradeoff between gripper complexity and generalizability. Adapting such models pre-trained on one type of gripper to a
ormance. In this work, we show that the widely used heatmaps are underestimated in the efficiency of 6-Dof grasp generation. T
network based neural controller (HPN-NC). To achieve this, HPN-NC consists of a hyper net and a low-level controller, where t
ate the sim-to-real gap with a combination of domain randomization, image augmentation, and segmentation tools. Our method
bject trajectory estimation for collision avoidance or estimating locations with the least odometry drift. In this work, we present A
fers from two critical limitations: a) learning the discriminative features for the unlabeled data; b) learning both global and local i
he poor contrast between cracks and their surroundings. However, most deep learning methods are designed for well-illuminate
e proposed binary point cloud Transformer network significantly reduces the storage footprint and computational resource requi
ch greatly limits the development of mmWave radar technology. In this paper, we propose a novel point cloud super-resolution
ce of extracted information, like the object detection models. In this paper, we propose to describe images by nouns and modifie
ng robust, view-invariant keypoints in a self-supervised approach requires a meticulous data collection approach involving preci
econstruction system which achieves fine scene reconstruction without prior information by utilizing neural implicit encoding. O
al workaround, but often requires extensive manual tuning and results in models that are brittle to distribution shift between sim
correct errors and adapt to changing situations. In this paper, we present Tactile Adaptation from Visual Incentives (TAVI), a ne
l anomalies. However, this task is challenging due to the small sizes of fixtures, the requirement for real-time performances, an
ntrained, densely cluttered environments. In particular, a dual RL model is introduced, which presents high resilience in handlin
Controller (DDMPC) to compute the motion command given the previous interaction experience and feature deviations in tactile
ement learning. The robot's workflow involves detecting the pose of the target object and the basket at each time step, predictin
anipulation is the inverse Jacobian, which can directly map from the desired in-hand object motions to the required hand actuat
ing tactile feedback. In this article, we make an attempt to explore how to pre-train a representation model for robotic manipula
e to reason about the stability of placement from very gentle contact interactions. Our results demonstrate that it is possible to i
es of controlling their contact-based interactions with objects, which are the final targets of manipulation. This paper investigates
to traditional methods, it achieves high precision and stable control even in environments with limited observation signals and h
ging contact state during interaction. In this paper, we extend the skill formalism by incorporating a behavior tree-based primitiv
thropomorphic hand is under-explored. We present DexDLO, a model-free framework that learns dexterous dynamic manipulat
In particular, we focus on minimizing the size of components to get proper performance without sacrificing compactness. A rob
wever, there is no guarantee that the relative position will be maintained in regions subject to joint limits and task-space reachab
atures designed by decoupling global deformation into low-frequency modes. Although modal analysis is widely adopted in com
discontinuous gradients, and (iii) the non-convexity of the planning problem. The stochastic nature of RL addresses (i) and (ii)
veraging contacts. It generates rigid-body motions and complex contact sequences. Our framework is based on Monte-Carlo T
er, enabling precise control and improved grasp success rates, particularly for small objects ranging from 5 mm to 25 mm. Our i
s or integrate these UAV modules into a whole. For practical applications, the R-UAV requires collaborative decision-making for
ocation-independent connections between robots and latency-sensitive robotic services. FogROS2-LS offloads conventional on
ifically, our simulated agents are connected in series by un-actuated, flexible, fixed-length tethers and use tether angle and stra
d DDP (ND-DDP), is a three-level architecture which employs ADMM for consensus, an augmented Lagrangian layer for local c
dered. The tasks include exploration, capture of dynamic targets, and perimeter defense over valuable resources. Since each r
ative, decentralized Voronoi-based coverage control approach to reactively navigate these complexities while guaranteeing safe
d; therefore, the computational effort, communication bandwidth, and memory for each agent scale like o(1) with respect to the
allocation process. The performance of the algorithms for multi-UAV collaborative visit missions is compared under pre-allocatio
ctored propulsion, combined with a model predictive control method, to autonomously achieve more agile and efficient moveme
evel control. In this paper, a constrained model predictive control (MPC) is proposed and applied to solve the attitude control pr
promising avenue for addressing this challenge by incrementally constructing random trees, thereby sidestepping the need to e
ther by also considering temporal robustness as a critical metric to quantify the tolerance of time uncertainty in STL. To this end
ficiencies. However, thanks to novel physics-informed DL architectures, such as Deep Lagrangian Networks (DeLaN) [1], it is p
del of these systems, where the soft links are modelled by Cosserat rods parameterised by assumed strain modes. This choice
+m^3) solver by Popov and Vereshchagin (PV), where n,m and d are the number of joints, number of constraints and the kinem
ne the interface variables between the communication points is essential for efficient and stable performance, especially for rea
models or static weightings assigned to obstacles, semantics, and other perceptual cues to estimate traversability. To address
contact dynamics that are difficult to model accurately in simulation, making it challenging to transfer simulation-learned policie
otions can be easily realized by taking advantage of high compliance wrist under the gravity. Our proposed PiH strategy is com
ose initialization, and pose refinement. In the detection phase, an existing detector is employed to identify the objects within the
eral-shaped assemblies. ASAP accounts for gravity to design a sequence where each sub-assembly is physically stable with a
sed simulator is developed that uses ant-colony inspired optimization to generate roadmaps tailored to the specific characteristi
, industrial warehouses). Existing solutions either depend heavily on the matching of elementary geometric features (e.g., point
gies that can generalize in different settings, providing good performance under different and uncertain scenarios. In this paper
thedomain of multi-robot patrolling, two approaches have received the most attention thus far. The first involves coordinating a
agent is less efficient than the others. Centralized control overcomes this bottleneck by leveraging global system state informa
or and servo stuck failures. The solution is used for designing a reduced attitude controller based on linear quadratic regulator (
on monitoring framework for distinguishing between expected and unexpected post-impact behaviors. The classification is base
efore, anomaly detection (AD) delivers a practical solution, using only normal data to learn to detect unusual events. We introdu
gned, the unpredictable nature of physical robot-object interaction may prevent success in object manipulation. Therefore, exec
odel-free reinforcement learning -- FT-Net, which is guided by an inverted pendulum model and support polygon. This pipeline
of commercial 3D printers. We achieve this by modeling the dynamics of the corrupted printer using a machine learning model
spread, preventing loss-of-control (LOC) is an ever growing concern. Unfortunately, LOC is not clearly defined for quadrotors, o
vehicles, pedestrians, intersections) and the sensed inputs of AVs. Existing efforts either assume for such semantic data to be
ach to generating contingency handling procedures by leveraging recent advances in Large Language Models (LMMs). Our app
ng exploration efficiency. To address this issue, we propose a method that utilizes a 3D dense map to generate Segmented Ex
omous power line inspection. The perching mechanism grips onto the power line, allowing the UAV to withstand external forces
n be communicated through low-bandwidth channels during large-scale mobile robot deployments. The generative property en
this, we develop a planar robot-terrain contact model and divide the robot’s motion into hybrid modes of driving and travers
ssue, this work proposes i) a robust multi-category (non-)localizability detection module, and ii) a localizability-aware constraine
t they come with high operational costs. Instead, Autonomous Underwater Vehicles (AUVs) have shown promise, but demonstr
plifications or limited onboard resources. We propose to combine source localization with a Gaussian Markov Random Field (G
ch is regarded as the baseline method. Subsequently, on the basis of our constructed turning datasets, a model is trained using
les (UAV), becomes more necessary and popular. Most of the previous works use a prior map for inspection viewpoint determin
gle- and dual-axis sun trackers automatically adjust the tilt angle of photovoltaic cells so as to directly face towards sun, but the
ly covered in the literature. Images collected by an Unmanned Aerial Vehicle are segmented using deep learning to extract ball
the future industry. Previously, the power consumption of the robot motion was the main factor considered in the evaluation of
ndings—especially amidst wind disturbances, obstacles, and unreliable global localization—we introduce a mobile hub conce
se of resources in response to incidents can be facilitated. In this paper, to incorporate the operational requirement on map gra
sensors, such as satellites, can provide a snapshot of the target phenomenon at regular cycles. Between these cycles, Autonom
strial-grade laser scanners used for high-precise 3D crop point cloud creation. The calibration method optimizes the transforma
ndividual particles with such high detail over the extensive domains associated with events such as wildfires or volcanic eruption
blem, we propose a method for traversability prediction in challenging outdoor environments using a sequence of RGB and dep
bilities into a single, integrated sensor unit. ViTacTip features a transparent skin to capture fine features of objects during conta
tile sensors for robots is still challenging. In this paper, a novel highly sensitive piezoresistive sponge based on multi-walled car
ng solutions. While these platforms continue to improve in regards to latency and power consumption, within recent literature on
orques transmitted from the grasped object to the fingers through extrinsic environmental contact may be particularly important
y for fine and reliable control during manipulation. Here we propose a deep RL approach to object pushing using tactile sensing
ral teleoperators reflect undesirable operational forces to the operator. The proposed X-lateral teleoperation framework benefits
ably enhances the three-axis force sensing capability of typical GelSight sensors. Specifically, L3 F-TOUCH sensor comprises:
olution tactile sensing and accurate proprioceptive sensing. As such, we present GelLink, a compact, underactuated, linkage-dr
we introduce RainbowSight, a family of curved, compact, camera-based tactile sensors which use addressable RGB LEDs illum
interpret the dynamics of both the objects they are interacting with and their own operations. While significant progress has be
er to acquire sufficient accuracy. In this letter, we propose AllSight, an optical tactile sensor with a round 3D structure designed
et versatile tactile sensor that offers 3D shape reconstruction and 6D force estimation capabilities. Conceptually, 9DTact is des
le sensor named GelFinger that closely resembles the human finger and is well-suited for detecting various complex surfaces.
nts, or without installing more sensors on the end effector. In this context, this paper presents StereoTac, a novel vision-based
n attractive option for implementing vibration sensing as they offer a fast response and can be sampled quickly, are affordable,
is crucial for enabling accurate and purposeful object manipulation. Humans retrieve this information primarily using the contac
on spaces. Here we introduce a dual-arm tactile robotic system (Bi-Touch) based on the Tactile Gym 2.0 setup that integrates
g coverage of these sensors. To this end, we present AcTExplore, an active tactile exploration method driven by reinforcement
ed as the robot traversed the ground. Many existing vision-based approaches are limited in directly perceiving the intrinsic prope
y. To address this limitation, we introduce TEG-Track, a tactile-enhanced 6D pose tracking system that can track previously un
(IR) tactile sensing and cloth manipulation heuristics to achieve in-air unfolding of randomly crumpled rectangular textiles by m
mbrane mechanics and air pressure. This paper presents a model-based method of reconstructing dense contact forces from th
y in labor-intensive tasks like harvesting. Although soft robotic grippers offer a significant promise for delicately handling fragile
novel multi-modal vision-based tactile sensor, SATac, which can simultaneously perceive information on temperature, pressur
and speed in dynamic touch sensing. This paper introduces a new framework that poses the touch sensing problem as an estim
ng a combination of stretchable fabric and hyper-elastic polymer, CushSense identifies contacts by monitoring capacitive chan
ng with real data. In addition, work on high-resolution sensors commonly focus on ones with flat surfaces while 3D round senso
erous manipulations. However, the prevailing GelSight sensors entail intricate calibration procedures, posing challenges in their
ensors generally focus on generating images without markers, owing to the challenges in accurately generating marker motions
a receding horizon fashion to control the pose of a grasped object. This approach consists of a discrete pose estimator that trac
nical significance related to bronchial anatomical structures and address issues related to singular points and discontinuities in
and enable continued progress. Even when these primitives are implemented in an open loop manner at clock-driven intervals,
ng method, which does not require complex analog components. A general Quadri-Terminal Impedance Network (QTIN) model
h vision feedback on garment unfolding usually suffers from occluded grasping points. In this work, we propose a dynamic man
physical world, enabling a more nuanced and adaptable response to changing environments. Nevertheless, much of the earlier
rs that operates at three different levels of granularity to ensure that robot interacts with the environment safely. The results sho
e propose a new generic approach for online camera-robot calibration that could deal with the least feedback input available in
 ility of the state of various multisensor systems with a particular interest in unobservable directions induced by degenerate mot
for contact state estimation and contact tracking control. In this paper, we propose a tactile-guided planning and control frame
s study, a flexure element comprising a structure that is easier to cut and process than conventional ones, as well as holes thro
ue to high computational time and reliance on precise surface equations. This study presents an ultra-fast IFS method that uses
rms of filtering foot-ground interaction noise and suppressing the accumulation of estimation errors. Additionally, our estimator
al dynamics, is established for the manipulator, and the Gaussian process regression (GPR) technique is employed to learn the
manufacturing, we propose a task-based tactile robot programming paradigm that uses an object-centric tactile skill definition th
ment, or user-defined task design changes. In cases where a modification is not possible, we provide feedback to the user and
arial player whose objective is to prevent achieving temporal and reachability goals. The aim is to synthesize a winning strategy
mely, the classical discrete-time linear system with linear state feedback controller and an extension with variable update rates.
s of predicates is usually defined in a model-free way, i.e., without including the system dynamics. Moreover, it is often nontrivi
dy orientations or the positive definiteness of stiffness and manipulability ellipsoids. Handling such geometric constraints effecti
sed to compute a relaxed specification automaton that captures all allowable relaxations of the mission specification and their c
te weighted graph, where the weights represent travel durations. Given the probability distribution over the durations of each ed
stance comes at the cost of a time delay due to network latency, resulting in past observations being used in the cloud to comp
n such environments with tasks expressed in Metric Interval Temporal Logic (MITL) - a rich robot task specification language th
To address this issue, by incorporating Transformer into reinforcement learning, we develop a Double-Transformer- guided Tem
agents. Previous works have considered reactive synthesis, where the human/environment is modeled as a deterministic, adve
tonomous generation of reactive tasks specifications during the runtime. As a first attempt towards this goal, this work develops
extend conventional atomic proposition to batch atomic propositions, which gives rise to the LTLT formula. The Task Batch Plan
rally induces commonalities among tasks while preserving safety guarantees. However, most prior work on reinforcement learn
atility. We analyzed this problem, and the solution was achieved by separately modeling changes in the film thickness distributi
atural language abstractor, are verifiable by construction. Leveraging recent advancements in motion primitives and probabilist
questions on the basis of reasoning about observations as they are made, interrelationships between multiple pieces of evidenc

of transparent objects. First, we propose a multi-scene synthetic grasping dataset named SimTrans12K together with a Gauss
and a notable feature of our hands is that they can vary their stiffness to match the requirements of the task, e.g. become stiffer
jects a robot has previously experienced. Existing approaches to solving this problem are typically restricted to a specific object
n difficult to observe and must be inferred through interaction, which can be expensive in both inference time and number of int
grasping framework to learn realistic and reasonable grasps in cluttered scenes, which tackles the problem in three items: 1) g
configurations parameterized by an object-centric continuous grasp function across varying instances. To ease the transfer proc
ther such systems are vulnerable to adversarial attacks, as that could lead to large economic damage. In this paper we propos
ch lacks the capability for deliberate and reliable manipulation at scale. Object identification (ID) unlocks the ability for large-sca
cts or the interactions of object features at a global level. This approach leads to imbalanced distribution of features and the ge
an external force and received by a microphone at the end of the channel. The sensor leverages regression and classification m
ning the operational space, an increasing number of works propose grasping datasets to learn from. But most of them are limite
used for the description of human grasps as well as the generation of novel, human-like grasps. However, the abstract synergy
perturbations of a seed grasp. The candidates are then scored using a novel metric which combines a learned grasp-quality est
ers, it alleviates the complexities tied to perception and grasp planning, especially when dealing with unpredictable object poses
e grasps. These grasps need to be compliant with the local object geometry. Second, for each proposed grasp, the robot need
asping trials. The GRM automates the process of resetting a grasping environment, repeatably placing an object in a fixed locat
 dvancements in interactive imitation learning have proposed a promising framework for human-robot teaming, enabling the rob
more effective representations based on existing pre-trained visual models remains unexplored. To better leverage the repres
ls, but navigating the contact-rich high-dimensional search space from solely sparse visual reward signals significantly exacerb
omprises three main components: object shape reconstruction, grasp sampling, and implicit grasp pose selection. GPDF accur
d to the multi-finger grasp space. Based on this representation, we develop a simple decision model that generates accurate gr
curate adjustment the desired grasping poses for objects of varying shapes and sizes. In this paper, we propose a novel dexter
irregularly shaped objects with unmodeled surface-mounted components. Vision-based separators, while effective for object ca
a robust understanding of object properties (such as weight, size, and shape), as well as a remarkable capacity to interact with
hen compared to those of other solutions, without taking into account the random variability across different experiments (e.g. d

ate and translate a grasped object, i.e., can perform in-hand manipulation. This in-hand manipulation is effective for grasping in
nd wire routing complexity. Finger-based sensors are limited to sensing external gripping force, and fingers cannot be easily rep
ive topology-optimized design of a compliant finger consisting of several multi-material phalanges connected by flexure hinges.
speed while remaining within hardware limits. By adding overload clutches in either the robot arm or gripper, tolerance to errors
timal vertical vibratory transport of a part. Namely, such a system must be capable of quickly realizing high accelerations, as we
s mechanism combining memory alloy actuators with particle jamming to achieve the desired bending angle and actuator stiffne
and 5.5mm in three dimensions. By integrating permanent magnets and elastic ropes as internal actuation elements into the fing
maximum repeatability during the gripping and releasing stages of a pick-and-place process for a chip-like silicon die. The design
structive processes such as tearing, cutting, or milling to speed up the disassembly. Currently, such destructive actions are only
e hand's interaction with the grasped object), but it is not anthropomorphic.The basic finger and hand geometry, governing kine
his paper introduces an innovative soft-rigid, tendon-driven gripper: the Double-Scoop Gripper (DSG). Its two-fingered design e
ronmental constraints to structure the co-design space for manipulation. We show that task-relevant constraints encode regions
eractuated gripper for passive landing on impact. We fix the geometry of the gripper and vary the joint stiffness and contact cond
pery, spiky and scaly objects. In this paper, we propose a novel robot end effector, called HASHI, that employs chopstick-like ap
skills. This paper introduces an inexpensive, modular, and robust platform - the D'Manus - aimed at resolving these challenges
imitate human force skills have been actively explored as a solution. In particular, bilateral control-based imitation learning ach
ing and restocking in retail and warehouse settings where products vary in size and weight and are packed tightly, which limits
odulation, accurately modeling orientation, handling trajectories with multiple options, and capturing aiming actions. In this stud
his model pays attention to both spatial and temporal features of tactile sequences from sensor arrays, which is verified by four
ertip via a highly under-actuated kinematic chain, exhibits many benefits in wearability and portability over other solutions. How
ponents provide high-frequency tactile cues for initial contact and surface information (e.g. texture) as well as pressure stimuli r
an affect its normal use. Determining how to produce desired vibration patterns at any contact point with only a few sparse actua
f a haptic- enhanced Virtual Reality simulator for surgical robotics training, featuring 8 surgical tasks that the trainee can interac
input devices, along with the general lack of haptic feedback available with such devices, can lead to non-intuitive control. In th
ring systems. The device focuses energy to a 1 mm^3 voxel. The current prototype was validated with a small, preliminary hum
These dependencies make training tactile perceptual models challenging. Additionally, the effects of limited sensor life and the
d a multi- stable curved line display inspired by physical splines. By using circular splines to initialize a discrete elastic rods simu
g the scanning process, which affects reconstruction quality. We propose VIDAR, which visually guides users with the streamin
y when the patient deviates from the correct movement pattern. Equally crucial is the need for the LLE to exhibit "transparency
data. Recognising the challenges posed by low-cost sensors, which are highly susceptible to environmental conditions and ofte
d in literature, however showing several limitations in realizing multiple grasping motion control. In this work, we propose a nove
 rs may result in protracted and awkward robot motions, contributing to unpleasant sensations by the human partner and affecti
ng of various objects. Creating datasets from scratch is a time and labor-intensive process. Moreover, existing datasets often c
ny requests can overload the expert. In this paper, we present a Reinforcement learning-based approach to this problem, where
g using modalities like simulators and VR and no performance assessments with line-of-sight UAVs. This paper presents a train
work to plan a policy of a collaborative robot that reduces the human fatigue in the long term, in highly repetitive co-manipulatio
wered by humans. In this work, we approach the sample-efficiency challenge by presenting a technique which synthesizes que
mpractical in every environment. Approaches that utilize human demonstrations through manual operation can aid in refinement
heavy traffic - can be costly and unsafe to collect. We propose a virtual reality human-in-the-loop simulator, JaywalkerVR, to ob
lancing and charging to meet users' travel demands still has been a challenge. Existing methods generally ignore human prefe
operators to fly the drone in a racing scene at a high speed. The whole system is structured hierarchically, consisting of both o
d force reference, facilitating the execution of arbitrary movements by the wearer. In case of a sudden contact/load loss, a powe
rst use sheet gel as a multi-dimensional passive elastic element and combine it with a two-finger gripper to design an elastic gr
behavior, this work addresses this challenge through the implementation of an energy-aware Cartesian impedance controller,
ne to fly over a table where important letters and the newspaper are stored, even though it is the shortest path to the goal. The
pedestrian paths is vital for ensuring safe human-robot interaction, especially in robot navigation. Furthermore, this research has
presentative training data. To ensure human safety and comfort, it is critical that these algorithms handle uncommon cases app
hen interacting with others. To gain a better understanding of these behaviors, this work presents new data that aims to verify h
now?", which is an important primitive for home robots. Its answer is challenging and inherently probabilistic as the robot only o
plays a significant role. However, existing pedestrian models often struggle in complex environments and do not have the capa
ehavior. Deep-learning-based models have recently achieved considerable success in trajectory prediction and many models in
proposes a Graphic- and Timed-Elastic-Band-based approach (GraphicTEB) with spatial completeness and high computing eff
xplanations. These explanations reveal the objects in the environment that are responsible for the failure, and how their location
rk via an Augmented Reality (AR) teleoperation system that can efficiently learn human demonstrations and provide 6-DoF gra
zed approach to recognize human-robot trust. This study addresses this need by introducing an EEG-based method for trust rec
we propose an easily implemented framework that can directly handle heterogeneous data sources for classification tasks. Ou
n individuals, yet successful interactions are expected across groups. To address these challenges, we formulate a set-wise pr
when to perform low-confidence actions, thus actively controlling the risk of trust reduction. In this paper, we propose Self-Confi
reate an action-aware costmap to perform effective path planning without fine-tuning. With the large models, we can achieve an
myography (HD-sEMG) modules combined with deep learning decoding have attracted interest by achieving excellent gesture p
for predicting hand motion and intended target during human-robot handover using Inverse Reinforcement Learning (IRL). A se
ation and team collaboration. In this paper, we investigated trust-aware robot policy with the theory of mind in a multiagent setti
shows that Haru is able to learn and imitate human different game strategies from human players in a human time scale. In add
to sports participation, such as their surroundings. Many related studies including our previous work for assisting users in doing
he front wheel assist. In order to solve this problem, this paper proposes a rear-wheel-swing assist mechanism that minimizes th
care recipients using one robot arm mounted on their wheelchair due to ergonomic and transferring challenges. This paper intr
multi-output variation of the classic autoregressive model. The network used an encoding-decoding method to construct conden
cted to facilitate user instructions to robots, its current implementation requires users to explain the details of the path for each n
sulting in cumbersome manual operation or inefficient exhaustive search patterns. We present a human-centered autonomous
eplace the superior ability of experts to make decisions in complex situations, thus making human-robot interaction systems a p
robotic solutions at these facilities that are reconfigurable, extensible, and compatible with pre-existing software infrastructure.
user experience questionnaires to access usability, NASA Task-Load for workload, two questionnaires for human-robot trust a
ncy while guaranteeing pHRI stability. Despite its significance, the real-time estimation of EoP remains an under-investigated top
such as household service, search and rescue, and environment monitoring. Despite many applications, the existing works do n
tate-dependent estimation of desired torques with an inverse dynamics model and a data-driven component in the first approac
advance in human-robot interaction whereby human interaction is within a machine that physically envelops us. We discuss the
ed from observable behavior like movements during the interaction with a large industrial manipulator. To this end, we report on
osed-loop human-following control strategy that combines model predictive control (MPC) and impedance control. The outer-loo
-based optimization. The framework is applied to the optimization of a shared controller for an Image Guided Therapy robot. VR
ely observing human trust in robots with precision is challenging, we adopt the widely used partially observable Markov decision
server) to choose the robot's behavior, but such models do not account for the influence the robot has on the human's actions,
viable tool for representing a machine Theory of Mind and for modeling communication in epistemic planning, with applications
e of uncertainty without requiring assembly skills and knowledge of robotics from the human operator. We demonstrated the eff
ent a collaborative 3D SLAM approach leveraging the power of augmented reality (AR). The system introduces a trio of diverse
e generation of new synthetic data based on existing training data has been enabled with the development of generative advers
by demonstration framework for generating reactive and modular behavior trees. CoBT relies on a single demonstration and a
nefficient behavior, such as unnecessarily reducing robot speed or unpredictably changing the robot path, which may negativel
mprove the overall surgical performance by supporting the surgeon effort during movements while keeping them in charge of co
a speaker. We employ this hardware in two forms of robot-to-human communication: synthesized text-to-speech (TTS-Soneme
addressed in this work: motion generation and shared autonomy. We propose a motion generator based on a time-invariant po
can be formulated as multi-agent RL, wherein there are two agents: a caregiver and a care-receiver. However, policies trained
al domain. We propose a reinforcement learning algorithm, transverse generative adversarial imitation learning (tGAIL), which
etrieve the right information in new settings and reduce the intervention rate, but also they would need to be able to respond to
. In the literature related to control strategies for robot manual guidance, it is usually proposed to control translation and rotation
to human-robot interaction forces to adapt to the task requirements, while generating virtual “ergonomic forces” that guid
' intent and modulate control authority to provide natural assistance when needed. We employ deep inverse reinforcement lear
per presents a novel extension to dual-arm tele-cooperation, leveraging the non-linear stiffness and passivity of FIC to adapt to
tion of an inflexible planning problem. Large Language Models (LLMs) have shown promising generalisation capabilities in robo
ations of the robot's policy. Much prior work in explainable AI and RL focuses on generating explanations for single-agent polici
s not the only aspect that must be taken into account. For many tasks, such as welding, it is crucial to ensure that the robot perf
ctless and physical HRCs. With the framework, the robot can perform contactless HRC under safety regulations regarding the c
particular, Power and Force Limiting (PFL) collaborative mode describes how the energy should be regulated during the collabo
xisting models, even the most advanced large language model GPT4, still struggle with dealing with multiple tasks by single-rou
a more natural and intuitive interaction. For this purpose, the main contribution of this work is a psychological and cognitive arch
considering human feedback as preferences between pairs of trajectories (sequences of state-action pairs). However, when si
e the Collaborative FRIDA (CoFRIDA) robot painting framework, which can co-paint by modifying and engaging with content alr
m, MateRobot, is established for PVI to recognize materials and object categories beforehand. To address the computational c
physically coupling their hand with the cobot's end-effector. These components, receiving a goal from the user, traverse a collisi
tead, the full potential of MPs remains to be attained via adaptable, large-scale MP libraries. In this paper, we propose a set of
eliminary observations from a pilot human study, we categorized post-fall recoveries into a sequence of statically stable poses c
We propose a method for targeted bedding manipulation over people lying supine in bed where we first learn a model of the clo
(recency bias), followed by information presented first (primacy bias), followed by information in the middle, collectively known
f human trajectories is challenging due to complex factors, including, for example, social norms and environmental conditions. T
 iori. However, each user's environment is unique and can continue to change over time, which makes these methods unsuitabl
The key is to enable the robot to understand human intentions, during the interactive creation process. We here formulate this
e who have little to no motor function, as well as being able to re-associate movement with brain functionality. Although BCIs h
nditions. Meanwhile, we develop a wearable environment-assisted system. In addition, we collect and annotate a new dataset c
simulation stage and a robotic drawing stage. In the simulation stage, geometrical contours were automatically extracted from a
al positions. The system mainly consists of a mapping and a control module. The mapping module is a real-time multi-agent vis
zability across skill conditions. Specifically, we learn the data-driven state transition dynamics of the robot in the formulation of e
ive force model for bipedal walkers on soft granular terrains by introducing effective intrusion depth correction. The enhanced f
angle of the terrain. In addition, a re-planning approach is introduced to correct the robot state off-track from the reference gait
man motions. First, we present a unified primitive-skeleton motion retargeting to mitigate morphological differences between arb
ose it into two sequential components: an Efficient Behavior Generator (EBG) and a Nonlinear Centroidal Model Predictive Con
ation transmission mimic the reflex arc in humans to improve bio-plausibility. The proper synaptic strengths between the senso
ediction while minimizing false positive rates. The proposed algorithm uniquely integrates the detection of various fault types an
spired by human body anatomy. The robotic mannequin can adjust the anisotropic deformation of its human-like skin to imitate
on-directly measurable effects, such as external contacts, thus leading to joint torques readily usable in control architectures for
e deployed to the real world, due to the significant domain gap between the simulated and real data. In this paper, we propose
nflicts. These conflicts encompass extrinsic conflicts caused by BEV feature construction and inherent conflicts stemming from
e cooperative 3D (VIC3D) object detection: $1)$ inherent pose errors when fusing multi-view images, caused by time asynchron
aset with ground truth annotations for different visual intentions to develop methods designed to classify the state of such syste
tereo-matching and dehazing, simply removing fog might not aid depth estimation because the fog itself can provide crucial de
ment image, the diversity of defect geometries, and the morphological ambiguity between classes. We propose a novel end-to-e
with networking being one of its supported core features. In the context of autonomous driving, V2V (Vehicle-to-Vehicle) techn
have extensively explored various aspects of this task, including view transformation techniques, ground-truth label generation,
ounding box and the accurate alignment of 3D point clouds with 2D images, making direct application in real-world scenarios ch
ely on a complex cascaded framework with relatively limited information to restore 3D scenes, including a dependency on super
. As a result, current state-of-the-art 3D detectors produce high rates of false-positives to ensure a low number of false-negativ
any, leading to the Distribution Gap of different private data for training distinct agents in multi-agent perception system. The da
s can be influenced by a range of factors, including wireless transmission and obstructions like buildings. Given the pivotal role
g methods have been made using curated data, i.e., with the assumption of perfect maps, detection, and tracking. This paradigm
pt of query cooperation to enable interpretable instance-level flexible feature interaction. To specifically explain the concept, we
of this background, this paper proposes a noise-distribution based defogging method for point clouds. Our approach hinges on
tial-wise fusion and adopt feature compression to alleviate bandwidth cost by high-resolution feature maps, which would inevita
munication to share perception information with others. However, most relevant works focus only on cooperative detection and
classification tasks. In this paper, we study these vision models with respect to the open-loop training to closed-loop generaliza
to enable robots to cope with long-horizon tasks. However, generalization to unseen tasks in a different domain with a human p
value function at a single time. Driven by the gap between maximizing the cumulative rewards of RL and the greedy strategy o
e most visually similar object experienced during human demonstrations, and then uses this object to align its end-effector with
performed. Moreover, this data can even be mined from video datasets or the web. Ideally, this data can serve to guide robot le
improves data efficiency. The information gain criterion focuses on precisely identifying all parameters of the reward function. T
possess the above-mentioned skill has been confirmed as a crucial and challenging task. In this study, a general framework is
an imitation learning approach. The optimization techniques include quantization, pruning, fine-tuning (retraining), and clustering
 a broader set of energy function constraints, such as gravitational and elastic potential energy of 3D deformable objects. Previo
constraint of the ball on the cloth is defined focusing on the lowest point of the ball. Next, the relationship between the input to
hand has three identical underactuated fingers, each of which contains an underactuated planar linkage, a spherical four-bar m
essible ones in the real world. However, the feature reduction from the simulation to the real world is conducted through an emp
unctions to learn how to accomplish tasks. Moreover, policies typically issue actions directly to controllers like Operational Spac
ainties in kinematic and physical parameters. Therefore, we propose a robust in-hand manipulation method, which consists of t
s work, we aim to improve the sample efficiency of learning dexterous in-hand manipulation skills using controllers available via
lation. Specifically, we propose Robot Synesthesia, a novel point cloud-based tactile representation inspired by human tactile-v
otion while maintaining contact with the object and, by staying within the hand's frame, ensuring that the object remains manipu
suction cups, called the perimeter cups, are attached to the vertices of the robot’s pyramid base, whereas three other suctio
ables and robotic grippers. We employ the Approximate Convex Decomposition technique to dissect the non-convex cable mod
C law is employed to compensate for the unknown residual error dynamics caused by elastic and transmission effects. The pro
It combines a sophisticated 3D perception pipeline for surface scanning and automatic defect identification, an interactive voice
os when idle in a specific one. Currently, a unified system capable of coordinating and using these resources efficiently is lackin
etection and image segmentation. As a result, improving the quality of low-light images is of paramount importance for practical
propose an autonomous robotic control method to constrain a robot-held camera within a designated FoV. Furthermore, we mo
aint parametrization to achieve superior efficiency compared to general-purpose nonlinear programming solvers. Additionally, w
ork, we propose a control solution for omnidirectional mobile manipulators in force-tracking tasks, interacting with unknown surf
omplex distributions (e.g., non-Gaussian or multi-modal) and cannot provide accurate guarantees of performance. In this paper,
ompound when predicting the future behavior of those agents. Alternatively, dense occupancy grid maps have been utilized to
ve continuously in time, the trajectory optimization problem is reformulated into a time-derivative form, whose decision variables
rake-specific Fuel Consumption (BSFC) map of the engine. We propose a pure data-driven method, Neural Predictive Control (
ment noise. The approach iteratively solves the robustified problem by incorporating updated state-dependent safety margins for
rrent optimal trajectory. This can potentially result in convergence failures and safety issues. Therefore, this paper proposes a
with complex logic specifications by co-learning planning and control policies. Notably, this approach significantly reduces the s
is to propose an output-sampling-based MPPI (o-MPPI), which improves the ability of samples to satisfy output constraints and
butions are three-fold: i) a method for discretizing the state and control spaces to amortize planning times across multiple queri
afety, convergence and optimality. Moreover, in simply-connected workspaces, our method is proven to asymptotically provide
of the model form (analytical, simulator, learned, etc...). However, real-world data is often expensive and dangerous to collect.
mitations of real-world perception systems lead to potentially of degraded perception performance when traversing such environ
ic system to learn how to perform a task. We propose the method Behavior-based Bayesian optimization and Planning (BeBOP
et to be leveraged for future planning. In this paper, we present a simple but efficient method called Motion Memory, which allow
sal confusion. Previous studies have explored the phenomenon of causal confusion, with a specific emphasis on the context of
fficient and optimal motion planning. The SOTA learning based motion planning algorithms utilize paths generated by sampling
hese methods need a short planning horizon to be practical. Thus, MPC is ill-suited for long-horizon manipulation tasks due to lo
biases the sampling distribution to improve the planner's performance. The key contribution is to formulate the sampling bias p
he robot's end effector. Most methods require the user to manually annotate correspondences on the CAD model, however, thi
ed set of measurements that can be obtained in an operational setting. These quantities exhibit inherent stochasticity and can o
we propose a compact, Internet-connected, electronic task board to measure manipulation performance remotely; we call it the
oundary conditions. This objective is achieved based on a model transformation scheme for homogenizing boundary conditions
g a deep learning model to detect and segment the opening, which represents the area of interest on the plate. Subsequently,
elatively large and flat finishing disk. In this paper, a planning method is presented that uses a variable contact point on the finis
ge, this paper presents a mathematical model framework for a mobile multi-skilled robot-operated manufacturing system that in
ng a majority of the algorithmic components to the edge or even cloud offers an attractive option to leverage massive computin
mixed traffic remains difficult due to the stochastic behaviors and uncertain intentions of Human-Driven Vehicles(HDVs). To tac
ause negative effects on the convergence of the learning process. In this paper, we propose a new contrastive divergence loss
nd ActorTypeTruck) to be queried via exact tag matching. Such a scenario database design has two major limitations, i.e. comb
all times? The work formalizes this scalable supervision problem by considering remotely located human supervisors and inves
generate human-like driving trajectories, accommodating multiple driver levels determined by model parameters. This approach
ble of reasoning about behaviors at different levels of granularity. To this end, we propose Dynamic goal quErieS with temporal
ng task of CAVs as an optimal control problem with safety conditions enforced as hard constraints, and propose a computation
e that autonomous vehicles take safe and efficient maneuvers in different interactive traffic scenarios, we should be able to eva
in the system. To address this issue, we propose a Cognitive Digital-twin-based Assistance System (CDAS) that issues warnin
back-annotate the criticality of the interaction to the object detector. However, those approaches do not consider the safety fact
and qualitative studies have found that Random Sampling may be less suitable for the autonomous driving scenario, since the L
he association task while either neglecting the model’s performance on state estimation or deploying complex heuristics to
namic allocation of communication resources. Moreover, in the context of collaborative perception, it is important to recognize
te networks, and then use the policy as one of several mutation operators in a evolutionary algorithm. We evaluate this algorith
nty and Disturbance Estimator (UDE) to compensate for uncertainties arising from imprecise model parameters and exogenous
ew procedures to manage the transition from human-driven to fully- autonomous vehicles while providing a feedback-loop mech
Object Tracking (MOT), the capabilities of 4D imaging radars remain largely unexplored. Recognizing the challenges posed by
. This gives rise to mixed traffic control, where robot vehicles regulate human-driven vehicles through reinforcement learning (R
obotics that transmute the outputs derived from perception models into linguistic forms, thus adopting a `pure-language' strateg
e a dif-ferentiable joint training framework for both ego-conditioned prediction and cost models, resulting in a direct improvemen
e scene. Moreover, to mitigate the increase in computational complexity associated with the number of agents in the scene, som
enchmark for evaluating and improving model robustness by applying perturbations to existing data. Specifically, we conduct an
or. We argue that such approaches are crucial for AVs' future, particularly when navigating alongside human drivers on highwa
einforcement Learning (DRL) presents a promising avenue in this direction, but learning DRL control policies that generalize to
base policy produces a near-optimal stable policy. In the routing setting, a policy is stable if its execution keeps the number of o
focus on generating high-risk driving scenarios and applying them to conduct safety-critical testing of AV models. However, lim
g as the target task, consisting of both perception and control. We first utilize information bottleneck analysis to build a causal gr
g dependence on historical trajectories, which yield unsatisfactory prediction performance and robustness. Additionally, their he
hable set for the longitudinal position of the vehicle along the lanelets of the road network, where lane changes are modeled as
it untrustworthy to apply them in practical use. In this work, we explore the interpretability of behavior prediction of target vehicl
d by human errors. However, the current design methods have several major limitations. Firstly, the warnings are mainly genera
to be dynamic obstacles and, as a result, they are incapable of capturing the full intent of the human drivers through this passiv
signs, which are unsuitable for complex scenarios. In this work, we introduce Informed Reinforcement Learning, where a structu
orcement learning (CRL). In our proposed framework, full-state references and regulatory factors concerning the relative import
nterdisciplinary approach that integrates principles of human cognition and observational behavior into trajectory prediction mo
ectory prediction. Built upon the backbone architecture of a timewise variational autoencoder, ContextVAE employs a dual atten
o meet the requirements of home-based rehabilitation, including affordability, portability, safety, and active participation. Althoug
on of motion. Our novel intent estimator, dubbed the Robot Trajectron (RT), produces a probabilistic representation of the robot
eft/right motion intention prediction, and ii) an online realtime prediction pipeline leveraging the trained classifier and integrated w
rate macroexpressions from a single microexpression image. In the first stage, we extract the expression information from a sin
In this work, we propose a novel method of utilizing both in-cabin and external camera data to improve state-of-the-art perform
rectives typically seen in robotics. Unlike most prior work where navigation directives are provided as simple imperative comma

propose a new metric for more effective ranking within the scenario-based evaluation. We conduct extensive empirical studies o
nature of hand motion and the diversity of HRC tasks. This paper proposes a framework that combines hand trajectories and ga
must smoothly absorb the energy of impact and simultaneously engage with the target in a short period of time. We present a bi
of these designs is their reliance on external sensors like rotary encoders, which restrict mechanical joint design and introduce
actuators over traditional and other modern actuators in mobility assistance. The working principles of the proposed Variable S
cuses on the use of series elastic actuators which can apply consistent trajectories of scrubbing force. Consistent force during s
oblem. This issue is critical as having multiple equilibrium states can lead to erroneous estimations of the manipulator's profile.
don transmission path to reduce tension loss and elastic deformation has become an important research direction. In this pape
c stability analysis of these robots for arbitrary planar designs. The buckling phenomena are investigated and analyzed using bi
ed Rocaspect, that has four compliant masts. The robot behavior and accuracy is evaluated experimentally and three different m
cts of cable pretension and shape on friction and torque efficiency. A custom self-designed testbed, comprising integrated actua
matic optimizer, underpinned by an autoencoder-based architecture, to address these challenges. Utilizing a neural network, ou
mation, and the difficulty of the water jumping increases with the robot's scale. This paper designs a miniature water jumping ro
al path using a binary search algorithm. Computing a trajectory from waypoint to waypoint leads to the fastest progress on the p
ulation of the neighboring orientation. For two-segments, a reachability criterion is proposed and the redundancy of its inverse
RPRR manipulator, we treat this as an inverse kinematics problem. The kinematic features of the 3D Dubins path are built into t
ally, both an accurate evaluation of the twist and wrench capacities of a robotic manipulator. However, this index is the minimum
ources of engineering time and costs. In this paper, we propose a novel concept for optimizing the fitness of a robotic arm to pe
its translation and neglecting its rotation. We propose a novel trajectory representation for rigid-body motions that incorporates
rt has been devoted to the development of intricate constitutive models for finite elasticity, we show that for the most part, the is
o come in contact with it at a non zero relative speed. The configuration of the robot and the speed at contact matter because th
pace by growing in height and density, but this can create challenges for the the robots that service them. In this context, we pre
o-two leg structure, the middle link that connects to the two proximal parallelogram four-bar linkages in each side only generate
ociated with the platform motions. In the present paper, we show that the static configurations of the sagging CDPRs are local e
havior of a CDPR featuring elastic cables while ensuring the Static Equilibrium (SE) of the Moving-Platform (MP). By modeling t
and speed limits, which are assumed to be fixed in the literature, but they are indeed coupled. This results in a conservative us
explore versatile locomotion and maneuvers on ladder-like structures. The robot’s anatomical resemblance to gibbons and
ved simultaneously. However, with certain conditions, there exists uncontrolled robot motion under the EIC-based control. We f
ve the control goal. To solve the problem of over-conservative controllers, we proposed the magnitude regularization technique
he lack of principled methods to design the class-$mathcal{K}$ functions, there are still fundamental limitations of current CBFs
imal policy. Nonetheless, safety-focused modularity in the acquisition of skills has not been adequately addressed in previous w
the physical environment. While one can train behavioral models capable of predicting human motion from past activities, this a
NNs) for processing each modality individually and a high-level RNN that integrates the multi-modality. This integration is achie
. Moreover, this deflection varies depending on the weight and length of the tool, making these models exceptionally complex. H
causal Bayesian Networks (CBN) are popular for modeling cause-effect relations between semantically meaningful environmen
n the other hand, Covariance Matrix Adaptation MAP-Annealing (CMA-MAE), an evolution strategies (ES)-based quality diversit
 hallenges: (1) the need to manipulate forceps using ports established on the body surface as fulcrums, and (2) difficulty in perc
es and/or accidents. In this work, we focus on reaching/point-to-point motions, where robots must always reach their goal, inde
target positions and obstacles. MPs enable their adaptability by capturing the variability of demonstrations. However, current M
ed to encode the task-relevant entities as nodes for a Pick-and-Place task declined at different levels of difficulty. During trainin
segmentation method that utilizes small characteristic fluctuations of 6-axis force/torque signals as features for motion segment
nvolvement. In the latter case, there may not be enough humans to quickly spread skills to many robots. However, propagation
Learning from Demonstration (LfD) approaches require a large number of costly human demonstrations. Our proposed semi-s
ile, the fast convergence of high-quality solutions is not guaranteed. Therefore, we develop a novel neural risk-bounded path p
ous agent must also avoid the high cost and potentially fatal critical mistakes. Traditionally, self-supervised training mainly focus
generalization to unseen worlds. Hence, we factorize policy learning as inference of such generic rules, which act as strong prio
thout delay. However, vertical joints are subject to gravity torques, which need to be compensated when disengaging the clutch
during the demonstration. This is accomplished by introducing a new reward function and employing a knowledge expansion te
an certain changes to a surgical technique result in lower forces and increased safety margins?” could lead to improved surg
wide range of demonstrated trajectories. We propose a novel DMP formulation with a generalized logistic function as a delayed
poses of the robots, we develop AirExo, a low-cost, adaptable, and portable dual-arm exoskeleton, for teleoperation and demo
utationally expensive and prone to instability issues. In this paper, we propose a novel Bayesian method that infers constraints b
demonstrations. Kinesthetic teaching requires physical handling of the robot, while teleoperation demands proficiency with addi
etworks with geometric inductive biases. However, such methods fail to consider that demonstrations for the same task can be
y concerns when facing perturbations. We propose SNDS, an imitation learning approach aimed at efficient training of scalable
problem to learn a stable and barrier-certified DS. Such optimization problem can be very complex or extremely conservative w
arly when those images are generated via simulation. A common method to tackle this issue is through domain randomization,
at each time step for the robot to follow, by combining tools from control theory and the target trajectory generated by the learn
y of the system. In response to this substantial challenge, this paper introduces a novel DS algorithm rooted in neural network t
, represented by a neural energy function, on the demonstration set. We propose a Polar-like space analysis approach to deriv
ure it out, we introduce a transferable constraint learning (TCL) algorithm that jointly infers a task-oriented reward and a task-ag
D), which can facilitate the identification of valid motions in cluttered environments by only using human skeleton information ex
onments due to a lack of long-term planning capabilities, and that planning in a temporally abstract model of the environment c
pacting a possibly large set of degrees of freedom composing the environment, mutual information maximization fails alone in p
erials which the robot must cut. To address this challenge, we propose a learning-based approach incorporating elements of in
RINT, a scalable offline policy pre-training approach which substantially reduces the human effort needed for pre-training a dive
ontrastive unsupervised learning use auxiliary tasks similar to those in computer vision, which does not guarantee the effective i
oaches have been shown to reproduce agile behaviors of animals, but typically have not been able to learn highly dynamic beh
rld model learnt. Due to its significant computational cost, standard domain randomisation does not provide an effective solution
ntractable problem into a supervised learning problem, where a fully-observable robot policy generates supervision for a partia
n transformer (ODT) is an effective and representative sequence modeling-based offline-to-online RL method. Despite its effec
ving the hyper-redundancy and flexibility of snake robots as mere challenges, there lies an unexplored potential in leveraging th
e scalability issues. This paper presents a novel safety-oriented reward-shaping framework inspired by barrier functions, offerin
or improving sample efficiency, it usually incurs training instability, making the sample efficiency worse. Moreover, manually cho
at utilizes graph hypernetworks to estimate the weights of multiple architectures simultaneously. Our method is capable of estim
regularization framework that modulates the robot's exploration throughout training, striking a balance between flexible improv
for position control gain tuning is required to achieve the best possible policy performance. We show that instead, using a torq
olicy is highly task-specific. Since all motors are controlled in a centralized way, out-of-distribution local observations can impact
h the addition of three new LiDAR sensors alongside the original camera, inertial, GPS, and proprioceptive sensors. We also re
ibuted to several factors: (1) Existing methods typically employ UIE as a pre-processing step, which inevitably introduces consi
escale terrains. To address this problem, we propose an efficient and high-fidelity vehicle mobiliy prediction method for a larges
ddress this issue, we present a novel method to learn [u]ncertainty-aware [n]avigation features from high-fidelity scans of [real]
rning. Without the need for feature point detection, D-GLSNet provides accurate pixel-to-point association between the LiDAR p
nvironment. This paper presents a new approach that utilizes Deep Multimodal Variational Autoencoders (DMVAEs) for estimat
ments are time-consuming and unstable. This work proposes F3DMP, which consists of two parts to alleviate the local optimum
narios. On the other hand, algorithmic data generation methods are usually limited by their model capacities, making them unab
o improve the training of multi-task reinforcement learning, which serves as the foundation for transferring. Then we conduct a
problem by leveraging planning algorithms to plan intermediate subgoals to augment GCRL. Their methods need two crucial re
e introduce KINet (Keypoint Interaction Network) ---an end-to-end unsupervised framework to reason about object interactions
k. By leveraging LLMs as an assistive intrinsic reward, IGE-LLMs guides the exploratory process in reinforcement learning to ad
y consider the history of observations and actions. Previous studies have shown that Recurrent Neural Network (RNN) can be
new tasks. Thus, we adapt language instruction to specify deformable object manipulation tasks and propose a learning framew
ented Policy Gradient (TAPG), a novel two-stage learning framework that synergizes reinforcement learning and policy distillatio
tructure present in the world and the robot, and to support sequential composition by construction, so that learned skills can be
s range from gently capturing living animals such as snakes or fish for biological investigations, to smoothly interacting with and
control. Those methods, while effective, rely solely on minimizing forward prediction errors to hope for better task performance
k training. Addressing the challenge of obtaining policies with generalization and stability represents a non-trivial problem. To ta
ctile feedback in the control loop. TEST's core design is to learn a skill transition model for high-level planning, along with a set o
ning with only visual tactile sensing inputs on diverse objects in a physical simulator. By training with diverse objects in simulatio
e formulated the problem as cooperative mTSP, a variant of mTSP where agents share cooperative costs, and utilized reinforc
ntation and policy architecture. We propose EquivAct which utilizes SIM(3)-equivariant network structures that guarantee gene
ead to catastrophic failures in real-world human-centric environments with large and unexpected perturbations. In this work, we
uch models are difficult to calculate analytically and vary among DLOs. Consequently, manipulating DLOs poses significant cha
when trained with new arrivals. For humans, researchers discover that there exists a memory replay mechanism in the brain to
sks. However, it is computationally expensive and inefficient to learn navigation end-to-end from raw high-dimensional sensor d
terms, each requiring a different trade-off. In this work, we take a constrained optimization viewpoint on the quality-diversity trad
aptability to various terrains. Our approach embraces a hierarchical reinforcement learning (HRL) architecture, wherein the high
ch conditions. In this paper, we propose RaLF, a novel deep neural network-based approach for localizing radar scans in a LiD
s, we develop VPE-SLAM, based on the proposed voxel-permutohedral encoding, which can incrementally reconstruct maps of
oss simple scenarios, yet fail to navigate in complex environments. On the other hand, end-to-end (e2e) deep reinforcement le
ing an explicit representation of environmental structure. This limitation hinders effective environmental representation and inte
e environment through learning-based approaches can greatly enhance task efficiency. While many such works build task-spec
t gym, a benchmark suite for safe RL in HRC. We provide challenging, realistic HRC tasks in a modular simulation framework. M
vely, framing the collision avoidance problem as a multi-agent framework, where agents generate dynamic movements while lea
stantly changing environment? To bridge this gap, we propose an adaptive reward function that utilizes visual attention maps to
nsists of three steps: the first step involves reward improvement update, the second step projects the policy to the neighborhoo
artifacts. For instance, a quadruped may exhibit different gaits when commanded to move forward or backward, even though i
op an effective object-to-arm task assignment strategy for minimizing the cumulative task execution time and maximizing the du
mous driving, large-scale expert demonstration data are available. We argue that extracting expert policy from offline data to gu
ensional observations from camera images, along with a cropped and rectified image patch centered on the current position wi
f the spatial representation, but are memory intensive and may vary significantly between scenarios, resulting in inferior genera
the true support surface in the presence of high grass or other penetrable vegetation. In this work, we present the semantic po
here the arm serves beyond manipulation. Since the system has 15 degrees of freedom (twelve for the legged robot and three
mework to generate dynamic motions for quadrupedal robots. First, a gradient-free evolution strategy is employed to discover s
. This mapping is then utilized in an actor-critic framework using Bayesian optimization (BO). The proposed approach is applica
ocomotion gait or a jumping motion. Instead of a hierarchical or mixture-of-experts architecture, we train a single composer polic
d is capable of a variety of human-like motions. Our framework is with a hierarchical structure. At the low level is a motion-cond
alistic motions, which will be more severe for robots due to different body structures and physical properties. Deploying such a
In this paper, we propose a novel deep reinforcement learning frame-work to enable a quadrupedal robot to walk with impaired
ake proactive actions. Maintaining the original motion patterns could lead to disastrous consequences, as the controller may pr
mework, Cascaded Compositional Residual Learning (CCRL), which learns composite skills by recursively leveraging a library
ocess. Our key insight is that stiff responses to perturbations are due to an agent’s incentive to maximize task rewards at a
Our framework first develops a conventional model predictive controller (MPC) using Differential Dynamic Programming and Ra
manage to bring both aspects of adaptive behavior to quadruped locomotion by combining RL and MPC while improving the ro
nd control systems to very low tolerances. This restricts them to tightly controlled settings such as a predetermined obstacle co
se a risk sensitive locomotion training method employing distributional reinforcement learning to consider safety explicitly. Instea
the generalization and robustness of locomotion policies. However, these methods are hard to handle uncertain scenarios such
PPO) to good performance. Prior works rely on hyper-parameter tuning on the weight of the reward shaping terms to obtain sa
marily address challenges underneath the robot, relatively little research has investigated legged mobility through confined 3D s
r bipedal locomotion push recovery, where the robot experiences unexpected disturbances. Existing recovery strategies often s
ed to take into account nonlinear Divergent Component of Motion (DCM) error dynamics that predicts the future states of the rob
sociated with the step-to-step dynamics. Here, the forward invariant set can be loosely interpreted as the region of attraction for
near-flat ground using a relatively simple, model-based controller that utilizes a novel combination of several interesting design
evious works, retains its validity and coherence during flight phases, and, therefore, can be used for planning running and jump
namics of bipedal locomotion from existing locomotion data. This reduced dimensional state representation is then used as state
ed and estimated. This allows for both position and velocity updates from kinematic measurement. Additionally, the use of kinem
reinforcement learning-based controller is trained to fully leverage the new hardware capabilities to balance and steer the robo
hen the thrust-to-weight ratio (TWR) is small. The robot is designed with a pogo-stick leg and a quadrotor, and thus it is named
itself as a latent contact state-estimator. Through a systematic ablation study, we show that the combination of phase observa

on of dynamic balance during trot gait based on the spatial relationship between the center of mass (CoM) and support area in
ar videos; 2) optimize a dynamically feasible reference trajectory for the robot offline that includes body and foot motion, as well
raverse it. In this paper, we introduce a new mechanism called the Carpal-Claw, which enables quadruped robots to negotiate
for investigating dynamic locomotion modes with extended flight-phases. Instead of gyroscopes or fly-wheels, the system uses
ance for each new system. Alternatively, attempting to train a single policy to accommodate different robot sizes, while maintain
ots, assuming a flat terrain for the purpose of the analysis. Utilizing mathematical models based on spiral theory, we examine th
red to standard legged robots. In this work, we explore pedipulation - using the legs of a legged robot for manipulation. By train
orld model with a policy network. We train a differentiable world model to predict future states and use it to train a Variational A
a-driven error model into traditional MPC for quadruped robots. Our approach leverages real-world data from sensors to compe
oped for trajectory planning purposes. At the higher level of the control architecture, two different model predictive control (MPC
on optimal control, covering a broad range of velocities and gaits. These reference motions serve as targets for the RL policy to
ns are described. Second, the condition for stabilizing zero dynamics is formulated as the time integral value of control input to
e dynamic characteristics of the original limb is proposed. The performance of the model is experimentally validated using an e
ng a reduced-order soft anchor model, the optimization-based motion planner generates the periodic reference trajectory. The

either sparse, insufficient for footstep planning, memory intensive, or too slow for dynamic humanoid behaviors. In this work, we
down). Meanwhile, these transitions can instantaneously change the system’s state, possibly causing perturbations to be m
g the quadrupedal platform CyberDog2 to achieve dynamic balancing and acceleration on a skateboard. An optimization-based
solves kinodynamically-aware trajectory optimization as the low-level motion planner. Through quantitative analysis of the expe
ng method to train the neural network to perform the vector propulsion control. Our approach coordinates the legs and propelle
y in multi-story environments. This study proposes a method aimed at addressing this limitation, employing reinforcement learn
locomotion. However,within the field of legged robot dynamics, the design of the spine and understanding how it positively influ
, we present CAMA: a vision-centric approach for Consistent and Accurate Map Annotation. Without LiDAR inputs, our propose
asets by simultaneously addressing several issues, such as environment diversity, motion patterns, and sensor frequency. It us
reased computational resource consumption for image reconstruction. To alleviate these issues, we propose a novel spatial-aw
ns can provide robustness to domain shifts, and, on the other, that fused information from sequences of images improves perfo

wever, this approach struggles to ensure accurate distance-based image similarity representation, particularly when training with
ents. Despite numerous research efforts aimed at improving VPR performance by addressing challenges such as illumination va
vironmental changes. To address this challenge, we propose a novel multi-modal visual place recognition method by incorpora
enchmark. However, existing fusion methods ignore modal interaction and utilize very simple fusion strategy, which suffers from
vices to provide accurate 3D motion tracking remains challenging. This is exacerbated by data transfer (between different proc
emantic occupancy, thereby circumventing the traditional need for concurrent estimation of ego poses and landmark locations. W
graph optimization. By regarding the neural feature volume as pose graph nodes and the relative pose between volumes as po
ntroduce an algorithm designed to achieve accurate self- supervised stereo depth estimation focusing on nighttime conditions.
lso incorporate lidar data which adds strong geometric constraints on the depth and surface normals. We exploit the trajectory
on or long-term map maintenance. In this paper, we propose a novel mapping, localization, and map update system based on g
l plays an important role for creating clean map. However, conventional dynamic objects removal methods usually run offline. T
novel multi-object mapping pipeline that does not rely on 3D priors. Given only monocular input, we use neural radiance fields to
in bottleneck of current CPU-based PVM, such as OctoMap, is determining voxel grids with occupied and free states using ray-
data. To mitigate the influence of illumination variations and the overfitting due to the dataset diversity, a series of data augmen
To this end, we introduce a novel, image-based self-supervised learning method for traversability prediction, leveraging a state-
enge for estimating traversability. This study highlights three primary factors that affect a robot's traversability in an off-road env
obile robots operating in unmapped locations. However, algorithmically defining these navigational patterns is nontrivial. This le
eatly affect VPR efficiency and precision. Existing methods overcome this by exploiting powerful yet large networks, leading to s
processing cost of BBS in 3D space, we propose an efficient roto-translational space branching. Furthermore, we devise a bat
epancies between instances to efficiently remove dynamic objects preserve more details of static map. It contains the Instance
ns. However, the existing methods are limited in two aspects: 1) poor generalization to unheard sound categories; 2) sample in
ent instance segmentation masks. 3DIML significantly improves upon training and inference runtimes of existing implicit scene
es, employing them as pivotal factors for enhanced robot loop closure and localization. Traditional geometric feature-based loca
 ces. To this end, we propose a localization system based on Block Maps (BMs) to reduce the computational load caused by ma
surface feature-based localization method that uses only GPR measurements with a known subsurface map. An efficient featur
In contrast, mapping methods based on LiDAR scans are popular in large-scale urban scene reconstruction due to their precis
ields. We project LiDAR intensity returns into an image, and present a novel image processing pipeline that produces filtered im
ses SVGD to collectively update particle states with gradient and neighborhood information, which provides efficient particle sa
gistration factors enables a smooth fusion of sensor ego-motion estimation and map-based trajectory correction that results in r
address this problem, this paper presents Same Place Opposing Trajectory (SPOT), a technique for opposing viewpoint VPR th
a reference image and matches to the 2D features extracted from the destination image by the coarse-to-fine 3D matching. Our
ental conditions. This renders them impractical for many robotic applications in which robustness in a wide variety of conditions
evisiting previous locations to perform loop closure, and second, the considerable memory resources required to maintain poin
e estimated depth as a reference. However, these existing methods neglect the influence of the depth uncertainty introduced b
s) that 1) leverages multi-modality by cross-attention blocks between vision and LiDAR features, and 2) includes a re-ranking st
ed to address scan-matching degradation. Our method enables degeneracy-free point cloud registration by solving constrained
o capture foundational semantic details that include the distinctive attributes for unique place identification. To address this prob
me-to-frame data association. In this letter, we propose a robust and consistent frame-to-frame LiDAR-inertial navigation system
r registering to the LiDAR map. Due to reconstruction errors or partial coverage of the camera scan, the reconstructed semantic
ancies between the wayfinding map and the real world make it hard to use conventional localization methods. Our objective is t
in inaccurate poses and large outliers. In this work, we introduce a novel vision Transformer based absolute pose regression m
camera pose estimation. Each stage requires different image features, with global features compactly encoding holistic image in
training data. We propose to leverage Neural Radiance Fields (NeRF) to generate training samples for SCR. Despite NeRF's e
alization problem. Nevertheless, obtaining sequences of labelled data is much more expensive than collecting isolated images,
er, current deep learning-based methods suffer from two major problems: poor generalization ability and catastrophic forgetting
for estimating the state of the robot's trunk. Leveraging joint encoder and IMU measurements, our Kalman filter is enhanced th
place recognition methods. P-GAT uses the maximum spatial and temporal information between neighbour cloud descriptors â
. For close-in-time images, where scene changes are minor, place recognition can be successfully managed with the recent tra
ng radiation, impedes its translation to clinical practice. To address this challenge, we propose a localization method that achie
ch to address these problems, which combines BEV (Bird's Eye View) feature extraction, coarse-grained matching and fine-gra
ely and efficiently implementing cross-modal localization. To solve these problems, a novel pipeline termed LHMap-loc is propos
ural network to learn a discretization-free distance field of a given scene for localization. In contrast to prior approaches, we dire
we propose a sequence-based framework for LGPR to address the aforementioned issues. Specifically, we first introduce a tra
ers from long-term drift. This paper proposes a framework to increase the localization accuracy by fusing the vSLAM with a dee
on. We propose EffLoc, a novel efficient Vision Transformer for single-image camera relocalization. EffLoc's hierarchical layout,
ch leverages semantics in odometry. The semantic information for the whole scan is timely and efficiently extracted by a 3D co
uncertainty modelling techniques are often coupled with a specific APR architecture, resulting in suboptimal performance com
es directly learn and regress 6 DoF pose from an input image. However, these methods do not fully utilize the underlying scene
e 3D poses. In this study, we propose four novel modules that improve localization precision by imposing the existing methods
ps. This paper proposes VOLoc, an accurate and efficient visual place recognition method that exploits geometric similarity to d
ork introduces a SNN for Visual Place Recognition (VPR) that is both trainable within minutes and queryable in milliseconds, ma
pose. In this paper, we theoretically analyze the influence of overlap and the solvability of 17-point algorithm. In addition, we sh
environments, rendering most approaches brittle to robust real-world deployment. In this work, we develop a universal solution
high-precision, millimeter-level localization without instrumenting, marking, or modifying the environment. The key innovations a
dress this issue, we design a real-time tightly-coupled Neural Radiance Fields (NeRF)-aided visual-inertial navigation system (V
insight from most nocturnal scenes is that the static and bright streetlights are reliable visual information for localization. Hence
ibration of the rotational extrinsic parameter between the GPS reference frame and the VIO reference frame. The behind reaso
ces on board. This work presents an online and onboard approach for localization in floor plans annotated with semantic inform
G massive multiple-input and multiple-output (MIMO) testbed and user equipment, audio recorded by 12 microphones, and acc
n and ignored biases of the IMUs. Instead, we study the most general case of unknown biases for both IMUs. Besides the deriv
widespread adoption in mass market applications. In this paper, we aim to use only a monocular camera to achieve comparable
g of several filter parameters and the fact that changes to how the Inertial Measurement Unit (IMU) is oriented necessitate mod
closed-loop control. The first FLS combines ultrawideband (UWB)-based range measurements and VIO to estimate the robot tr
is often due to the careful implementation of exception code for a multitude of corner cases along with significant skilled tuning
ility of real-time multipath determination, especially in challenging conditions such as when there is no direct line of sight. We ac
ntrast, the proliferation of smartphones and WiFi-enabled robots has made crowdsourced radio maps – databases pairing loc
marks with the same object category, followed by inlier extraction using RANSAC or brute-force search. This approach become
stance observations of carrier phases in the global navigation satellite system (GNSS). However, when such likelihood function
ical, topometric semantic map representation that uses polygons to demark areas such as rooms, corridors or buildings. This r
and bounce dynamics upon contact with the ground. In this study, we introduce an innovative approach that combines a multi-
velocities, is calculated taking into consideration its kinematic model which provides a constraint between the velocity measure
ation in low-visibility conditions (dust, smoke) where lidars and cameras typically fail. With the advent of high-resolution Doppler
ure detection) and pose refinement (i.e., point cloud registration) problem both in isolation or combined. However, few have exp
s data association. However accessing the marginal covariance of the state variables of interest in optimization-based VINS re
e PL allows for tuning the coupling parameters of the DMP, tailoring the interaction to each participant personally, and allowing
gorithms with large amounts of prior data, keep learning continuously, and ensure safe navigation in applications. Conventional
tegrate a forgetting factor to refine a variational sparse GP algorithm, thus enhancing its adaptability. Subsequently, the hyperp
entation model to tackle these challenges effectively. The proposed representation model combines a PointNet encoder and a c
C). This paper provides an answer by introducing a new framework called Actor-Critic Model Predictive Control. The key idea is
een evaluated for open-loop prediction, we show that they can be used to parameterize autoregressive closed-loop models with
ing world. To address this issue, we introduce GenDOM, a framework that allows the manipulation policy to handle different de
hallenge, we present a sim-to-real reinforcement learning approach for training general box pickup and carrying skills for the bip
ng approaches offer higher model capacity and better generalization but require large amounts of state-labeled data. This paper
dictive control (MPC) continually re-plans at each time step to remain robust to perturbations and model inaccuracies. Howeve
domization addresses this problem by randomizing simulation dynamics to train policies that are robust towards modeling error
se advances, robotic RL remains hard to use. It is acknowledged among practitioners that, separately from the fundamental tec
mics, and 2) physical limits on the set of possible motions, in the form of nonholonomic constraints. Crucially, we do not assum
ge prior data into the robotic learning pipeline. However, these methods have a "type mismatch" with video data (such as Ego4
regolith are prone to flow downward under small disturbances, making it challenging for autonomous rovers to traverse. Moreo
he exploration, and makes the algorithms substantially less rewarding. In this paper, we propose a method to construct a boun
problems. In this study, we propose and investigate planning high-quality action sequences for solving long-horizon tabletop rea
the tasks. In such scenarios, the object must be reoriented and re-positioned into intermediate poses that facilitate accurate pla
iated forces (force goal). Together, these two components form a visual-force goal. Prior work has demonstrated that deep mo
lation, is identified and addressed. Specifically, a tactile-sensorized tool stick that well fits in the narrow space is utilized to prov
. Grasping with multiple grippers can create closed loops between the robot and DOO, and If an obstacle lies within this loop, it
 excellent performance and generalization capabilities in articulated object manipulation. However, a significant challengeremai
movable objects to rearrange, where to move them, and how it may do so. In recent work we introduced E-M4M, a graph searc
plan a minimum-time rapid trajectory that can evade obstacles and intercept the projectile, and (iv) execute the planned trajecto
he effect of SD maps for real-time lane-topology understanding. We propose a novel framework to integrate SD maps into onlin
smooth surfaces. This is particularly problematic in large-scale applications with sparse inputs, as is common in robotics use ca
memory, and the catastrophic forgetting phenomenon. To counter these challenges, we introduce the Robot-centric Implicit Map
nges that can affect scene appearance and shadow regions, causing a degraded pose optimization process. In this paper, we p
drics to represent diverse primitives with only 10 parameters and propose the first end-to-end learning-based framework, namel
ddress this issue, we propose ERASOR++, an enhanced approach based on the Egocentric Ratio of Pseudo Occupancy for eff
nd real-time capability even on edge computers. Specifically, we propose a novel hierarchical hybrid representation that leverag
es as input and outputs dense 3D shapes and 9-DoF poses (including 3 scale parameters) for detected objects. The core idea i
e propose a novel topological representation of an environment based on `image segments', which are semantically meaningfu
t this is enabled by the ability to represent the environment abstractly as interconnected navigational behaviours, e.g., “follo
anning while a small grid size lowers efficiency and causes gaps and inconsistencies. To tackle this problem, this paper genera
e static world to ensure safety. In this paper, we address the problem of jointly estimating moving objects in the current 3D LiDA
ating normals from a single scan in a robust manner poses difficulties. In this paper, we address the problem of estimating norm
ese two types of sensors is poorly performed in many existing frameworks, leading to unsatisfactory mapping results, mainly du
potential problems caused by the errors of maps and a lack of uncertainty quantification, an adequate uncertainty measure for
cy state of regions instead of the traditional ray casting method. Secondly, we introduce an efficient on-tree update strategy on
ework based on 3D region-wise hash map structure (RH-Map) is proposed, consisting of front-end scan refresh and back-end
d images resembling those taken by conventional cameras. Nevertheless, the typical effective global refinement techniques em
s, novel jet propulsion mechanisms, and multifunctional spherical shell. Its architecture is modular, appendage-free, and largely
h has been developed, featuring the capability of adaptable bistability. Through tuning the bistability, the robot can acquire two
this research is to design an alternative manipulator for deepsea suction sampling. Instead of relying on heavy hydraulic rigid m
ping new hardware and algorithms for autonomous underwater manipulation is challenging. Simulations do not capture the full
ven by a piston variable buoyancy system (PVBS) is proposed to reduce the excessive underwater drag caused by aerial struc
ontrol at limited speeds, mainly within a confined environment. To address these limitations, we designed Snapp, an integrated
ot inspired by the whirligig beetle, with the design goals of good stability, adaptability, and maneuverability. First, the design of t
keeping capabilities, low disturbance of the substrate, and precise position control. The sediment sampling system has been de
ize underwater terrain's deformable properties is hard. Here, we propose a method to perceive the interaction forces and feed t
ngly challenging environments, necessitating methods that are able to safely navigate robots in real-time. Though, most solutio
ale ambiguity. To allow prior inputs of arbitrary sparsity, we apply a dense parameterization method. Our model extends recent
water conditions. An approach is to employ sonar or laser sensing to acquire 3D data, however, the data is not clear and the sen
spondence capable of withstanding viewpoint variations. The inherent complexity of the underwater environment stems from th
and reliable outcomes, especially within the demanding underwater environment. Existing solutions for estimating uncertainty a
g the effects of the medium and neural scene representations. Our approach models water effects as a combination of light att
water cave scenes. It contains pixel annotations for important navigation markers (e.g., caveline, arrows), obstacles (e.g., groun
he CUREE AUV, uses four hydrophones to determine the bearing to biological sound sources through beamforming. These obs
feature space among nodes varies significantly. To address this problem, we innovatively represent the LiDAR points in a two-
s of surge motion, sway motion or a combination of both on a cable. Experimental studies are carried out with eight cables of va
s condition of a UV operating at or near the free surface. We report a Monte Carlo-based hybrid model approach for calculating
 me, the cross-domain window, i.e. the vehicle is at the wave crest and heading downstream, is characterized and defined. The
n complex environments such as in close-proximity to coral reefs. However, they are inherently power-inefficient and produce si
ynamics and control constraints. Nevertheless, achieving robustness in changing environments like oceans and rivers, which ar
ve reliability and reduce communication pressure, our unique design enables the robot to rely on a kind of propulsion device nam
ens, e.g., wide-angle field-of-view (FOV), the refractive effects at the enclosure (water-acrylic and acrylic-air) interfaces, and off
rk, we curate and release DotsonEast, a semi-synthetic MBES registration dataset constructed from an autonomous underwate
s is challenging, however, because of the lack of available standardized software frameworks and pipelines. Previous works off
heoretical results, we show it is possible to design the behavior of the MRS by the selection of a relatively small set of paramete

cking capabilities to recharge in situ at Wave Energy Converter (WEC) with dock recharging stations. However, the power gene
where multiple ASVs are navigating in the presence of static obstacles and strong currents, due to the complex interactions. T
ainty prediction of future poses into the planning process and choosing actions that reduce uncertainty. However, performing b
new type of remote-controlled hydrofoil marine vehicle, named Sea-U-Foil, which is suitable for different marine scenarios. Sea-
me while maximizing classifier accuracy. We introduce a novel active perception framework for multi-view adaptive surveying a
pulators face limitations in terms of size, accuracy, and scalability. Therefore, in this letter, we present a novel electric underwa
ces for their respective types - if not even the most affordable ones on the market. The software runs with low computational p
tively untouched. We address the state estimation problem that arises when trying to rendezvous a chaser AUV with a dynamic
clearance. Area water clearance is important especially for path planning and obstacle avoidance. Line water clearance is critic
or locating and tracking a DS exist, but most rely on expensive acoustic sensors, or are vision-based, which is significantly affe
y. A deep learning-based semantic segmentation algorithm is applied to detect the navigable region in camera images, and the
dex n and radial pixel coordinate) and a virtual pinhole model (as a function of n). We derive the self-calibration formulation leve
s of IMU orientation propagation are used to efficiently transform magnetometer measurements between frames producing rela
stem into an optimization-based visual-inertial SLAM. In our proposed Acoustic-VINS, the reprojection error, IMU preintegration
e, employing volumetric techniques to create grid-based submaps of the unoccupied water volumes through images collected fr
y of the robot but significantly complicates the planning problem by introducing binary decision variables. To address this challe
enhance the safety and efficiency of elevated wood structure inspection using the resistography method, which involves drilling
king problems. By utilizing the mathematical efficiency and compactness of dual quaternions, a unified representation of the UA
overcome these issues, we propose a new fully-actuated multirotor unmanned aerial vehicle platform capable of translational m
ensors, being both cost-effective and lightweight, are capable of sensing contact information including force distribution, as wel
s a common method of take-off for birds, jumping take-off has a great ability to adapt to different terrain and high energy efficie
g controller to avoid rotor saturation and mitigate overshoot during transition between free-flight and perching. To enable stable
ve connection mechanisms such as rigid links or cables for transportation and manipulation tasks. However, these systems are
h experimental tests to accurately model the perching manoeuvre. The resulting model is used to develop a realistic hardware-
e rotorcraft falls into its wake flow. This is known as the vortex ring. Therefore, the safe descent velocity of consumer UAVs is u
meters available to an established communication-less force-based controller. First, we find the closed-loop equilibrium configu
i-modal aerial robot system that integrates tensile perching for energy conservation and a suspended actuated pod for data col
mely challenging to ensure the high-precision performance of aerial manipulator. Building upon the philosophy of disturbance re
able sensor information onboard a MAV, such as accelerometer data, pose estimate, and throttle commands, which makes our
mic coefficients that leads to polynomial differential equations of motion amenable to semidefinite programming optimization. T
d on proximity to surfaces. Unlike rotors, they have no fragile and quickly spinning parts that have to avoid those surfaces; takin
d dextral turns, respectively. The addressed problem generalizes the classical Markov-Dubins problem and the asymmetric sini
 etary value terms as a Human Value at Risk (HVaR) measure. Additional operating and UAS Capital Value at Risk (CVaR) cos
ss proper relative tangential velocity. However, quadrotors' attitude and relative velocity errors would become significant when th
omputed per the trajectory. To demonstrate this, a Monocopter in the form of a Stable Invertible Coaxial Actuated ROtorcraft (S
e often used for placing NDT sensors in contact with a surface to be inspected. Aligning the NDT sensor and the work surface w
pitch rotation without mechanical constraints like cable winding, significantly enhancing its capabilities for various tasks. To con
manifold. Our particle filter with stable embedding enables accurate state estimation by maintaining estimated state values in clo
oop controller —an abstraction that, in theory, constrains the optimality of the trained controller, necessitating margins to coun
cond mode of locomotion, tightly coupling the attitude of the robot to the rolling cage. The benefits are precise rolling and turning
ctuated joints that increase the topology and control coupling, making the design process more complex. We propose a co-desi
tion of the quadrotor. In this paper, to keep the quadrotor stable even under suddenly drastic disturbances, a novel control fram
with a performance figure of merit (PFM) that takes into account both the aerodynamic-torque control input and attitude-error s

rk surfaces. When interacting with such surfaces, the coupled gravity compensation and interaction force generation of undera
ly placing vertical poles, blending load transport with precision. Geranos minimizes the effects of the poles' large moment of ine
teracting with the environment. Hence, it is challenging to design a generic and robust control strategy adapted to various robo
ollaborative transportation of many payloads by many drones remain to be considered. In this work, we formulate the collaborat
es of freedom (DOF). Therefore, existing work often compromises between mechanical complexity and the controllable DOF of
e deflection angles of the platform, which uses only onboard IMU measurements. The state estimator is an extended Kalman fil
OAR path planner which addresses the complexities of evolving risks during missions. It offers real-time trajectory adaptation w
otational power with a predetermined gear ratio, in contrast to traditional mechanical gear-based transmissions. This approach
een their low specific shear stresses which restrict gen- erated holding torque, limiting many applications. Here we show how co
ionship between the panels are obtained. Bionic foldable wings utilise shape memory alloy to drive wings to fold, and embedde
on the ground. However, creating scalable and steerable bristle-bots remains a challenge. Here, we focus on developing a new
, the scarcity of commercially available VSMs requires that such systems are mostly designed in-house. We propose a new typ
mentation to evaluate the concept viability. We designed and characterized Shape-Retention Bellows (SRBs) that constitute the
gn parameters affecting the dynamic performance of the actuator and a full specification is provided. Fundamental parameters
ehabilitation is designed in this paper. The exoskeleton combines gravity compensation with left-right arm switching function th
ard as a human interface and for workspace optimization. In this study, we propose a novel flexible omnidirectional driving gear
ogram remote center of motion (RCM) mechanism to implement reconfigurable features. With the RCM mechanism, the gripper
on state of modularized body elements. Robots and we can use the ALDMag to attach and detach mechanical and electrical co
hanisms. Using screw theory, the comprehensive kinematics of the proposed deployable kirigami-origami robot were analyzed,
s paper. Owing to its unique foreleg design, the robot could operate either in the quadrilateral four-wheel mode, collinear four-w
challenging due to increased complexity, heavy weight, and vulnerability to impact. In this paper, a novel leg and toe mechanis
ovided by a single autonomous sailboat, this scheme is not suitable for the floating equipment with large mass. This paper prop
An water wiping experiment confirmed that the finger wiping motion removed 35% more water than the translational wiping moti
nsemble and implement a control strategy consisting of two nested loops, the inner one performing high-bandwidth joint torque
s, and unfold to cover larger areas efficiently. The folding pattern has a deploy-to-stow ratio of 3 in the width dimension, and a k
ced. Regarding the structural anisotropy, the rotational axis of the motor is nonparallel to both principal axes of bending stiffnes
the tip’s eversion naturally conflicts with grasping, and the addition of grippers at the tip would limit navigation inevitably in
al walking and wing-assisted incline running (WAIR) to climb steep surfaces. We took inspiration from nature and designed a m
performance of eSKT is far from that of natural skeletal muscle tissue. In this paper, an optimized design method for eSKTs ha
ces may limit patient participation and hinder motor learning. To address these challenges, this work proposes the use of two lo
e been implemented to improve performance. Firstly, the magnetic catheter model and hybrid visual/force servoing control sche
tic properties to the target object, or drive the robot to contact and then manipulate the object, both probably inducing physical
handheld surgical dissector also provides suction and visualization for surgeons to enhance effectiveness. The steerable sectio
c platforms to overcome the limitations of manual surgery. However, specific tasks, such as instrument insertion through the tro
litional base controller within a hybrid volitional control framework (PVI-HVC). The gait progression estimation and biomechanic
ging, methods for autonomous robotic guidance have been proposed. In such systems, the robot’s cognition is often govern
mplex and narrow environment of the colon. Moreover, these robots are typically equipped with a single camera, limiting their ab
e than 6,700 systems in current clinical use worldwide. To accelerate research on robotic-assisted surgery, the retired first-gen
reducing the sonographer's physical efforts. In this paper, we address the control of the interaction of the probe with the patien
nce hand-eye coordination and instrument triangulation in narrow and shallow rectal spaces. Key innovations include the hybrid
s. Robotic ultrasound systems have emerged as a promising solution, offering standardized acquisition protocols and the poss
trial-and-error experience with the background that additional imaging modalities such as iOCT remain inaccessible in most op
scan trajectory on a generic atlas to the current setup for different patients, particularly for thorax applications with limited acou
d size of the heart, lungs, and ribs differ between individuals, which makes it difficult to find the optimal position of the US probe
is still challenging. To address this challenge, this paper proposes the DopUS-Net and a vessel re-identification module that lev
esthetics. In this study, a non-invasive image-to-patient registration method for navigation/robotic CMF surgery based on intraop
urrent commercial electrical 6-A F/M sensors to adapt the electromagnetic environment in the operating room, and status chan
he epiretinal membrane (ERM) which have an average thickness of only 60 μm and 2 μm, respectively, making it challenging
mittance control algorithm incorporating a model-free adaptive nonlinear force control and fuzzy control algorithms is proposed t
and wrist motions. Also, for the better outcome of surgery, the proposed controller has a force feedback function which is not a
cognizing the irreplaceable role of a skilled surgeon. This paper presents a collaboratively controlled robotic system tailored for
gate in densely populated environments. Hence, in this paper, we design a new disinfection robotic system consisting of a mob
ding may occur during lymph node removal caused by collisions between surgical instruments and delicate blood vessels (arter
s and needles outside the FOV. This paper presents a laparoscope-holding robot system capable of employing both vision-drive
e constant curvature assumption has shown modeling errors. This causes the current geometric kinematic model to become un
objective of this paper is to design a haptic interface that can simultaneously capture the force-position information of the interv
d-loop design, enables two-degrees-of-freedom transmission from the proximal end of the robotic arm to the distal end. This fea
mplifier was assembled after machining the components using a 3-axis computer numerical control machine. A flexible hinge wa
eventing them from accessing hard-to-reach areas. This is due to the fact that under a global magnetic field, fixed magnetizatio
on of the external magnetic field, the spiral structure on the outer surface of the shell can actively moves the robot back and fort
s surgical manipulation deep in the brain. Taking into account the uncertainty of the control model, an image-based visual servo
ting (FBG) sensors for shape sensing has shown great potential in estimating the CDM's tip position and subsequently reconstr
ime, and cannot be vectorized. This affects the behavior and result of downstream methods in control, learning, workspace est
novel vibration module inside a capsule robot for colonoscopy. The capsule’s motion was controlled by applying an externa
ng magnetic field applies torque on the swimmer and makes it rotate. The shape of the swimmer, combined with the rotational m
ts, resulting in surgeons needing to switch attention between performing surgical procedures and moving endoscope camera. A
analysed for features that may be indicative of biomechanical changes in the lesions. Different combinations of the features we
e development of cancer. Meanwhile, the use of interventional Ultrasound has gained popularity, due to its well-known benefits
ts and the small workspace and singularity issue of parallel robots. Besides, by analyzing the workspace of the robot, the equat
en paid to the impact of non-predicted and sudden movements of the patient and the environment. Therefore, analyzing environ
 d tremor cancellation, hand motion scaling, and safety constraints that enable them to perform these otherwise challenging and
t of self-adaptive motion scaling has been proposed in previous work. However, previous frameworks required extensive manu
in more complex, potentially risky interventions. To surmount these limitations, this study introduces a flexible capsule endosco
oupling and impacting tissue during the insertion and coupling procedure. In this paper, we proposed a wireless MAGS consisti
iotic-resistant strains, necessitating alternative treatment strategies. Photodynamic therapy (PDT) emerges as a promising pros
these problems, a flexible tactile-sensing gripper (FTG) is designed in this study. Firstly, a catheter grasping model is construc
xible robots for accurate sensing results. Another challenge lies in online acquiring external forces at arbitrary locations along th
izing oncologic outcomes. This letter presents an innovative autonomous system for tumor resection (ASTR) and conducts a fe
gned and successfully demonstrated a fully functional shared control scheme for an autonomous tensioning and tele-cutting of
, and need the surgeon to modify their operative skills and strategies. In this paper, we introduce a modular robotic system for
ramework for vitreoretinal surgery must be considered as an essential component. Such a framework, ideally, considers the com
m that addresses these challenges by using magnetic fields and computed tomography imaging to precisely navigate minimally-
re real-time and precise tracking in the deep-seated region over abdominal organs, while considering full-wave propagation. To
e for this issue. In this work, we proposed a dual-segment slender surgical robot arm for TNS, which is a Co-Axial Slender Tub
dy, we investigated robotic autonomous guidewire navigation methods. An electromagnetic system was used to track the real-tim
d flow. However, clinical micro-catheters that deploy embolic materials used today face limitations, primarily their rigidity and the
esign space by which the deformation-behavior space of a TD-CM can intuitively be analyzed via its geometrical design param
even with robot assistance and medical imaging guidance due to needle deflection caused by contact with soft tissues. In this
manipulated robotic arms that can be equipped with a variety of tools, a common scenario in industrial but also clinical and resea
rohand is presented for fabricating magnetic soft microrobots. Our proposed system can automatically pick and place compone
blocking force, and a free displacement of 30 μm is experimentally characterized. The actuator leverages bare NiTi film (∼
urrent microrobot design, where the underlying intelligence and subtlety spurs related modern artificial systems. Inspired by air
hile harnessing opposing torques on the magnets to induce internal stresses needed for gripping. Our microrobot design featur
We propose the microswimmer composed of a self-propulsive disk-shaped module with multiple channels using biofuel cell (BF
aped, tendon-driven continuum microrobot designed explicitly to navigate through brain tissues. The microrobot has a cross-se
king the advantages of geometry/wettability heterogeneity of the microrobots, etc. Using the step-out behavior of helical microro
otic mi- crosurgery platform that generates explants of specific tail tis- sue from growing zebrafish embryos, a popular model or
owever, human operators seem to be able to complete the task more successfully with an inherent motion and control strategy
ace consumption with better dexterity, higher precision, and well biomechanical compatibility. In this paper, we propose a novel
rectly infer depth from microscopic images and suffer from poor resolution. Alternatively, micromanipulation tasks obtain accura
body-based drugs. However, conventional OET platforms are limited by issues such as low autofocusing accuracy, restricted im
sed method consists of several steps. An attention mechanism improved (AMI) tip localization neural network is proposed to de
and fabrication methods. Here, we propose a 3-RRS micromanipulator with an origami-inspired spherical joint based on the PC
his method can construct compact and functional miniature origami mechanisms through planarized fabrication and folding ass
ed, the forces and torques have yet to be systematically investigated within biomedical context from underlying physical princip
as robotic intracytoplasmic sperm injection. However, conventional image processing methods have limitations in analysing sm
k encompasses an active microscope-camera system equipped with both auto-calibration and multi-scale visual servoing capab
ulator. Even though the focal depth of the microscopy was limited, the proposed method functioned the three-dimensional posi
. The merging of the patches into a single image is done through a Generalized Procrustes Analysis Iterative Closest Point (GP
ccurate tracking of the target sperm and precise alignment between the sperm tail and the micropipette. Manual sperm immobi
llow a “detect-then-segment” paradigm. However, due to sperm’s slim shape, their segmentation suffers from large
e been devoted to sparse image reconstruction while little attention has been paid to sparse sampling trajectory design required
icro tissues, suffers from mechanical damage and low efficiency. Several techniques have been introduced to manipulate cells
ion imposes a strict constraint on both functionalities. This study addresses this challenge by substantiating the feasibility throu
benchmark synthetic dataset, Sim-Suction-Dataset. It comprises 500 cluttered environments with 3.2 million annotated suction
e Markov decision process, a novel formulation that models task-level planning where uncertainty pertains to object- level attrib
nning. To bridge the gap, this paper presents a motion planning method capable of generating high-quality, safe, agile and feas
n which a robot has to open doors to navigate rooms and search inside cabinets and drawers to find target objects. These new
cus is on fast online planning in response to new and changing environments. Our main contribution is a whole-body constraine
(AMPC) is an approach for controlling robot motions while adapting to uncertain or changing dynamics. However, most of the e
pability-aware mobile manipulation approach powered by an online grasping pose fusion framework that enables a temporally c
for cooperative control of legged robots and robotic arms. This work decouples the control of the robotic arm and the legged ro
ation in unstructured and cluttered environments, such as households, remains challenging. In this work, we introduce an active
hanges. In this paper, we present a self-supervised learning approach that actively models the dynamics of nonlinear robotic sys
tly deform under stress and are constantly changing shape. In this work, we investigate each of these challenges using the task
f such events. These maneuvers are a distinctive feature within the recently established motorsport discipline of Autonomous R
response to environmental changes, such as variations in terrain elevation. To this end, we propose a method for learning terra
anticipate the operating environment a priori due to complexities that arise from numerous factors including imprecision in man
epresents how close the pseudo-inertia matrix is to a given reference on the feasible manifold in the regression problem. The pr
ons of multi-robot manipulation and direct measurement of DLO configuration (state). This work focuses on single-arm manipul
ods do not comply by design with the problem's physical constraints, especially in states that are not densely represented in tra
estimate the non-linear effects of tire-terrain interaction on robot velocities in a probabilistic fashion. Using the mean estimates
ally measured and realized via the optimization of human-likeness metrics. However, the abundance of such metrics and the ab
been a challenge. In this paper, we present MiBOT, a head-worn massage robot that actuates two soft tactors to produce touc
this paper, we build a new dataset, which aims at the long-term prediction with the inconspicuous state variation in history for t
-sensitive. In this paper, we introduce SynthAct, a synthetic data generation pipeline designed to significantly minimize the relia
equipped with optical tactile sensors to acquire multimodal information, including hand positions and postures, the relative dista
hology-specific behaviors and computational cost. This paper presents a framework to efficiently evaluate the suitability of a wo
with a reduced number of muscles. There is also a lack of algorithms capable of controlling over 600 muscles to generate reaso
ctors lead to redundant parameters, large calculations and low efficiency, while lightweight solutions to address these challeng
suring in vivo muscle force is challenging, leading to the development of numerical methods to estimate them. Due to their high
this specific class of robots remains an open problem in the literature. This paper aims to present the first solution, to the best o
ation, and for demonstrating that, for a range of input commands, the catheter can be used as an actuation mechanism for prop
d to compensate hysteresis of a soft actuator. Employing a novel dual pneumatic artificial muscle (PAM) bending actuator, the i
is subject to external forces or even gravity. In this paper, we establish a three-dimensional (3D) kinematic representation of a
planning, control, and calibration for any robot structure. For CRs, the need to avoid obstacles in tightly confined workspaces gr
hat has shown promise in controlling soft robots. However, manually tuning the many hyperparameters in the learned dynamics
e tracking of large continuum objects/manipulators. The physical occlusion of markers can often compromise accurate shape e
thod and proper orthogonal decomposition technique. Based on this model, an adaptive learning control scheme is developed w
using a prominent feedforward control action to preserve the soft and safe robot behavior. To tackle this challenge, this letter pr
d flexibility that is important for wearable devices. In this study, we introduce the control implementation for a FEA to the applic
bodies to the soft robot or inserting soft sensors made of materials different from the robot itself. Both approaches result in chan
sk. To deal with the high non-linearity of the dynamics of soft robots, we deployed a neural network to learn the relationship betw
and-drop soft robot control interfaces by leaving the 2D computer screen for 3D Virtual Reality (VR). VR-Soft IK uses a simulat
g of these tasks, as exists for rigid robots, is still lacking. One of the difficulties comes from the modeling of these robots, which
using kinematics (constant curvature) and mechanics (virtual work). Modelling considers the potential misalignment between th
y estimating its shape. In this work, we present Proprioceptive Omnidirectional End-effector (POE), a tendon-driven soft robot w
ontrol of the system's internal configuration. This method constructs a series of connected models comprising actuator and loco
d hybrids with intrinsically coupled elasticity between various degrees of freedom. In this work, we seek to address the issues o
ough an optical tracking system, while a force-torque sensor quantifies the resultant fingertip loads. Employing this setup, we s
on environment with musculoskeletal models. 20 inverse optimal control (IOC) problems were solved for cost function weight tu
latively flat surfaces such as walls. We present a novel free-climbing robot to bridge this gap through innovations in gripper des
verlooking payload capacity. To address this issue, a novel adhesion foot structure utilizing micro-suction tape (MST) and a det
ons (e.g. arm drag in low Reynolds environments and arm momentum in high Reynolds environments). While thrusters can in t
n planning to high-dimensional systems. In this paper, we propose a gait-based coordinate optimization method that overcome
hybrid mobility robots offer high flexibility in traversing challenging terrains by leveraging the advantages of both wheeled and le
ements are computationally complex and not scalable to high-dimensional or systems with complex nonlinear dynamics. Reinfo
ork has identified a tractable case where the flow is modelled as piecewise constant. Extending this method to allow for comput
s where found trajectories are restricted to only a single homotopy class of a given path. Distinct paths can also be used to guid
ximated optimal action distribution in closed form, i.e., without iterative solution updates, it struggles with the multimodality of the
It has a compute and memory complexity of $mathcal{O}(n)$, where $n = n_{x}times n_{y}$ is the size of the grid, and requires
ble challenges, primarily stemming from the lack of a closed-form clothoid expression. To address these challenges, we introdu
t continuously take on new tasks. In this work, we make it possible to extend key ideas enabling state-of-the-art (SOTA) metho
ore time in areas with high information density. However, existing literature on ergodic search does not consider the robot's ene
as Informed RRT*. However, with the growth of spatial dimensions, the time consumption of randomly sampling the entire state
e lies in the fact that CRs can have multiple possible configurations when in contact, rendering the forward kinematics not well-
one based on spatio-temporal planning, and one based on target swapping, to increase building efficiency. The first method ca
ce, rule-based informed approaches sample states in an admissible ellipsoidal subset of the space determined by the current p
nning times into the range of microseconds and solution rates into the range of kilohertz, without specialized hardware. Our key
nformation and risk in a way that allows for local sampling to generate Pareto-dominant solutions over a receding horizon. We p
tives, emerging problems in robotic trajectory optimization frequently involve dynamically evolving objectives and stochastic mo
ng A* in order to reduce the overall computation and define appropriate cost metrics to ensure winding angle constraints are sa
ocess and shows that the previously published attempt at optimization of these path plans was too coarse to be practically usef
motion planning the bottleneck for achieving reliable and adaptable autonomy. Recently, a class of constant-time motion planne
tion. Our policy uses height and intensity-based cost maps derived from 3D LiDAR point clouds, a goal cost map, and processe
datasets. However, without a prior understanding of what diverse valid trajectories are and without specially designed cost func
zation. In this work, we present an efficient method for approximately covering complex configuration spaces with a small numb
solution quality. This paper presents a novel asymptotically optimal algorithm for generating a joint motion strategy for the pursu
, which is inconsistent with the human intuition that the path segments should be straight in wide-open space due to the absen
as data and triangulates the manifold to ensure its existence in the obstacle region of the configuration space. In this paper, w
rse data structure but retaining its topology. The bubbles also enforce safety directly in the roadmap structure. We first demons
form of a path or waypoints. In this work, we consider the multi-goal problem of visiting a set of pre-defined targets, each of wh
acles is generated, which is then inserted in an OctoMap. The unoccupied areas are identified, thus resulting to a sum of safe l
ask. This formulation often leads to a conservative choice of paths through the gates, as the spatial potential of the gates is not
ck-end that optimizes this path to respect dynamics and input constraints. While there are many different choices for each stag
mous aerial transportation, this paper presents a systematic solution to address these difficulties. First, we present a real-time p
ways that are beyond the reach of existing methods. The trajectory prediction of the ball is achieved by integrating forward the
d by fusing the measurements of a downward-facing camera and a load cell with an Unscented Kalman Filter (UKF). Additiona
f the detected objects by modeling spatio-temporal reward states and coordinating a team of vehicles to collect the rewards. Th
ive tasks. The presented work demonstrates the feasibility of deploying the aerial robot to physically work with the human opera
ng paths involved in the aerial 3D printing process without physically depositing the overlaying material. The flexible multi-DoF a
Arm near the base are implemented as dual-motor joints (DMJs), while the remaining four axes are single-motor joints (SMJs), w

-free to rotor failure without controller switching. To achieve the purpose of the passive FTC, the rotors’ fault is modeled as
hod solely requires the reference trajectory and the robot's kinematic constraints to output kinematically feasible joint space com
-effective transition from testing in a simulation environment to a physical system. However, the existing scale vehicle platforms
ement is partially observable through data and can be even time varying. A data-stream-driven ambiguity set is firstly constructe
e settings where smaller robots may be deviated from their task because of unmodeled physical interactions. This is all the mor
lenges due to errors and uncertainties. To address these issues, the authors propose the use of ILC, a control strategy that mo
e can occur due to the inertia of the manipulator and adjustable equivalent stiffness of the controller. This paper proposes an FF
oller. The top-level planner generates a safety-ensured trajectory for each agent, accounting for the dynamic constraints in the e
drupedal robot in distillation columns, locomoting on column trays and transitioning between these trays with a roller arm. To ad
eplaced to complete different tasks, the structural information related to the robot manipulator and its rod-like end-effector may
enough for high-precision systems to be robust with respect to unknown input-dependent noise and stable under safety constra
hesis (PS) scheme for retractable worm robots. The scheme combines an undulating gait inspired by caterpillars with three-dim
without requiring acceleration signals and the inverse of inertia matrix. Then, based on the USDE framework, an adaptive-gain c
ed controllers of fine stages are designed according to the so- phisticated models and identification techniques used. Due to the
this paper, we further extend the NRIC in such a way that an optimization scheme can be embedded in the control structure. T
control method by intelligently allocating control authority, with the surgeon remaining the leader during the surgical procedure.
nded Cartesian impedance control algorithm is developed, which includes geometrical constraints and enables explicit force tra
ability-preserving fashion, and resourcefully exploit the knowledge acquired during the interaction process. In this article, we pro
ess this demand, we have designed a cable-driven serpentine manipulator (CDSM) that combines force and precision motion c
y in recent years due to computational advancements. It can achieve high performance in specific tasks, but it lacks physical in
onding infinite dimensional Koopman bilinear model over the lifting function is equivalent to the original control-affine system. Fu
uires a significant amount of data, and the irreversible nature of the removal process makes data collection expensive. This pa
model tuning effort, thus resulting in low adaptability and robustness. In this paper, we propose a real-time adaptive multi-conta
Data-driven Robot Input Vector Exploration (DRIVE), a protocol that enables characterizing uncrewed ground vehicles (UGVs) i
ve the underdetermined control and planning problems. The state-of-the-art approaches arrange tasks in a hierarchy and decou
es of the robot through factors of simultaneously manipulatable joints of an object. Based on this factored state space, we propo
utions quickly, (2) reduce wasted compute, and (3) improve data efficiency. The key insight of the approach is to make the mos
es. We develop local plans, closed-loop control algorithms for planning the connection of two structures at desired faces. The g
lable constrained path following. Firstly, based on control points, a recursive computational model and an equivalent planning a
ractical constraints. The algorithm explores the inter-traffic gaps in the time and space domain using a breadth-first search. For
ectory into shorter segments. Each such division introduces a reconfiguration, in which the robot deviates from the reference tr
 the goal while preventing collisions with obstacles. The proposed global navigation scheme is experimentally verified on a quad
re of manifolds makes the CMP problem challenging, and the state-of-the-art methods still take several seconds to find a path a
sed on the Explicit Reference Governor (ERG) framework to ensure the safety of a flying quadrotor UAV. From a small set of e
g algorithms. We particularly focus on attacks on the physical environment, and propose the first such attacks to motion planne
distribution of environmental information while optimizing robot trajectories; however, current methods lack the ability to construc
ions, where the robot and the obstacles are in contact. Our algorithm seamlessly traverses widely free regions, follows semi-fre
lanner has been capable of generating SEF paths for tethered differential-driven robots lacking an omni-directional tether retrac
based steering controller to reduce the number of samples needed in a sampling-based motion planner RRT. Our method com
es mobile manipulator motion planning more challenging. In this paper, we tackle the mobile manipulator motion planning probl
based motion planners. We propose to adapt a pre-trained VQ-MPT model to reduce the search space for constraint planning
ased stage finds a long-term global plan consisting of a contact schedule and a sequence of stable whole-body configurations. S
k-inspired aerial active exploration system, which utilizes the flexibility and view field advantages of Unmanned Aerial Vehicles (
ving in different indoor settings. We train models that predict whether there is a moving person nearby and then their location. W
er intention to interact, which can be trained in a self-supervised way. Our main contribution is a study of the benefit of features
designed a protocol to include non-verbal cues in the social robot Vizzy (head and eye gaze, hand gestures) and test them in a
the handover scenarios, including one or more of (a) conducting handovers at fixed locations, (b) not adapting to human prefer
ands to robots demands intricate mathematical formulations. Furthermore, the transition between tasks does not quite possess
g and vibration feedback for enhanced HRI. The dual-modal tactile e-skin offers multi-functional tactile sensing and programma

i-view human body reconstruction. Nevertheless, these methods have not fully utilized multi-scale semantic information from m
patiotemporal feature learning, our main idea is to use sequences of 2D skeleton heatmaps as inputs and employ Temporal Co
n unintended collision that requires safety manoeuvres. Many published methods classify discrete interactions using more adva
sambiguation questions, resulting in performance reduction in realistic interactive scenarios. In this paper, we propose TiO, an
he complexities of identifying user-specific objects. To address these limitations, we introduce Zero-shot Interactive Personalize
ey their intentions by relying on context to achieve goals, we introduce a new IOG task, Pragmatic-IOG, and the corresponding
rtifacts. This study presents a model-driven multimodal convolutional neural network (MM-CNN) for joint EIT-based force and sh
mmand is certain (i.e., clear) or not (i.e., ambiguous or infeasible). Once the command is classified as uncertain, we further dist
ns. The system consists of three modules: a dialogue assistance module, which facilitates Haru to speak to users and answer q
erent assistive strategies based on the performed task. From this perspective, human-machine interaction is required to safely
(LED). This novel controller design allows users to free up their hands for more immersive experiences. To track the controller's
n. Incorporating the physical constraints of these multiple aircrafts to improve the accuracy and reliability of integrated aircraft p
ave utilized the object's semantic invariance to generate a semantic graph to address this issue. However, previous works lack
ystem that is designed to be scalable, flexible, decentralized, and sparse, which are all key properties in swarm robotics. Our s
AVs. Nevertheless, when faced with unknown environments or the loss of Global Navigation Satellite System (GNSS) signal, m
g-inertial state estimation in expansive environments, enabling robotic swarms to operate without the need for anchor deploym
relative motion or stationary situation, we design a rotating ultra-wideband tag to provide the persistency of excitation condition
mon practice is to transform points among sensors while requiring strict time synchronization or approximating transformation a
ction and vulnerable to erroneous data associations without accurate GPS assistance. We present AutoMerge, a novel framewo
and the quadruped robot, taking into account the nonlinear robot dynamics and the uncertainties associated with the cable-towe
so advantageous for robots to wait for teammates and tasks to become available in order to form better subteams or reduce the
mation fusion. The majority of existing works either assume a passive all-to-all observation model to minimize the summed unce
zing behavior, a strict decentralized approach, and limiting the robot-robot communication. Despite applying such strategies, an
s in the team. To optimize this solution we develop a CSLAM back-end based on Consensus ADMM called MESA (Manifold, Ed
t predefined space. However, the dynamic distribution of the reward signal or quality of the MAB with diversity limits the perform
we propose a hierarchical control scheme, which uses an RL-trained policy as the high-level controller to generate waypoints for
-task plans and task space waypoint paths, which are used by a multi-arm motion planner to accelerate trajectory planning. We
otor teams maneuvering in simple, obstacle-free environments. However, the addition of obstacles increases the number of po
k assignment, which puts high requirements on efficiency and stability. However, robots may often get damaged even crash in
offer computationally efficient approaches to guaranteed policy improvement; however, this improvement is with respect to a sta
p up with others. Inspired by this, we propose a competitive learning framework which is able to help individual robot to acquire
erwater or in a lot of radio noise environment. OWC is a method of communicating only with other agents in visual range using
munication and sensor limitations. In this work, we propose a probabilistic approach to allow coordination among robots in comm
e future adaptations in a persistent coverage task. We consider a heterogeneous multi-robot team, where robots are equipped w
aph describing the probabilistic structure of all of the observations robots make internally or of each other, and is flexible for any
ior studies proposed incorporating $L$-hop delayed states into the computation. While this approach shows promise, it can lea
e introduce time synchronization into mutual localization, to recover the time offset and relative poses between robots simultan
Multi-robot Opportunistic Communication for Heterogeneous Collaboration), a framework for resilient multi-robot collaboration th
ge by both robots while minimizing the coverage time. The UGV can also serve as a mobile recharging station. The UAV and U
be has been explored in robotics. Thus, here we roboticize a non-cubic space-filling shape: the rhombic dodecahedron. This ge
ntainment achievement and an attitude control component to govern rotational channel. The proposed optimal control protocols
nd their unique skills in a team is challenging. Dynamic runtime conditions often make static task allocations infeasible, requiring
ories. We formulate the local MPC formulation for each individual robot that includes both individual and shared objectives, in w
e completion. To extend existing methodologies, our framework adeptly manages interactions among multiple agents. This exte
mes as many objects as comparable methods. We achieve this improvement by decomposing the planning space to consider m
n. Additionally, in an on-orbit servicing operation various failure modes of the target may result in unreliable a-priori knowledge o
ns have failed to account for realistic satellite operational constraints and have used static utility for repeat observations of the s
t clouds into range images, compresses them using an encoder-decoder framework, and subsequently reconstructs the range
ed areas in the map, that must be traversed sequentially, i.e. a route. Each semantic label represents one or more constraints o
fully known 3D structures, our motivation is two-fold: first, 3D structures separately distributed in 3D environments are complex
y status of neighboring space in the near future, as the environment representation. Based on this representation, we extend th
os, such as those where there are unknown points of contention that require negotiation between robots. Each intent message
the belief space. Stochastic dynamic games can be solved to a local Nash Equilibrium using a game-theoretic belief space vari
ssigned to monitor disjoint partitions of the environment, reducing the burden of exploration. This also limits applicability when th
ncertainty. Additionally, we employ an iterative expectation-maximization inspired algorithm to assess the potential outcomes of
ease the problem complexity by: 1) considering battery-limited robots that need to schedule recharges; 2) tasks that can be dec
t lends better explainability compared to an end-to-end (learnt) neural network based policy for MRTA. However, deriving suitab
and highlight issues relevant to an engineering audience to raise awareness of AI and automation trends in mining. It assumes
follows: 1) the Top Layer (or Topological Layer) allows to model the traffic of vehicles among the different areas of the environm
emergency rescue scenarios. We present a distributed framework considering various preferences between robots and tasks to
ensus Filter (ICF) to estimate the state of the target(s). The ICF can be proven to be uniformly globally exponentially stable und
er this requires that the formation framework be bearing rigid. Rigidity may be checked numerically for a given formation frame
nal complexity operations for the global optimal solution, or recourse to simple yet effective heuristics for approximate solutions
 een the robots, our formation control algorithm aims to deploy a dense robot swarm that mimics the behavior of tornado school
based on a binary pattern projected on the surface. We use a decentralized Bayesian decision-making algorithm and deploy a
struction plan may thus build a staircase-like scaffolding of blocks to reach other blocks at higher levels. The order of block plac
lly limits the application of anthropomorphic hands. In this letter, we propose a fast force-closure grasp synthesis (FFCGS) met
ve and hybrid actuation systems with lockable joints, facilitating transitions between the execution of delicate and forceful tasks
any robotics projects. To this end, we propose an accessible robot hand capable of tactile sensing, which can be produced throu
hallenging problem due to the high dimensional working space. In this work, we propose a novel temporal-parametric grasp pri
to capture hand motion information. However, this method suffers from limitations as the collected information cannot be direct
with any robot and maintainability, the mechanism must be contained within the small space of the fingers and palms. In this pap
hands evolved greatly, hardware platforms still lack capabilities, particularly in tactile sensing and providing soft contact surfac
gh-skill policy, whereas the mixed dataset is collected using both expert and non-expert policies. We found that vanilla behaviou
orresponding finger mechanisms. Moreover, we propose a manipulation strategy for grasping shaft-shaped parts with the propo
AV-USV landing control algorithm via heterogeneous-feature matching. Specifically, the heterogeneous landing features of diffe
Barrier Functions (CBFs), whose primary objective is to make minimal adjustments to a nominal controller while ensuring constr
ion of robots in a group. Our analysis shows that the input constraints on the leader are necessary for stable formation control.
hould influence the others for the most efficient behaviour. In this paper we use the example application of distributed informatio
a notable absence of real-life implementations and studies concerning their application in time-critical robotics scenarios. This a
d rotatable parallel jaw gripper and a dual camera setup for experimentation. The system design is focused on continuous oper
rn a menu of cloud compute services that tradeoff latency and cost. Because it is infeasible to try every hardware configuration
nce of existing communication infrastructure. Different from previous works, our approach forgoes heuristics for network perform
connection, as well as the Time-Domain-Passivity-Approach (TDPA) for the communication channel. We reduce conservatism
ught solution. IFFNeRF utilizes the Metropolis-Hasting algorithm to sample surface points from within the NeRF model. From th
a lower budget and higher performance. In this paper, we propose a 4D object detector, VeloVox, to establish accurate object d
instance-aware sampling strategies to discover multi-class road users and propose an occupancy map tool based on kernel de
vely learn depth information for accurate feature lifting and object localization. Different from previous methods which directly pr
d to adapt the detection network trained on clear weather images to adverse weather images. While previous methods do not e
nting a novel symmetry-aware multi-view 6D pose estimator called SyMFM6D. Our approach fuses the RGB-D frames from mul
d ignore uncertainty quantification. Addressing this gap, our novel study integrates the principles of conformal inference (CI) with
m monocular observations. In this paper, we propose a novel pipeline that decouples the 6D pose and size estimation to mitiga
nce. In this paper, we propose a framework for category-level pose estimation from a single RGB image in an end-to-end mann
pixels corresponding to the presence of unknown objects. To address this problem, the pixel-wise anomaly-detection approach
his limitation, we propose PVTransformer: a transformer-based point-to-voxel architecture for 3D detection. Our key idea is to r
and missing visual details. In this paper, we propose a cross-modality feature distillation method that can focus on regions whe
rmer restrains its ability to learn features at different scales. Such limitation makes transformer detectors perform worse on sma
ct detection results by fusing history detection frames. In FrameFusion, we ``forward'' history frames to the current frame and a
these, pillar-based models are most suitable for onboard deployment due to their light architecture. Despite their advantages, p
mpressed video. Specifically, our framework draws the inspiration from the principle of Wavelet Transform, and thus we design t
onal demands and increases receptive fields, it will compromise the preservation of crucial non-local information for accurate 3
ic, instance, and semantic bridged network to delve into the reciprocal relation. To make semantic and instance benefit from ea
simultaneous backbone refinement and hallucination generation. Specifically, spatial alignment is proposed to deal with the ge
ges. However, shifts in a robot's deployment environment can also impact the likelihood that different objects will occur, termed
ages inaccurate depth prediction and massive pseudo points for perspective-to-BEV transformation while the latter only fetches
steps. However, this is often limited as its gain does not scale well with long-term past observations. To address this, we advoc
ace of improvement and the computation cost is large, which is bad for real-time object detection. An improved Yolov5 model: H
ly as natural language expressions. In this paper, we tackle this problem by using referring expression segmentation, which is a
ared to regular object detection. These challenges encompass the need to detect extremely small-sized objects, contend with s
nferior performance compared to their LiDAR-based counterparts due to inherent depth ambiguities in images. In this work, we
h in the form of a lightweight module that leverages entropy measures, segmentation predictions, and spatial context to charact
on large-scale data to learn RGB/RGB-D feature embeddings, where cluttered environments often result in inaccurate segmen
mplicated by dynamic elements and localization uncertainty. To this end, this paper presents (1) an uncertainty-aware multiage
nce understanding. In this paper, we introduce a new open-vocabulary affordance detection method in 3D point clouds, leverag
nce of several weakly supervised frameworks for 3D object detection which can automatically generate pseudo labels for unlab
riors, a unified and general pipeline for human-object interaction scenes by leveraging recent advances in differentiable physics
y and higher resolution, radar is emerging as a viable alternative to lidar as the primary sensor for mapping. In this paper, we p
orm unsatisfactorily due to the huge domain gap between low-resolution and high-resolution point cloud classification. Some sc
ach is to incrementally estimate high confidence free-space areas by modeling and accounting for sensing, state estimation, an
on and mapping (VI-SLAM) point clouds unexamined. In this paper, we create and release two datasets: VIP500, 4772 VI-SLAM
To alleviate these problems, we propose a new LiDAR and Camera Range-view-based 3D point cloud semantic segmentation
ving instance tracking in sparse radar point clouds to enhance scene interpretation. We propose a learning-based radar tracker
.g., roads, lane lines, vehicles and pedestrians) can be assembled into progressively more complex shapes to form a BEV repr

ion hyper-parameters and simpler than previous techniques. For neural networks used in safety-critical applications, bias in the
pplied. Thus, we propose EfficientDPS, an efficient, end-to-end, and unified model for DPS. In our method, query features extra
time delays, and a comprehensive framework that can simultaneously address such issues is absent. In addition, the feature e
. After analyzing its properties, we demonstrate how CID can benefit point cloud analysis without the need for meshing or norm
OT research field have been constructed around datasets functioning at 10--30 frames per second (fps) which can be difficult to
al for real-world applications. On the other hand, weakly-supervised domain adaptation (WDA) is an underexplored yet practica
LiDAR and cameras often comes misaligned due to the miscalibration, de- calibration, or different frequencies of the sensors. A
, an end-to-end framework that overcomes these limitations. By utilizing a transformers, TerrainSense detects lane semantics a
r information. In this paper, we propose Radar-Camera Multi-level fusion (RCM-Fusion), which attempts to fuse both modalities
cess. Active learning holds promise as a method to reduce the human annotation effort needed for semantic segmentation. Ho
ion module to adaptively fuse feature maps via a gated network. We perform extensive evaluation on nuScenes to demonstrate
work we explore AL for LiDAR semantic segmentation. As a human expert is a component of the pipeline, a practical framewor
R semantic segmentation. First, they overlook the severe class imbalance inherent in LiDAR semantic segmentation datasets. S
ctory prediction, and UGV control. Firstly, an estimation algorithm based on the Extended Kalman Filter (EKF) is employed to m
S) with four wheels in parallel to each other can maintain the vehicle's heading when moving along a curve. Symmetric negative
follows a specified path while satisfying up to third-order constraints to ensure safety and smooth motion. One significant challe
es of such systems. However, fast and stable convergence of iterative methods is reliant on the application of a high-quality pre
terative methods to solve the underlying moderately-large and sparse linear systems, are a natural fit for parallel hardware acce
r even hours) can hinder its ability to respond quickly to changing scenarios. Additionally, modeling errors and external disturba
s invariant descriptors for force trajectories by exploiting the duality between motion and force. Two types of invariant descripto
aints. While trajectory retiming is most often formulated with the minimum-time objective (i.e. traverse the path as fast as possib
 mobile manipulators. From a software standpoint instead, complex hierarchical strategies are often used to define a set of beh
ork, we first introduce a continuous-time planning formulation with an average-value-at-risk constraint over the entire planning h
ency of this energy storage and release mechanism remains one of the major challenges in controlling elastic robots. A possib
als with applying DDP to articulated soft robots by using numerical differentiation, in addition to using pure feed-forward control
sensors while circumventing explicit modeling of the contact force evolution. Our approach is based on the online estimation of
alization of these popular formalisms can be found in geometric algebra. The aim of this paper is to showcase the capabilities o
nces. Our main insight is that a single controlled trajectory of this embedding system corresponds to a controlled forward invaria
e challenges, recent work has explored learning CBFs using neural networks via neural CBFs (NCBFs). However, such metho
(integral) control barrier functions. The result is a stability-safety-passivity (SSP) filter implemented as a convex optimization con
s by finding good approximations of the value function. We summarize a unified SOS framework to synthesize both under- and
article Filter with Exploration Particle (MCP-EP) is employed to establish real-time feedback of multi-contact information. Then t
with at least 3 DOF in 3D) systems. Recently, a machine learning-based Graph Neural Network (GNN) simulator has been prop
s nonlinear systems to minimize a non-convex cost function. Specifically, TO can get stuck in poor local minima when the sear
rning method. In particular, our approach uses a nominal, global, nonlinear, physics-based model with a local, linear, data-drive
ts such as collision-avoidance and control-effort constraints. We provide an algorithm with bounded dynamic regret, i.e., bound
LIPM). More precisely, an induced norm can be taken to formulate the fact that the balance for a biped robot is achieved if its z
ccount throughout the learning process, wherein the system’s dynamics are also assumed unknown. These challenges lead
than degrees of freedom and thus require for more elaborate control systems. This paper proposes a novel co-design algorithm
Most existing ergodic control approaches assume the robots/sensors as individual point agents moving in space. We introduce
p, weight-tied neural network with rectified linear unit (ReLU) activations. This reformulation enables the deployment of ReLU-Q
rst decentralized system that achieves coordination between a heterogeneous fleet of black box robots for which the internals o
rning-based methods to effectively design the cooperative plan. To this end, we develop a Stackelberg game-theoretic approac
nvoy, wishes to travel from a starting location to a destination while minimizing some accumulated cost. The convoy may traver
different sensor and motion models, to optimize coverage in a target region, based on a measure called ergodicity. In this pape
g, they frequently overlook robot dynamics in simulations, leading to a simulation-to-reality gap. In this paper, we bridge this gap
efficient algorithms for multi-agent path finding (MAPF), though often in environments with simplified kinematics. In this work, w
quentially selects the optimal level of compression, guided by the robot's path, by balancing map resolution and communication
cannot ignore the effect during planning. In this paper, we extend the reservation system in autonomous intersection managem
an process-based approach that uses gradient descent to optimize paths in continuous environments. Our approach efficiently
h Finding (MAPF) methods to precedence-based algorithms. However, it is still unclear how human presence impacts different
nsures optimality but suffers from increasing complexity as agents or maps grow. To tackle this, we leverage learning-based ap
ch SFCs can be overly conservative when dealing with a quadrotor swarm, substantially limiting the available free space for qua
hile reasoning about each agent's kinematic and dynamic constraints and actuation limits using MPC as the low-level planner.
od, and discontinuity-bounded A*, a single-robot kinodynamic motion planner. Our method, db-CBS, operates in three levels. In
ng-based MAPF approaches typically only let each agent make decisions based on a limited field-of-view (FOV) around its posi
or all the robots in each horizon, significantly increasing the computation burden in large workspaces with many robots. As a re
aisles where coordination between robots is required. In single-robot settings, topology-guided motion planning methods have s
oted Min-Max Tree Cover (RMMTC) problem. For the first time, we propose a Mixed Integer Programming (MIP) model to optim
score for queried 3D points within the sensor field of view. Then, this network is translated into an algebraic symbolic equation
event cameras, our ultimate goal is to design a computationally lightweight event processing method that persistently identifies
gripping mechanism grasps the powerline cable during landing after which a control circuit regulates the magnetic field inside a
 ne as a polynomial by solving an optimal control problem. A modified Q-network, termed LPNet, is presented to predict the acti
nvironment and assume that the map built during exploration is accurate enough for safe navigation, which is usually not the ca
itation, we propose the Sim2Real-based Multi-Scale Context Mask-RCNN (MSC-RCNN) network, specifically designed for slan
r via motion compensation, accumulation, projection, and upsampling processes. The generated 3-channel point cloud tensor a
n and in the multiscale setting. Our contributions are threefold: (i) a Bounded Rectification technique to prevent tagging many no
nobserved obstacles due to unpredictable disturbances, sensor limitations, and control uncertainty. Instead of completely avoid
nting direction, psi_d, which utilizes control barrier functions (CBFs). First, we generate a spatial density function, Phi, which lev
ed in this work utilizes Gaussian mixture models (GMMs) to represent the environment. While prior GMM-based mapping work
that uses Deep Learning (DL) to predict the most likely indoor map given the previous observations, and Deep Reinforcement
eton-guided planning framework that can achieve fastaerial coverage of complex 3D scenes without pre-processing.We decom
n set of rooms in a structured building, keeping track of which rooms are already explored and sharing this information among
rd UAV perception, research focus has been shifting toward higher-level reasoning for multi-robot missions. However, autonom
l depth images using supervised learning such that it compresses the high-dimensional depth data to a low-dimensional latent s
nboard sensing and detect the gap from a single onboard camera. The training problem is challenging for two reasons: a precis
apping to guarantee safe navigation in complex and unstructured environments. Our system relies solely on onboard IMU mea
current work advances flow detection by introducing a strategy to distinguish between two flow sources applied simultaneously
vy Light Detection and Ranging (LiDAR) sensor and their high computation cost for learning-based data processing make those
Perception-Aware trajeCtory gEneration framework for quadrotors aggressive flight, that takes into account feature matchabilit
em, mobile robots are deployed to work just like in a warehouse, moving plants grown in containers to make sure every plant re
and the estimation of their properties - a key capability to let autonomous unmanned aerial vehicles soar as well. In contrast to
rence time, and mitigate the computational burden.Furthermore, we optimize the Spatial Pyramid Pooling (SPP) module to enh
nse, we propose Natural Quick Response codes, or N-QR, which enables rapid and reliable correspondence between large-sca
ely impeding their grasping accuracy. Furthermore, scene reconstruction methods have primarily relied upon static techniques,
ation robots with tactile awareness. This will facilitate safe and natural human-machine interactions while enhancing object man
pplications, are the suction cups. The rudimentary act of grasping an object and relocating it, devoid of a cognizant awareness
d rotation pre-training strategy is adopted to deliver robust initialization for RGB and depth networks. We then propose a globa
onstruction. Recent work has shown neural radiance fields (NeRFs) work well for depth perception in scenes with transparent o
ainly straight as commonly found in the real world. To reconstruct such structures in an efficient manner, we employ straight edg
ifferent liquids demand different manipulation policies.We command the robot to rotate the container and monitor the liquid resp
clusions by the hand and object. We turn this challenge into an opportunity by utilizing the hand shape to constrain the possible
ble clutter, which is how they are commonly stored and transported after harvest. The method consists of a deep learning-based

itional sensors, such as depth camera, the sparse and weak visual features still hinder the improvement of detection accuracy.
the localized, dense, and HR representation of contact surfaces. In particular, we build a Gaussian triaxial tactile sensor degrad
s from previous, possibly simulated, experiences. In contrast, many other works estimate articulation by observing the object in
rom grasp to grasp. Tactile sensing offers a window in the contact mechanics and provides information about friction. Notably to
point correspondences, which can encounter optimization problems for objects with symmetries, since each point may have mu
To address this limitation, this paper presents a self-supervised pose estimation method based on a novel pixel-wise weighted
ose of unknown environmental contacts encountered during the manipulation of a DLO, classifying the induced constraints as u
e propose AnyGrasp for grasp perception to enable robots these abilities using a parallel gripper. Specifically, we develop a den
y, In phenotyping, the detection, association, and maturity estimation of tomato trusses and individual fruits are accomplished th
 main of assistive robotics, we seek instead to utilise the user's cognitive abilities for enhanced satisfaction, grasping performanc
sual obstructions arising from the robot's own manipulation activities. Yet, there exists a substantial opportunity for robotics to e
eal world is time-consuming, can induce costly failures, and often requires extensive human intervention, which causes operati
introduce extra costs to the robotic hand. Also, the common approach of visual perception may not be available due to occlusio
ation for objects with known 3D models, even when they are very small and with challenging reflective surfaces. We present a
ment pose of each object based on the single goal image and observations from the current scene. In order to establish sufficien
ing. We propose the EDOPT algorithm, which maintains real-time operation with a variable event-rate (which occurs due to var
ches and neural policies. In this work, we introduce a novel attention-based neural architecture capable of solving a smoothing
s. This has the benefit of providing a single, quantitative metric to evaluate state estimation algorithms against, while being com
ts. To address such limitations, this paper provides autonomous driving datasets and benchmarks for offline reinforcement lear
e community the best practices to learn from errors. An ideal benchmarking protocol should encompass the different aspects u
Additionally, the benchmark is designed to be easily reproducible in the real world, making it accessible to researchers and pra
llowing such a regiment provides a little insight into the properties of the models both in terms of how well they can handle diffe
ated as false positives. We therefore propose variants of the 3D AP metric to be more permissive with respect to depth estimati
n robotics simulators like Gazebo. The main goal is to facilitate the development and evaluation of human-aware robot navigat
two distinct benchmarking approaches: black-box testing, which measures performance by eliminating upper layers and replac
and comprehending them in real-world applications remains elusive. This study sets out to address two primary objectives: (1) t
y stable way, by bounding the amount of energy allocated to the controller. Nevertheless, when the workspace is shared with hu
sm for human performance augmentation. A core challenge in deploying robots that interact physically with humans in dynamic
novative method for achieving adjustable loads during resistance training by leveraging physical Human-Robot Interaction (pHR
o a hierarchical multi-objective optimization problem. A major contribution is the prediction formulation, which allows to consider
of Z-width, which indicates the region of passively renderable impedance by a haptic display, is also important for rendering a
f the garment make the task a challenging problem for robots. In this paper, we propose a bi-manual control method for garmen
As a result, an at-home-based game design combined with squat workouts is essential to enhance lower-limb performance. We
ds of physical care and emphasized the need for a deeper understanding of the complex interaction between humans and the r
he granularity nature. Individual variances can introduce significant bias in the rating scale results. A high granularity in the sca
data from a motion capture suit, the current contact forces. The latter predicts -- taking as input the same kinematic data and th
a major bottleneck is the reliance on human motion capture data, which is expensive to acquire and difficult to scale to arbitrar
nicate the desired destination of motion through back-and-forth force exchanges. Inherent to these exchanges is also the ability
orimotor control. There are many common interests and interconnections between the fields of human modeling and robotics. I
ur previous predictor by achieving a success rate of 93.8% in testset and 90.9% in real experiments with 21 volunteers predictin
mobile pHRC but often gets neglected is the impact of non-rigid components e.g. deformable tyres, suspension systems and u
als and bias strap forces can be modulated to optimize interface power transmission. Few methods currently exist for estimatin
In addition, they all lack adaptability to individual requirements. To address these challenges, this paper introduces WANDER, a
uilt by different agents when a single frame of overlap is detected, which can lead to incorrect merging due to the false-positive
g the uncertainty of the predictions. To obtain prediction regions, we use conformal prediction, a statistical tool for uncertainty q
ces uncertainty, which is rarely considered in trajectory generation. In this paper, we propose Distributionally Robust Safe Corr
has been observed that MPPI tends to make poor decisions if unmodeled dynamics or environmental disturbances exist, prev
tacle-aware ego robot plans. However, planners that ignore the inherent uncertainties in such predictions incur collision risks an
ot support hybrid beliefs directly. In particular, these solvers do not address the added computational burden due to an increasin
, known as fully observable data association. To address this limitation, we propose a planning algorithm that maintains multiple
h the goal of maximizing the expected return. But the resulting policies cannot provide safety guarantees which are imperative f
dmap must be sparse so that the planning problem is tractable, but still contain edges that are likely to contribute to low-cost pl
f the robot or environment, or requires per-object knowledge of noise. Instead, we propose a method that directly models senso
or continuous problem domains. In many large robotic domains, hierarchical decomposition can simplify planning by using tools
lity of model predictions has to be considered for robotic exploration and planning in the unknown environments. To address th
presence of an unexpected obstacle). These situations require the robot to use information gathered online to correct its plannin
or only reactive collision avoidance is carried out. We propose a novel approach to not only avoid dynamic obstacles but also i
mapless environments while ensuring comprehensive coverage of all traversable directions. Our formulation uses a Conditional
inistic planning problem using Rapidly-exploring Random Graph (RRG) to construct a nominal trajectory graph. Then, an inform
nto a control loop due to the uncertain nature of human motion, which can make predictions of these models imprecise. In this w
ures. To this end, we propose a Gaussian Process (GP) based model for proactively detecting the risk of future motion planning
nization, where one converts a stochastic world into a de- terministic one and plans under this simplification. This makes the pl
d reasoning, the measurements are susceptible to several environmental effects such as the presence of occluders, variations
environments, our work combines the previous notion of Time-Varying Markov Decision Processes (TVMDP) with partial observ
high-quality features, but comes at the cost of decreasing performance in other, less favorable environments. For robotic system
novel sigma point-based approximation to accurately estimate mutual information (MI) for general, non-Gaussian distributions
cision insertion tasks as planning problems under pose uncertainty, where we effectively utilize the occurrence of contacts (or th
While visual perception can deliver a more accurate representation of the environment, its use for safe motion planning is limite
s used in the Belief-A framework, and inherits the completeness guarantees of Belief-A's low-level sampling-based planner. We
cted because of the measurement uncertainty of onboard mapping and motion uncertainty. This study presents an uncertainty-
ety, particularly when accounting for uncertainties in robot perception and control. To address this issue, this paper presents th
d multi-agent path planning problem by converting the original Environment Graph (EG) into a Joint State Graph (JSG) to impli
average task waiting time by allocating tasks to cranes and optimizing their execution sequence. Unlike prior research focusin
existed algorithms that explicitly characterised non-homotopic paths using topological invariants such as â„Ž-signature and win
re its applications in layout design for multi-robot path planning. Roughly speaking, a well-connected set over a connected grap
s on scenarios involving homogeneous robots, but where tasks vary in the number of agents required to start them. For exampl
execution. Potential task failures can add further contingency tasks to recover from the failure, thereby causing delays. This pa
nce of uncertainties. Our idea is to consider traffic flow equilibrium in path planning to relieve any potential congestion and incre
s paper presents a formulation based on Mixed Integer Linear Programming (MILP) for allocating tasks to robots by taking into
tric gait pattern, slower walking speed, and higher cost of transportation. Although powered prosthetic legs address the aforeme
nners (implantable markers and CT scans), which introduce unnecessary trauma and prolong the preparation time for patients.
on of body joint positions based on EMG features has remained largely unexplored. In this paper, we address this gap by propo
r IMU sensors attached to the thighs and shanks on both legs were utilised to collect data during trials on ten healthy subjects f
ever, these recent efforts only use continuous impedance control during stance and instead utilize discrete transition logic to sw
d three-loop linkage for function of continuous energy absorption-release in gait stance phase, which can facilitate ankle plantar
control, particularly on inter-joint coordination, remains relatively uncharted. This paper investigates the effects of generic low-a
ecoding methods of surface EMG signals remains challenging, which limit the application of robot hand in real life. Herein, we p
t force transmission, and giving autonomy to users when walking. In this paper, we propose a novel BSE with a differential seri
n elliptic rolling joint as an artificial wrist joint, mimicking the two-row structures of carpal bones. The joint offers two degrees of
ematic compatibility with the user. This paper presents a novel ankle exoskeleton frame design that allows for 3 DoF ankle moti
ing design approaches and determines the optimal anterior-posterior (A/P) CoM position considering factors like actuating pow
hancing user experience. We present an approach for real-time transition detection, aimed at optimizing the processing-time pe
se to external forces. We map the step duration and position to a human-like foot trajectory, which is then translated into joint tra
w sensitive participants are to their unique torque profile. To investigate, we applied previously optimized hip-knee-ankle torque
an replicate normative joint kinematics and kinetics, active control also holds the potential to shape system-level characteristics
rders (WMSDs) by ensuring continuous joints support. However, achieving synchronization between the exosuit and human m
way of distance candidates selection method which enables the use of a very small number of candidates and reduces the com
ed for fisheye cameras, catadioptric cameras, and super fisheye cameras. But in both cases, few models offer the possibility o
ely); such point clouds can then be used for accurate environmental mapping, navigating unknown environments, and other rob
mantic class. In this way, each LiDAR scan is reduced to a compact collection of four-number vectors. This abstracts away imp
thods need to be further improved. In this paper, we propose a novel method for degeneracy detection using local geometric m
-time. To address this issue, we present the i-Octree, a dynamic octree data structure that supports both fast nearest neighbor
as measuring beam-to-surface incidence angle. This fact has recently motivated researchers to exploit traditional filters, as well
the surrounding terrain is important for aspects of the drivability of a surface, path planning, and obstacle prediction. In this artic
D visible models which generally reconstruct one viewpoint of objects. We propose a method to leverage both the general Morp
metric map representation, with a novel hierarchical and parallelizable obstacle avoidance algorithm formulated through Rieman
general poor visibility conditions, that can cause the performance of state-of-the-art vision-based methods to decline. In order t
sequence and solves it using a sampling-based method. Results by MPPI significantly depend on sampling noises. To quickly
eded. Instead, we formalize the problem and propose a novel field- based reactive planner that explicitly accounts for rotationa
ounter. In particular, our proposed feedback control system is theoretically proven to achieve convergence within a predefined f
and robot manipulation. However, despite its versatility and effectiveness, RAMS demands expertise in robotics, leading to a h
obotic systems to acquire novel manipulation skills. However, simultaneously learning geometric task constraints and control p
of a reactive circular fields motion planner with the focus on simultaneous control of multiple robots in robotic soccer games, wh
navigation subgoals around the robot. Subsequently, a cost function, which prioritizes the safety of the robot in terms of keepin
This study investigates the potential of Evolutionary Computation (EC) methods in robotic design optimization, with an underact
he design are most relevant for successful translation to the community presents a challenge. Here, we design a portable rigid
required during assistance. This research focuses on achieving mechanical transparency within the context of shoulder motion
e admittance controllers that allow humans to adapt the motion. However, when the patient's limb has limited mobility or is pote
ed according to different safety constraints. During the single support phase, ground contact constraints are enforced. In the sw
arm movement with success in simulation. However, challenges remain in translating the successes into real-world settings. On
o find the right balance between robotic assistance and FES that will provide personalised assistance, prevent muscle fatigue a
nce to self-guided PT, but prior robotic systems have been large and immobile, which can be a barrier to use in homes and clin
se constraints in complex environments, such as safety constraints. Some of these approaches prioritize goal attainment withou
ation to manage this trade-off effectively. Our approach embeds safety constraints within the action space of a modified Markov
l estimations as concrete samples, which undermines the integrity of learning. While several studies have addressed this issue
ance safety, feasibility, and performance. In this work, we delve into the problem of adaptive safe learning for multi-agent system
ast experiences for data collection. Independent of the underlying RL algorithm, we introduce the concept of a Contrastive Initia
policy optimization framework is introduced where the policy is optimized towards optimality and safety, simultaneously. The op
erceptive DRL-enabled systems remains incredibly compute and memory intensive, often requiring huge training datasets and
ing information across tasks will depend on how well the structure underlying the tasks is captured. In this work, we introduce o
eal with complex scenarios, such as occluded, blurred, and unaligned scenes. In this study, we introduce an end-to-end framew
d demand for augmented reality displays in vehicles. In order to enhance driving convenience and safety, automotive Augment
plying actions that are off the feasible control manifolds can quickly lead to undesirable divergence. In such cases, most sample
ern within the realm of embodied AI, as the robot accesses substantial personal information. However, the issue of privacy leak
scenarios, restraining their widespread applications. To fill these gaps, we propose MAexp, a generic platform for multi-agent e
obotic applications, an inaccurate simulator is often available. However, the data directly collected from the inaccurate simulato
extremes and many other external physical interference, are highly dynamic and difficult to model. REFORMA can perform a re
eactHD, an HDC-based framework tailored for perception-action-based learning for sensorimotor controls of robot tasks. React
on policies. However, previous research limited their scope to single-material DLOs and engaged in time-consuming data collec
table for real-world event detectors. Such uncertainty in an event detector generates multiple branching possibilities on LTL ins
egration in human perception, we introduce MOSAIC (Multimodal Object property learning with Self-Attention and Interactive C
d a time-consuming exhaustive search to find a solution and often make unrealistic assumptions, such as having identical, simp
, and powerful ability to represent unbounded scenes. However, existing NeRF works that adopt MPI representation for novel v
d distance functions of either the robot geometry or the swept volume of the robot motion. Building on this, we present a novel n
metry and appearance in a way which is compact and ideal for robotics applications. However, limited prior methods have inves
common aid from auto-encoders to establish the self-supervision by the self-reconstruction schema. However, the previous au
entations for robotic control. Building on theories of "what-where" representations in psychology and computer vision, we use s
nd structures, bringing two challenging issues: (1) determining the task-relevant geometric structures across different objects a
ning approaches use unprocessed, coordinate-dependent robot state data for learning new skills, thus inducing challenges reg
t representations from RGB images and SlotGNN for predicting their collective dynamics from RGB images and robot interactio
ation using three different robots and two different policy learning paradigms. From this effort, we can arrive at three insights: 1)
ork rethinks this approach; we present a representation learning mechanism that identifies visual features that best correspond
types of visual changes for RGB-D and point cloud based visual control policies. To perform these experiments on both model
ents the environment as a Signed Distance Field (SDF). State of the art representations of SDFs are based on either neural net
ers that are better suited to robotic tasks than their generic counterparts. We propose Scene Objects From Transformers, abbr

acting planar regions from the environment, they often suffer from issues such as low computational efficiency or loss of resolut
ne reconstruction task challenging. In this paper, we present a method for scene reconstruction by structurally breaking the prob
se attention aware mechanism that not only requires the model points to predict the correspondence, but also explicitly models
ribution. To address this issue, we propose a novel stereo-LiDAR depth estimation network. Our network includes a deformable
nt approach. We use cycle-consistent keypoints as salient points to enforce spatial coherence constraints during matching, impr
countermeasures. MMAUD addresses a critical gap in contemporary threat detection methodologies by focusing on aerial dete
annotating 3D data for supervised learning is time-consuming. In this work, we explore zero-shot instance segmentation (ZSIS)
hift clustering algorithm is not differentiable, making it difficult to integrate it into an end-to-end neural network training framewo
tically segmented by a pre-trained 2D U-Net and a dense depth prior is estimated from a depth-conditioned pipeline fueled by D
astive learning, to tackle performance degradation in complex weather conditions. Concretely, we first present a progressive cu
ected agents' observation is very challenging. In this paper, we propose a novel collaborative decision-making method that effic
usly reconstruct how all objects moved between the two point clouds. This setup generalizes standard 3D registration where on
nformation. To address these problems and provide a more robust DLO state estimation for downstream tasks like tracking and
ensor. Typically, a sensor processes the transient histogram using a proprietary algorithm to produce distance estimates, which
wn that Vision-Language Models (VLM) pre-trained on web-scale text-image datasets can learn fine-grained visual concepts tha
s in 3D representations. However, these approaches tend to produce maps with per-point feature vectors, which do not scale w
The algorithm improves on previous approaches by addressing three common scenarios which cause tracking failure: tip occlu
 imbalance is neglected. In this work, we address the challenge of 3D semantic segmentation with a long-tail distribution of class
ever, utilizing the continuum bending mode of materials brings critical problems, including undesired movements and joint radiu
rease that comes with solar tracking often far outweighs the amount of energy required to move the panel, which makes it a co
d to reorienting or resizing the limbs of a predetermined and static topological genus. Here we show policy gradients for design
e steel shell (BSS) as the stiff element was inserted into a 3D printing soft helical exo-skeleton to achieve coiling around and ho
by trivial geometric and algebraic conditions. These conditions can be readily obtained by analyzing the determinant of the rob
ed, fabricated, and evaluated. The motion of the robot was characterized and different combinations of actuation amplitude, fre
ng it to move smoothly in different directions. The robot is designed to have a narrow cylindrical profile to lower drag and impro
ble workspace. In addition, type II singularities, which critically limit the orientational workspace, can be fully avoided by utilizing
cal challenges, primarily related to their wire-driven structure, such as wire-routing and wire-friction problems. Therefore, this re
nt for many applications. Can such a consolidation happen in robotics? Conventionally, robotic learning methods train a separa
converts pre-trained or already fine-tuned Transformer-based robotic policies of quadratic time complexity (including massive b
stinct grasping strategies: tap grasping using adhesion forces such as electrostatic and van der Waals, fingernail grasping leve
elevant internet-scale data may be used for pretraining other modalities such as tactile sensing. Such pretraining becomes incre
ch directly learn how a robot should act from interaction data, we adopt a factorized approach that can leverage large-scale hum
ory optimizers. In addition, the set of feasible configurations becomes a measure zero set, which presents a challenge to samp
imation power of neural networks, there is a growing trend toward representing CBFs using neural networks, leading to the not
obability distribution over possible states. We propose belief control barrier functions (BCBFs) to enable risk-aware control synt
tem can move from any state in the control-invariant set to any other state within the given time horizon. In this paper we develo
minimum uniform scaling factor that results in an intersection between two convex shapes and to also compute the Jacobian of
and under-actuation nature of cloth present significant challenges for robots to effectively interact with them. Many recent work
y wherein data is collected online and incorporated into a reachability analysis approach to check in real-time that the system av
ucture and equipment. For decades, the approaches we have been testing legged robots were rarely standardizable with hand
n faced with novel or out-of-distribution inputs. Such errors can cascade to catastrophic system failures and compromise syste
able to capture uncertain dynamics or disturbances). This work proposes a different approach: training a state-conditioned gene
control. The proposed approach constructs an inverse perception contract (IPC) which generates a set that contains the groun
rk that (i) identifies safe frontier targets for multiple robots while taking into account the system dynamics of each robot to ensu
ge. A well-known tool to make controllers (either Model Predictive Controllers or Reinforcement Learning policies) safe, is the s
of states from which safety can be ensured. Unfortunately, computing such a set for a nonlinear system is extremely challengin
-world environments where objects can appear, disappear, and shift over time. In this paper, we present a closed-loop percepti
wn that low inertia and high bandwidth (> 30 Hz) magnetorheological (MR) actuators have the potential to improve the safety of
ble states. Namely, Particle Filters (PFs) can handle arbitrary non-Gaussian distributions in the robot's state. In this work, we de
ing to be high throughput but low error, to the choice of trusting a nominal policy versus switching to a safe backup policy at run
l inputs while satisfying driver intention. By leveraging {em backup controllers} designed to uphold safety and input constraints,
n-uniformly scaled point clouds, requiring a 9DoF transform for accurate alignment. We propose HEGN, a novel equivariant gra
quiring real or synthetic OOD examples at training time. From a shared per-pixel latent feature representation, a classification ne
emendous progress in recent years, they still face resource stringency and numerical instability on embedded systems when im
e degradation in scenes with sparse points. To break through this limitation, we introduce Sequence-to-Sequence tracking para
epresentation. To ensure global consistency, our system incorporates loop closure in the tracking module and maintains a glob
zation to unseen tasks. Prior task decomposition methods require task-specific knowledge, are computationally intensive, and
iffnesses. The gripper consists of three suction-cups located on the palm, along with in-hand perception. To evaluate the grippe
 ompetitive advantage. Strawberries, as fruits with a short shelf life, require sorting and packaging of fruits of different weights a
wheat, rice) are tiny with heterogeneous shapes and appearance, manually performing GAI is time-consuming and error-prone
hotogrammetry to enable inter-day precision agriculture analytics for intervals of up to a month. Our work starts by presenting a
computer vision-based method to measure pruning weight using a smartphone camera and structured light which produces res
s on a small subset of the leaves present in the breeding plots, while still requiring substantial manual labor. To automate this pr
behavior. In this study, we present a framework for learning the deformation behavior of tree-like crops under contact interactio
h. The proposed algorithm is evaluated in scenarios occurring during representative agricultural operations: harvesting and trans
ng providing reproducible and high-frequency measurements. Today's perception systems use deep learning to interpret these
abor-intensive procedure that faces the challenges of sparse sampling and human variability in determining SWP. In response
h the gripper and the crop, which is mainly caused by uncertainties from the sensors and the manipulator. To solve the problem
esting, disease detection, and yield estimation. However, achieving robust and precise infield fruit segmentation remains a chall
ent growth stages of plants. In this paper, we present an approach to perform correspondence estimation of different plant orga
obotic platform equipped with multiple laser and camera sensors for high-throughput, high-resolution in-field plant scanning. We
ntroller enhancing transient performance for an aerial manipulator extracting an object wedged in a static structure. This task inc
he scanning task in cluttered environments. To this end, we propose a cooperative system where an Unmanned Ground Vehicl
s and external disturbances. The controller handles the ascending and descending phases of the navigation using a visual lock
cene-scale reasoning) and the edges between them represent relationships. While these representations have shown great pro
pose a challenge for the development of efficient data recording and provisioning pipelines. Furthermore, accessing specific info
e development of a sweet pepper harvesting robot and reports its experimental verification, which mainly includes end-effector
and a robust control method of a transplanter are proposed for accurate transplanting. In this system, the path of transplanter i
or soil pH prediction, a key parameter of soil status for farmers. The EdgeSoil 2.0 predicts the pH value of the soil in real-time,
ning point determination, and manipulation that must operate under variable lighting conditions and in complex, highly unstructu
y, width and height) with wheeled mobile robots in the field. Our method is based on the fusion of 2D domain-specific data (norm
ler fruit sizes. To overcome these limitations, our method combines viewpoint sampling around semantically labeled regions of
her plant parts. We formulate this problem as a local next-best-view (NBV) planning task where the robot has to plan an efficien
lore the use of cobots (i.e., collaborative robots) to automate two key farming operations, namely, the transplantation of sapling
ns which results in degraded performance. In this work, we address this limitation with our proposed Injected Noise Discriminato
alleviate the performance degradation by manually labeling more data obtained in the new setup, this procedure is labor and co
peration. This work is motivated by the acute shortages of human operators and the need to improve the efficiencies of timber-
zation. To map and reconstruct the environment, the hector mapping approach and circle-fitting method are adopted for the oc
oward efficient and sustainable foresting practices. In this paper, we address the problem of automatically producing a forest in
er, the relationships of this data with the quantity of interest are often not known ahead of time, limiting the ability to leverage thi
se. In the surrounding phase, a custom artificial potential field is employed to simultaneously guide the robots to encircle the he
ular structure. In this paper, we first report on the design of the robot hardware including the dowel and hydraulic circuit, then pr
It samples trowelling sequence rollouts from a new plaster simulator and weights them according to the flatness of the finished
mbled and reassembled into more complex structures. Robotic modules are made by combining actuators with a functional voxe
he semi-static scenario where objects are added, subtracted, or changed in position over weeks or months. Our system combin
y model can be combined with the noise characteristics from multiple devices. For the task planning problem, we propose a lay
Reinforcement Learning (RL) automates this process without the need for collecting additional human-annotated data. We inve
vation tasks concerns obtaining skills adaptable to various conditions. When the conditions of soils differ, the optimal plans for
tructive evaluation (NDE) method can be insufficient in providing a comprehensive evaluation of the structure's condition. In ad
t two granularity levels. A coarse-resolution inference is formulated as sequence generation through a large language model, w
merous grasp datasets, their object diversity remains limited compared to real-world figures. Fortunately, foundation models pos
asks. State-of-the-art methods for task anticipation use data-driven deep networks and Large Language Models (LLMs), but th
bed using natural language. By using the same set of tasks as Meta-World, Language-World results can be easily compared to
presence of uncertainties is challenging as it requires enquote{chain-of-thought} reasoning, aggregating information from the en
s. We consider a scene graph model of the environment and utilize a large language model (LLM) to convert a natural language
pose a novel approach (CAPE) that generates corrective actions to resolve precondition errors during planning. CAPE improves
he existing semantic knowledge is typically constructed based on closed-world concept sets, restraining the generalization to no
t) architecture which allows us to seamlessly integrate agile development and reliable operation in a fully automated fashion. W
novel adaptive non-singular fast terminal sliding mode strategy is introduced to manage the visual servoing error. Unlike other s
ally unknown system, in tandem with a simple controller designed for trajectory tracking of a point located along the UAV’s
ecent Parameter Identifying Extended State Observer (PIESO) which employs a gradient adaptation law to actively identify the
e sampling and control application. These delays limit the practicality of MPC in resource-constrained systems when engaging
s with proof of asymptotically stable tracking error convergence, as well as a preliminary simulation study demonstrating 6-DOF
ely identify vehicle model parameters governing lateral tire slip effects, which can evolve over time due to factors such as tire w
behaviors. Leveraging recent results in risk-sensitive optimal control, this paper presents a risk-sensitive Extended Kalman Filte
onal terminal sliding surface, the sliding surface with global and terminal sliding motion is proposed, and the convergent time by
asks or to the constraints are chosen. In this paper, we address such instability shortcomings. First, we highlight the non-robus
ology by tailoring it to incorporate comprehensive and robust safety concepts. This paper introduces a ``safe-by-design'' digital t
tasks, specifically, those requiring precise replication of demonstrated trajectories or strict adherence to constrained regions ev
his algorithm runs in real-time and provides the safe and fastest control input for every cycle with respect to ISO standards. How
bot's view. This occlusion raises the robot's risk of unintended physical contact with the person, leading to discomfort and safet
responsibility in avoiding collisions, and (ii) individuals should behave legibly, that is, move in a way that clearly conveys their in
e complexities, we propose a hierarchical prediction strategy that combines data-driven internal state inference and planning-ba
ntroduce a methodology for quantifying this relationship, taking into account factors such as detection rate, detection quality, an
ose a novel control architecture that allows a torque-controlled robot to guarantee safety constraints such as kinematic limits, s
o develop methods to improve the models' robustness against them. Adversarial training is one approach to achieve this, but its
tnessed comparatively smaller levels of growth due to limitations in bi-directional communication and significant technical hurdle
ent tasks were recently added to LineRanger’s capabilities. While previous applications were focusing on non-contact inspe
ts of its mobile base cannot be controlled with the same system as the movements of the robot, and mobile base oscillation inc
rative robotic systems have emerged as invaluable assets, assisting first responders in several tasks, ranging from victim local
ework is structured around a novel goal oriented graph representation, that consists of, i) the local sub-graph and ii) the global g
ch proposes a Computation-Aware Search for Multiple Objects (CASMO) algorithm to further consider the computational time i
ing function to efficiently dispatch robots. The intersection system is composed of routing and clustering constraints. The exper
e the relay placement, establishing network connectivity to support robust flood-based communication routing. The proposed m
suspension. It operates in unknown and GPS-denied environments and on indoor and outdoor terrains. The exploration consis
d in that previous work, we first refine the theoretical model of NaivPhys4RP in terms of integration of variables as well as perce
identifying and mapping large semantic regions. The present work proposes a method for semantic region mapping via embod
LM-Grounder, a novel zero-shot, open-vocabulary, Large Language Model (LLM)-based 3D visual grounding pipeline. LLM-Gro
ations frequently for new circumstances. In this paper, we introduce a method for learning traversability from images that utilize
e model to favor semantics at the near side with the majority of points and object instances. Both the popular cylindrical and the
 sparked a paradigm shift leading to the advent of large foundation models that can be trained with completely unlabeled image
y, we propose a hierarchical gated convolution unit, which simultaneously expands the receptive field and restores edge texture
ormers are underexplored and face two key inherent difficulties: over-attention and noise from different modal data. To overcom
tation of its surroundings. Lifelong LERF maintains this representation over time by detecting semantic changes and selectively
eling, and present a technique for translating scene graphs into knowledge bases. To this end, we take advantage of the Unive
network based on a SLAM system with 3D occupancy mapping. When segmenting a new frame we perform latent feature re-pro
ntersections, pathways, dead-ends, and unexplored frontiers, which constitute the structural semantics of an environment. The
ueryable scene representation using RGB-D data. Open-Fusion harnesses the power of a pretrained vision-language foundatio
Former is the first transformer-based approach unifying semantic instance segmentation and tracking of sparse and irregular se
fferent instance IDs. To describe the dynamics of the surroundings, 4D panoptic segmentation further extends this information
ches extract scene priors from the instant visual observations and solidify them in neural episodic memory to achieve flexible n
o as expectation) on any given 3D scene graph, which is then used to strategically add new nodes (referred to as blind nodes) t
isadvantage of such Dirichlet-based VTM algorithms. We propose the Guided Gaussian-Dirichlet Random Field (GGDRF). Our
ependent encoders. The extracted features are subsequently fused and decoded to yield semantic predictions of collision-free s
egment novel classes, especially when the training data are contaminated by noisy samples. To addressing these limitations, t
overcome these limitations and provide Doppler velocities, delivering direct information on dynamic objects. In this paper, we a
ods assign high confidence to the entire map even when they are incorrect, leading to miscalibrated outputs. Several methods
clusion by nearby vehicles, the performance of these methods significantly declines in complex scenarios and long-range detec
inputs. Therefore, the networks easily become over-reliant on a single modality, making it difficult to learn complementary and
multiple agents are still challenging problems. In this work, we present Collaborative URBan Scene Graphs (CURB-SG) that e
per, we propose a BEV fusion (BroadBEV) that aims to enhance camera BEV estimation for broad perception in the pre-defined
stacles, the movement of the air-ground robot often suffers a suboptimal trajectory under existing mapping-based and learning-
erformance LiDAR is notably expensive and is not commonly available on vehicles. This paper presents mmEMP, a supervised
performance of 3D target detection through cross-modal schemes, deep learning algorithms, and training tricks, we devote atte
ment or guide the RGB modality; however, these methods typically adopt asymmetric network structures that rely predominantl
ds require strictly paired data during inference, which may not be available in real-world scenarios, while 3D-to-2D fusion metho
ional highlights and capturing detailed textures from both modalities. However, these techniques encounter limitations when fac
network that learns missing points or dense point density distributions to compensate for the sparsity of the LiDAR point cloud
or network feature-level fusion and only consider short-range HD map generation, limiting their deployment to realistic autonom
nd spatial differentiations are further provided. Based thereon, a new sliding-window spline fitting scheme is established for asy
ments. However, leveraging the full capacity of GNSS is cumbersome because of the diverse choices of formulations, error mod
they neglect the potential spatial tailing effect of moving object events caused by excessive motion, which may affect the struc
esides, discrepancies in the two data representations further complicate fusion methods. We introduce AYDIV, a novel framew
undaries, and suboptimal accuracy. To circumvent this issue, we learn to augment versatile and robust monocular depth predic
l anticipation ability, which can enhance spatial perception through semantic understanding, necessitates the availability of app
ch can result in lower accuracy when the RGB information is unreliable. To address this issue, we propose a novel deep neural
the camera motion is calculated, camera images are projected onto the point cloud. The variations in the projected intensities a
ent works leverage novel event cameras to complement or guide the frame modality via frame-event feature fusion. However, e
homography estimate along with an uncertainty. Uncertainty information opens the door to adaptive filtering approaches, post-
and staircases, where substantial accelerations and angular velocities can occur if the robot loses stability and tumbles. These
on jolting robot platforms (only 20-60% success rate). To improve the accuracy, a Point-Video-based Transformer Tracking mo
 ventional cameras are used. This work proposes a Spiking Convolutional Neural Network, processing event- and depth data fo
g impact forces have a negative influence on multiple robotic tasks. Recently, preemptive impact reduction techniques that expa
ally define and synthesize a quadratic CBF over a point cloud. This CBF is used in a CBF-QP for control and verified in simulatio
ered, we provide closed-form equations for the closed-loop system by obtaining the analytical form of pseudo-inverse for the in
ition scheme with image bracketing patterns is proposed. Images with different exposure levels are continuously captured to ex
be structure. This structure can help increase the spatial resolution of MagicTac to perceive external interaction contacts. Moreo
. To address these issues, our approach has been structured into two training steps: the first step focuses on microphone pair t
a speech module to rotate the mobile robot. It has two components, namely, a distant automatic speech recognizer and a soun
ts have primarily focused on supervised learning of a single instrument, limiting the performance of robots in the realm of cross
peaking, such a shape completion can be realized by exploiting prior knowledge about the geometry of the fruit. This is typically
ich is accurate but time-consuming as data is relayed between planets. This paper describes a global localization algorithm tha
n open challenge. In this paper, we present a method for legged locomotion control using reinforcement learning and 3D volum
023) based on multiple foundation models. This system is both generalizable to variations and adaptive by prompting each mod
detection and the Segment Anything instance segmentation models, the proposed method can accurately recognize the states
ean keypoint correspondence, and suffer from limited convergence basin or weak feature error robustness. Recent learning-ba
se a neural uncalibrated VS policy (NUVS) that can adapt to calibration disturbances with an adaption mechanism and a control
or distributing equal medical services, facilitating remote consultations, and maintaining the integrity of medical education, espe
mpared with traditional manipulator, the continuum manipulator has the advantages of a small moment of inertia and high dexte
urements, designed to be installable in tight spaces. We introduced two types of lenticular angle gauges to enhance pose estim
a specific sequence each time it receives a new perception task. Generating proficient combinations of vision methods to solve
y in achieving a high area coverage performance. Mass customization of a robot’s morphology would improve its productivi
d to task and motion planning for robots. We propose a system architecture that orchestrates a seamless interplay between mul
liant contact to model large body deformations. We rigorously characterize our approximations and present implementation det
g crowd simulation methods typically rely on either handcrafted rules, which may lead to overly aggressive navigation, or learnin
me this challenge, some research proposes simulators that incorporate hardware acceleration. Although these simulators impro
modeling and implementing different mechanical systems such as rigid and deformable bodies as well as their interactions via co
ulations, thereby missing the subtleties of authentic fish movements. In this study, we utilised Computational Fluid Dynamics (C
quality simulators that will enable the development of lunar robots. In this paper, we propose Omniverse Lunar Robotic-Sim (O
generation typically use rules-based algorithms that lack realism. In this paper, we propose SceneControl, a framework for cont
multiple frictional contacts. We use continuous collision detection to detect the time of impact and adopt the backtracking strateg
y hinders the deployment of virtual developed solutions without additional effort. Modeling complex systems like highly dynamic
work, we propose a tightly coupled keypoint detector and descriptor (TCKDD) based on a multi-task fully convolutional network
have short trajectories without loops that are unsuitable for the evaluation of longer-range SLAM or odometry systems, and da
M), the proposed Direct Imaging Sonar Odometry (DISO) system is designed to estimate the relative transformation between tw
dometry is designed by leveraging the Continuous and Ultra-compact Representation of LiDAR (CURL). Termed CURL-MAP (M
ust odometry solutions are of key importance. In this work a method for tightly-coupled LiDAR-Radar-Inertial fusion for odometr
ras and in any arrangement. Our SLAM system uses the generalized camera model, which allows us to represent an arbitrary
onments. We tightly integrate RGB-D images, inertial measurements, wheel odometer and GNSS signals within a factor graph t
articularly in challenging or data-limited situations, remains an unresolved issue. This paper presents HERO-SLAM, a Hybrid E
n-varied environments. We propose ONeK-SLAM, a robust object-level SLAM system that effectively combines feature points a
be limited while their influence and the long-lived landmarks are appropriately maintained. To do this, we propose a new two-ste
ts with longer lengths of sequences and agnostic of the scanning pattern of the sensor. While many approaches for loop closur
on. However, estimating normal and distribution of points are time-consuming tasks even with the assistance of kdtree or volum
o-point ICP approach that leverages the velocity information provided by 3D radar sensors. The second one is advantageous f
ingle outlier measurement, with possible catastrophic and meaningless results. Although recent works on robust optimization ai
eature extraction and leverage hierarchical refinement strategies to generate matched correspondences. The primary focus of t
ting the modeled object residuals into bundle adjustment. However, few have demonstrated better results than feature-based v
cross-task feature representations and lead to notable performance degradation. Regrettably, existing self-supervised SVO lac
east squares method for 6D pose prediction that employs rational regression. Our method can accommodate fixed data points
ational complexity of mapping and the limited accuracy of tracking. In this paper, we improve our previous direct pipeline Event
onment that encodes its different geometric elements with semantic attributes and the relational information between them. Sp
nition. It unifies the terminology of VPR and complements prior research in two important directions: 1) It provides a systematic
and low-light scenarios. The emerging 4D imaging radar technology is capable of providing robust perception in adverse conditi
establishing accurate tentative point correspondences between consecutive frames. Additionally, our semantic descriptors enric
per proposes a novel deep SLAM network with dual visual factors. The basic idea is to integrate both photometric factor and re-
no direct correspondences are assumed between point clouds. Instead, all points are merged into a global point cloud, whose
assumption without accurate model of ranging error. In this paper, we propose a high-precision and robust LiDAR odometry (LO
t, LONER uses LiDAR data to train an MLP to estimate a dense map in real-time, while simultaneously estimating the trajectory
ed by sequentially registering individual LiDAR scans and provide a robust pose estimate. Although effective, LiDAR-inertial od
generated. This paper introduces a real-time, computationally efficient multiple LiDAR-inertial odometry system (Multi-LIO) that
jects. Dynamic SLAM is an emerging field in SLAM research as it improves the overall system accuracy while providing additio
sentation and corresponding covariance estimation. Considering total map will contain a large number of coplanar features (kid
res to represent the scene geometry, and directly minimizes the natural Euclidean distance from each raw point to the respectiv
object removal methods are restricted to utilize only local scope information and have limited performance. To address this chal
e localization module first utilizes the preprocessed sensor data from the receiver, estimates the sensor pose online by register
ing a new robotic platform, researchers and practitioners alike usually mimic standard configurations or maximize simple heuris
by the feature attributes. In this paper, we propose incorporating hybrid point features efficiently into a single system. By integr
earning, the estimate still contains errors due to several factors. To compensate for this, we perform a shape alignment betwee
remove a portion of measurements. However, reliable data, such as odometric measurements, can be lost if measurements are
Jacobian and error vector resulting from linearization and hence, additionally affect the stability and convergence of the non-line
on properties when the wavelength increases, thus are not affected by low visibility. Hence, this letter presents ultra-wideband
ndence estimation, and residual functions are all strengthened by integrating the physical properties. We conduct experiments
en spaces, RGB-D SLAM approaches, such as L-SLAM, fail to estimate robust 6-DoF camera poses due to the limitations of th
d on incremental and general bag of words. Incorporating an adaptable keypoint and 3D local feature extraction method, we em
ultaneous Localization and Mapping (SLAM) system that harnesses the open-vocabulary nature of Large Language Models (LL
the-art Pose Graph Optimization (PGO) techniques. Departing from the conventional PGO paradigm, which primarily relies on
ual-Inertial SLAM system and 3D mapping framework applying local submapping strategies to achieve scalability to large-scale
the same map is even necessary. However, localizing e.g. a ground robot in the map of a drone or head-mounted MR headset
ne coverage and often necessitates extensive viewpoint sampling, particularly in complex scenes. In the paper, we propose to
amework which attempts to maximally reduce pose graph uncertainty during exploration and improves occupancy map quality.

e different measurement times is a major challenge of using event-based cameras. It is often addressed in visual odometry (VO
e design of an equivariant filter for the problem of visual-inertial odometry (VIO) including IMU bias, and camera intrinsic and ex
and relieve the incrementally accumulated pose errors. Firstly, we introduce a lane line tracking method that enables multi-fram
heir walking, while some pedestrians, denoted as non-cooperators, may try to block the robot. It is also challenging to identify p
LMs) have recently been used to enable commonsense reasoning, grounding this reasoning in the real world has been challen
mented strategies have been developed based on an ethnographic study on assisted shopping and tested in a lab experiment w
neration for robot navigation as an automated planning problem considering different possible explanation attributes. Our visua
tes when crowd configurations change, i.e. become larger or more complex. Thus, it is crucial to fully understand the complex,
assess both user experience and the technical feasibility of this approach, experiments were conducted with human participant
iment with 28 participants, we assessed response latency and utterance lengths to interactive story narrations delivered by eith
ross-cultural communication. This framework paves the way for a social robot to act as a mediator between groups of schoolch
arge, it requires a large amount of user feedback. ChatGPT is a powerful generative AI tool based on large language models (L
and reconfigurability in electromagnetic devices. The prototype employs a continuum robot as its supporting structure, featurin
nt of time for their motion planning, posing a hurdle to their practical implementation. To tackle these challenges, efficient motion
paper, we report an ultrafast gripper based on tunable bistable structures. Different from current designs which are only suitab
ming effect.Here,we present the design, fabrication, and tests of a novel soft actuator with good restorability and compliance. To
uipment, such as specific diving robots, engineered to withstand the high pressure. Our research aims to reduce the costs of de
h mechanism and flexibility of vine robots, while keeping the tip of the vine robot free of encumbrance. Our implementation is a
or estimators for the peg-to-hole pose. In contrast, we use a partially observable formulation and deep reinforcement learning f
nt is performed at the tool-center-point of a soft continuum manipulator and is realized by a hybrid force-position control. For th
ape and size, and delicate harvesting is among the first concerns for fruit quality. Various soft robotic grippers have been propo
resses this challenge by presenting a lightweight (2.57 g) soft wall-climbing robot with integrated linear, turning, and transitionin
gn concept for fabrication of modular, soft truss robots and demonstrate its varied capabilities in a tetrahedral robot (Tetraflex).
hieving high payload and positioning accuracy remains challenging in existing soft manipulator designs. In this work, we propos
ith living beings through the deformable interface. However, mechanical and morphological requirements limit the design and im
nfolding. This paper proposes a solution in which an extensible continuum robot (ECR) carrying the unlocking device enters the
otic intervention. Inspired by the motion of the chrysalis during its transformation, we designed an innovative pupal-mode actua
uipped with wheel-legs capable of multimodal locomotion to negotiate various obstacles. The soft robot consists of a pneumatic
g conditions. These challenges in maintaining the blackberry supply motivate the development of an autonomous robotic solutio
r robots often tethered, i.e., it cannot integrate energy and control systems onto the body, which greatly limited in its working ra
displacement behavior of planar flexible-link mechanisms. To do so, we replace a subset of rigid with flexible links in an existing
etic field in ideally any manner. A promising concept to realize such a surface leverages signal refraction via antenna edges, bu
t soft palm, which enables more surface area contact for the objects that are pressed into it. Moreover, this design, along with t

adaptive grasping of small surface asperities is overlooked. In this paper, we propose a novel method to achieve adaptive grasp
a rapid design and fabrication method for string-driven origami structures and robots. We developed an origami design software
d robot composed of ferrofluid that shows promise for small-scale magnetic manipulation applications. The objective of this wor
tions, such as soft grippers, walking robots, crawling robots, climbing robots, swimming robots, etc. Although previous research
reinforced by lattice metamaterials. We develop an analytical framework to model the twisting, bending and elongation finite de
driven manipulator with continuous stiffness control capability and multiple motion patterns (bending, rotation, and elongation). F
gn of a multi-material 15 DoF robotic hand with five fingers including an opposable thumb. Our design has 15 pneumatic actuato
ension glove with thumb assistance poses a challenge for researchers. This paper introduces a fully wearable soft hand extensi
 o silicone rubber laminates. When inflated by compressed air, the folded corners of the flat tube tend to unfold, resulting in the
ety of retraction mechanisms have been developed, many necessitate the installation of extra rigid hardware at the distal part, c
rgery are costly. One promising inspection device inspired by plant growth are “vine robots” that can navigate cluttered
ensors. Additionally, soft pads are strategically placed around the ball joint to adjust the module stiffness irrespective of its posit
ting widespread adoption of this technology. To tackle these problems, this work details the design and experimental testing of
UAVs applications. In this paper, we designed, fabricated, and characterized a whisker sensor unit based on MEMS barometer
d resolution and are difficult to scale. Here, we present a soft robotic gripper made out of two cubic lattices that are sensorized
ucture of pneumatic actuators using Bézier curves and Gaussian Mixture Points, to have a simple representation of the actuat
they have high flexibility and safety, there are problems such as poor fit and low output force. In order to solve these problems
n lines for considerable amount of time. In this paper, we adopt the abundantly occurring spiral motions encountered in nature a
is designed to achieve a wide range of stiffness while maintaining a compact form factor. The structure incorporates rigid retain
gical retraction. Using origami-inspired design and fabrication principles, the OSM is easily folded and integrated within a soft ro
atures in two bending directions, enabling 3D shape sensing and tip tracking of the continuum robot. The robot is able to bend a
ctricity to classify finite gestures, rather than continuous (infinite) force detection, which greatly limits the use of prosthetic scena
plish this, we use a widely available, off-the-shelf resistive sensor and multi-tap this sensor, i.e. make multiple electrical connect
c imitation of contact from human demonstration using a wearable finger-tip sensor. By developing a multi-modal sensor (provid
motion control framework for DEAs capable of accommodating different configurations, materials and degrees of freedom (DO
f novel control algorithms, these devices provide exceptional opportunities to visualize the internal structures of the human body
astomer tube filled with different types of fiber. The three types of fibers play different roles in maintaining the structure, variable
allowing an organism without a central brain to successfully search for resources (e.g., light). Yet to date, vine-inspired robots h
o date, design of the morphological shape (inverse problem) has not been fully addressed. This article outlines a geometry-me
compatible with soft robots have been identified such as combustion and chemical reactions. A further milestone to such syste
nhindered patient/user movement is a priority, to move towards personalized and decentralized healthcare solutions. In this wor
on exchangeable multi-rate real-time control components with standardized interfaces. On the hardware side, the use of EtherC
n successful in other domains such as large-scale customer-oriented web services (e.g., Netflix). We recommend to employ sim
easier and model sampling efficient. It also provides a language for specifying the relationship between the model and the actu
ents, and task descriptions. Our benchmark suite is especially useful for modular robots, where the multitude of robots that can
adopted and standardized testing methodology. Existing GSL evaluations vary in environment size, wind conditions, and gas si
s the temporal richness of cooking recipes could enable the extraction of unambiguous, robot-executable plans. In this work, w
ke ChatGPT, to automate this process, leveraging their ability to understand natural language instructions, generalize examples
dently. We argue that this approach is better suited than the state of the art (i.e., synchronous execution) for flexible/adaptive ro
ce of obstacles and/or situations not commonly found in indoor settings. One such example is given by tall grass and weeds th
vant to resolve, however, the spatial resolution of the available image data is often insufficient - particularly in the poorly illumina
oduces an innovative approach that combines Proximal Policy Optimization (PPO) with Model Predictive Control (MPC) in the z
oF (thus complex) system -- for a wide range of launch scenarios. Here, scenarios are defined by the relative location of the de
e. OST aims to close this gap by training a pose estimation NN online using incoming flight images during Rendezvous and Pro
problem and empowering the robot to learn pre-grasping policies through trial and error. Our methodology incorporates an off-
o real target datasets, algorithms are often trained using synthetic data and applied in the real domain, resulting in a performanc
ntation, our approach effectively analyzes hazards such as obstacles and hazardous slopes in both terrestrial and aerial naviga
s OD using data from an inertial measurement unit (IMU) and a low-cost RGB camera. Machine learning techniques are used to
which identifies all parameters required for the reconstruction of system dynamics despite unmeasurable sloshing and modal s
on. This paper presents a comprehensive solution for on-orbit construction tasks, encompassing a novel standard interface, do

ot becoming an advantageous option for all-terrain exploration and surveying. This paper focuses on a novel iteration of such a
nterconnect is used to mechanically couple the involved systems while providing electrical power and data transmission. The p
se flat surfaces, the gripper must ensure contact between the gecko adhesives and the surface before applying the appropriate
f a closed spherical shell structure, which makes the capacity of carrying workloads weak. At the same time, the single point co
of flexible appendages, such as solar panels, or fuel sloshing may introduce significant end-effector errors. Therefore, controlle
c monocular RGB image. It is robust, initialization-free and runs at 10 fps, thus forming a promising tool to improve robotic and a
ost of simultaneous imaging processing, identification of mechanical properties, and motion planning, leading to a need for less
work for sensorless transparency optimized four channel teleoperation. It is sensorless because forces are estimated from exis
s of kinematic and video data. We conduct an ablation study to evaluate the impact of fusing different input modalities and their
sk into a sequence of sub-tasks with increasing complexity and utilizes an actor-critic algorithm with deterministic policy output.
nprehensile manipulation by implementing a vision-based reinforcement learning agent for rolling a block, a task that has comp
rs based on visual feedback. In recent years, supervised deep learning algorithms have shown increasingly better performance
ries with improved outcomes through smaller incisions. Recent efforts are working toward making robotic surgery more autonom
oal-conditioned tasks due to the intricate compositional structure, which requires decision-making for a sequence of sub-steps a
roperty, pose challenges, including the risk of bleeding. This paper introduces a 4 mm diameter robot-assisted bronchoscope w
ion of robotic manipulators and create a significant challenge for applications that rely on precise interactions between manipul
g in shape and appearance throughout the surgery. To address this problem, we exploit optical flow to naturally capture the pix
al constraints and allowing for model prediction and control. However, real soft objects in robotic surgery, such as membranes a
sing. In this work, we propose an automated sensing method that maps 3D subsurface anatomy to provide such geometric kno

bility to generalize to unseen subjects and tasks. Kinematic data from the surgical robot is particularly critical for safety monitor
With a high complexity, this task is an ideal candidate for Reinforcement Learning (RL) in simulation. First, we present an inter
k with photorealistic rendering in NVIDIA Omniverse. We provide 14 benchmark surgical tasks for the da Vinci Research Kit (dV
uracy 3D-2D registration techniques and computational efficiency requirements in intervention robot applications. We categorize
ated and translated, the backbone elongates, twists, and bends with a dexterity that is advantageous for confined spaces. Tube
PPR kinematic models assumed two tubes, planar shapes, no torsion, and no external loads. In this paper, we develop a new, m
nd reinsertions. To address these shortcomings, we introduce an optimized needle shape-sensing method using a fully distribut
scopic imaging for catheter localization during navigation, which risks creating radiation-induced injuries to both the patient and
s manuscript presents our recent work on force-perceptive safe semi-autonomous navigation of the ValveTech platform prior to
ture operations. In this paper, a dual-arm robotic needle insertion system with guidance of intraoperative 2D ultrasound (US) an
ent distances. It does not require an external localization system or inter-agent exchange of state information. Our approach us
e set of rules to identify and trail a dynamically changing source gradient. To validate the effectiveness of our approach, we con
oblems, as demonstrated by the ARK system (Augmented Reality for Kilobots), which enhanced the experimentation possibilitie
alized shells. Few current platforms fulfill all the above characteristics. Here, we present Morphobot, a robot platform for morph
d attachment reliability while also reducing the number of motors from six to three. We also expand the sensory capabilities of F
exert attractive and repulsive forces on each other. Moreover, by controlling agent-specific coupling forces on-demand, we can
nteracting only with a static environment and without the involvement of other active entities. In this paper, we investigate the d
rones scales, possibly into the hundreds. Software-in-the-loop simulation is an important component in verifying that drone flee
that are scattered in the environment, carry them to a goal location with load constraints, and optionally have access to a conta
 obot. However, state-of-the-art TAMP algorithms do not scale well with the difficulty of the task and require an impractical amou
ormal geometric algebra, we aim to further enhance the geometric expressiveness and thus simplify the modeling of various ta
torial challenges, limiting their applications for online replanning in the real world. To address this, we propose Dynamic Logic-G
ta is classified as potential objects, walls, and obstacles after outlier removal. Segmented point clouds are then transformed int
hs, or full optimal manipulation trajectories). We consider the general problem of how to schedule compute effort in such hierar
nd quantitative aspects. Most existing model-based GSL algorithms rely on lengthy stops at each sampling point to ensure accu
GP does not scale well to high-dimensional systems (e.g. mobile manipulators) and can suffer from obstacle avoidance issues d
planning problems. In this paper, we propose to search over sequences of easier pick-and-place subproblems, which can lead
d on Large Language Models (LLMs); 2) the seamless generation of temporal parallel plans for multi-robot systems through a s
must consider its past, i.e., historical data, when making decisions in each step. Nevertheless, recent approaches in EIF often n
We discuss essential parts of the logic, the architecture and working principles of NARS, and the integration of NARS as a ROS
responding to solution trajectories—as a hybrid constraint satisfaction problem. The proposed algorithm follows several design
e this problem into a Traveling Salesman Problem (TSP) variant with timing and precedent constraints, and we solve it as a Mix

challenging problem. In this paper, a Probabilistic Integrated Task and Motion Planner (PROTAMP-RRT) is presented. The pro
Our interest in this problem is motivated by applications in precision agriculture, where robots are used to collect measurements
anning approach that considers reachability conditions as task constraints of both perception and manipulation tasks for the mo
nges in perception, reasoning, manipulation, diagnostics, fault recovery, and goal parsing. RAMP has been designed to be acce
ently, the standards specify well-reasoned force and pressure thresholds for blunt impactors and advise avoiding any edges an
g over an agent's lifetime by integrating abstract strategies: a generalizable planning experience encoding introduced in earlier
rior to this work, little has been done to show how plans acquired from FOON can be executed by a robot, as the concepts in a
(LLM) to retrieve recipe instructions and then utilizes a fine-tuned GPT-3 to convert them into a task tree, capturing sequential
onal methods aiming to find globally optimal solutions, especially for some large-scale problems, incur extremely high computa
iciency of the system. In the task of multi-robot exploration of unknown environments, the strategies of frontier point detection a
nd easy-to-use human-robot interface is paramount. Recent works introduced a novel teleoperation concept, which relies on a v
rn and generalize. In this work, we propose to enhance learning performance by introducing a network-aware confidence weigh
e workspaces. The system consists of a pair of lightweight and compliant anthropomorphic dual arm manipulators (LiCAS) in le
sly to decrease the demand on the operator and increase the task completion rate. However, decision-making for mode-switch
sive environment for interacting with robots, effectively reducing the mental and physical workload of operators during teleopera
continuous transmission of environment impedance. Moreover, it allows the local model to generate accurate force feedback an
user workload and required prior knowledge through the automation of secondary or auxiliary tasks. The introduced novel meth
workers using a long stick tool. Thus, implementation of robotic teleoperation system is in demand to ensure safety of workers
mensional input signals. However, designing SCTs is a laborious task requiring robotic expertise. To facilitate their design, we p
n, we propose TELESIM, a modular and plug-and-play framework that enables direct teleoperation of any robotic arm using a d
e data and a deep learning model. By integrating (i) and (ii) in a task-space whole-body controller, we achieve full-body synchro
ly, we focus on two aspects of the place task: stability robustness and contextual reasonableness of object placements. Our pro
an eye-to-hand camera. Such algorithm is instrumental in contexts where quickly adjusting the perspective to see objects obstr
ubiquitous devices. We demonstrate and evaluate the effectiveness of iRoCo in practical teleoperation and drone piloting appl
truggle to complete tasks as competently as they could in person. We present a framework for incorporating open-world shared
eeded, relying on behavioral observations. The framework presents an intention estimation technique at hierarchical levels i.e.,
tion, built upon three key contributions: Firstly, a teleoperation framework employing a Human Machine Interface (HMI) and a b
gate such limitations by assisting the users in teleoperation. It aids in teleoperating 3D motions, such as helical motions, which a
urden, which becomes inefficient and unacceptable as the size of pretrained models scales. Although efficient fine-tuning appro
m significant class-imbalanced performance, restricting their adoption in real applications. This imbalanced performance is main
as typically focused on domains with similar state and action spaces but differing in other aspects. In this paper, our primary foc
s from three aspects. Firstly, a PromptLearner module is devised to replace hand-crafted prompts with learnable contexts to au
erception, we figure out the significant domain gaps existing in typical real-world cross-domain scenarios and make the first atte
el serves as an objective function, that we maximize to identify successful grasps. Key to our approach is the utilization of learn
controllable solids, existing RL algorithms meet difficulty to accomplish training within a reasonable timeframes. In this paper, w
sical environment is often infeasible. With the development of increasingly realistic simulators, it has become feasible for indust
lf-expression, enhance social interaction, and assist users in their movement and performance. The “Yes We Dance!” in
und Reaction Force(GRF) effectively. Each arch of the proposed foot consists of passive joints and springs and is designed wi
hich allows a chain-like robot composed of these modules to assemble diverse robotic configurations through folding and exhibi
riable curved components, particularly for non-magnetic components such as aircraft skins, wind turbine blades, etc. The adso
ous task insertion with a robotic manipulator. The plate does not go into the groove inside the frame straightforwardly and need
use of proprioceptive actuators in the dual-arm manipulator, enabling force controlled, dynamic and dexterous cooking actions
t previous Wi-Fi sensing technologies require two+ separate devices, for instance, one transmitter and one receiver. These mu
e machines and robots are environmental guardians, forming Earth's artificial immune system. Powered by AI and democratized
rticularly in the industrial context. While shown at scale, this works seeks to contribute in addressing the difficulty in interacting
ment in both moving and stationary platforms. Many existing 3D human perception systems have limitations. Depth-aware sens
especially suited for Human-Robot Interaction (HRI) given their sensing capabilities. This video shows the implementation of thr
terity never before shown by any humanoid. This is especially true for autonomous dextrous manipulation with its multi-fingered
d hoc agent's behaviour. It is often difficult to pursue such an approach in complex domains due to the lack of sufficient training
ns) and signal calibration, which is typically not feasible in many robotic applications. Moreover, the wireless signal is often bloc
ject. Communication with the robot is done using a LAN cable. By connecting this cable between the ground-based PC and the
animal tracking, particularly those equipped with multibeam imaging sonar (MBS) unaffected by underwater light and turbidity. I
t grippers are not applicable to objects with sharp sections such as broken valves and glass fragments, especially on fire. In thi
a patients. However, the number of caregivers available to provide such care is not rising fast enough, so any new forms shoul
ery. Nowadays percutaneous procedures are becoming more popular because of less trauma and shorter hospitalization time,
ents in the personal robotics field. In this work, an affordable (parts cost of around 550$) 3D printed robot consisting of a mobile
hanics and robot dynamics. To address this issue with LfD, the Dynamic Movement Primitives (DMP) framework offers a promis
only one active degree of freedom (cable-driven knee joint), BALLU is capable of forward and backward locomotion, jumping, tu
must collaborate, share data, and offer services, such as interplanetary communication. However, the unpredictability of outer s
can be a significant physical and mental burden on patients. One potential solution to this, inspired by the reflexive functionality
but operated in Unity. The kinematic model of the robot provides chain to move the manipulator, and the Unity model updates
wer stations, and ships. Existing climbing robots have many shortcomings that limit their practical use. Wheeled climbers have
objective, we envision establishing a service humanoid robot named GARMI that can be utilized in various application fields, inc
ns. In this video, we present a soft robotic gripper that is to our knowledge the first to integrate the superelastic nitinol wires wit
mpared to their traditional rigid counterparts. We present a tensegrity robot comprising three rigid struts, six motor-cable pairs fo
xpected that movement will be generated with less kinetic friction with the outside world. The world's first device to propose this
ists of a silicone rubber shell encasing paraffin phase-change material. It is heated by Nickel-chromium alloy wire. Heat is trans
surgical robot force sensing, visual reconstruction, and autonomous robot control. We hope this can provide valuable insights fo
sion tools was proposed to better adapt to variable tasks, which is multi-assistive for the wearers. The conversion time between
laboration between the robotic exoskeleton and its user. In particular, the stiffness parameter of impedance control, which dete
where human limbs are heavily occupied with primary tasks, proposing additional means to control and perceive SRL poses a
arned structured latent space to perform planning. Given multiple LSR framework instances, that differ either on their latent spa
ured by event camera. To enable reliable and accurate tracking of objects, we use a new event representation and predict brigh
applications in robotics due to a lack of sufficient accuracy. To remedy this, we propose a new method which boosts the accur
, cost, and increased points of failure. Alternatively, deploying an external camera in the environment is very easy to do, and m
c visual cue such as bounding boxes. This compositional-semantic-contextual hierarchy is flexible to be integrated in different a
amplified by extreme camera motion, sudden illumination changes, and strong occlusions. Most modern trackers rely on the ap
atisfactory results by fusing UWB ranging measurements with visual-inertial system. However, these approaches still have som
g accuracy and robustness, our approach offers a different solution. We realize the crucial impact of precise gyroscope bias es
obotic mapping which, due to computation constraints, limit map resolution or scale, and GPU-based reconstruction systems w
syntactically combine these components. Particularly, we propose Exploitation-Guided Exploration (XGX) where separate mod
antly focused on motion planning as opposed to robust motion control. In this paper, we propose a novel T&R system based on
omplex object-level queries. In addition, we design a novel LLM-based augmentation and prompt templates for stability during tr
ent is the localization of each mobile furniture. The challenge is to assess the state of an arbitrary living environment so that the
roach takes inspiration from BADGR, an LSTM-based network that primarily concentrates on environment modeling, but disting
ps? In this work, we propose the Coarse Map Navigator (CMN), a navigation framework that can perform robot navigation in un
tance recognition necessitates increasing marker size and camera magnification to ensure the required resolution. Furthermore
previous works have focused on developing low-cost optical flow strategies, their estimation quality has a noticeable gap with m
his approach ignores objects in the environment and learns data bias to fail navigation. According to our preliminary findings, mo
e is still limited by the scarcity of suitable training data and they suffer from poor computational scaling. In this work, we present
ators often rely on discrete categorical one-hot vectors and vote counting strategy to construct graph representation of the scen
. We further enhance the framework's practicality by introducing POLoNet, a neural network trained to approximate the comput
ing approaches. The gap between RGB space and map space makes it difficult to directly transfer the knowledge from foundat
h data for new training data can be prohibitively expensive, especially in complex or GPS-deprived environments. To address t
or structured geometric obstacles but have limitations regarding the semantic interpretation of different terrain types and their af
erest (OOI) while avoiding obstacles, all without relying on localization. UIVNav utilizes imitation learning and draws inspiration f
es due to lighting variations remains a fundamental problem in visual terrain-adaptive navigation. Existing solutions either requi
ware network (ASP-LED) with structural priors, including high-frequency and edge, to enable effective image representation lea
ary category after learning from a support image. An off-the-shelf petrained ViT is directly deployed for generalizable and transf
on process and sometimes ambiguous labels have severely constrained the usability and scalability of 3D occupancy models.
ded in the learning framework. This paper presents a novel optimization method based on recurrent neural networks to further
e from in-the-wild images and videos without ground truth supervision. Given a pair of images, it estimates the dense flow field e
engineering challenges and costs. Recently, Multimodal Large Language Models (MLLMs) have emerged as novel backbones f
ion. To address these issues, we propose a self-supervised learning method that leverages the cross-modal relationship betwe
d from inertial measurements. In particular, we introduce a Full-Scale depth prediction network named FSNet. FSNet contains fo
l Sensor (APS) devices such as RGB cameras, DVS suffers from a dearth of ample labeled datasets. Prior efforts to convert AP
f scenes has the potential to enhance models' ability to deal with the adverse effects of occlusion and the dynamic nature of the
mains challenging in robust generalization due to limited data resources and spectral differences between thermal and RGB ima
open-set scenarios. Such algorithms often falter in the presence of test samples originating from classes unseen during trainin
 g inference. While existing self-supervised methods have made strides in this field, they often suffer from inefficiencies arising
er, they have limited performance in practical applications due to their inherent noise in event data. This paper advocates fusing
earch topic in both robotic and computer vision communities. This paper introduces an innovative and elegant framework name
g the daytime, but perform poorly at night. To address the limitations of both image modalities, recent methods have employed t
ents bring significant challenges to the annotation. Simply projecting the Lidar-labeled 3D boxes to the image leads to non-trivia
non-annotated videos. We show that our self-supervised models can reach the state-of-the-art performance 'without bells and
m BEV maps is essential for portraying complex traffic scenarios and enhancing driving algorithms. Concurrently, diffusion based
often lead to diminished quality in reconstructions and extended durations for both training and rendering. To tackle these chall
, we utilize motion, geometry, and density priors behind event data to impose strong physical constraints to augment NeRF trai
ing 3D positions and camera orientations. On the other hand, LiDARs excel in generating accurate 3D maps but often come at
cations demand an extended view that faithfully represents the scene. To achieve this, we formulate a new problem of faithful F
or two limitations. First, few representations on $SE(3)$ are smooth and differential, making them difficult to apply in deep lear
t states and the model is typically trained with a singular diffusion loss. This paper explores the potential enhancements in such
emonstrations, as well as generalizing them to categorical objects in novel cluttered scenes remain unsolved challenges. In this
proach that leverages NeRF to synthesize new viewpoints within the expansive lunar environment. By blending 3D hash grids a
rmation capture strategy and heavily impacts the quality of the 3D model generated from the captured data. Recent methods h
aces, often face challenges like damage or contamination. These issues can lead to compromised marker integrity, resulting in
provide localization and clock signals. However, hardware-generated signals could be vulnerable to noise and potentially malici
limitations, we propose a method to reconstruct the clear 3D scene implicitly from multi-view images degraded by water- drops

een rendered and target images. However, during inference, they require long converging time, and suffer from local minima, m
are not visible, is still challenging. We propose an occluded part-aware graph convolutional network (OP-GCN) to address this c
static scene and their performance tends to degrade in dynamic environments. To address this issue, we present Motion-Aware
ds tend to perform well in synthetic hazy scenarios but struggle in real-world situations. It stems from the fact that these method
s of full-supervised methods limits the capability of the visual BEV semantic segmentation, which usually needs HD maps, 3D o
man-level cognition capability. However, there are few open-vocabulary auto-labeling systems for multi-modal 3D data. In this p
on tasks, termed the Segment Anything Model (SAM), has been introduced. This model substantially addresses the dilemma o
enhances vision-based reconstruction under extremely low light conditions. The approach locates features with well-defined sca
sparse spatial information for pixel-wise tasks and difficulties in achieving paired depth labels for temporally intensive spike stre
the provided input point clouds. The PillarGen model performs the following three steps: 1) pillar encoding, 2) Occupied Pillar P
metry devices, should be relocated in response to rapid change in mechanical structure and hardware integration. Nevertheles
singly popular. However, these methods encounter difficulties such as high noise sensitivity, intricate computational procedures
er than the manual annotation of real data, we intend to mitigate common challenges in state-of-the-art VMMR datasets, namel
SAC methods to calculate the poses of ordinary objects. However, the metal parts in industry are reflective and texture-less, ma
ustic field's position and orientation, consequently diminishing the adaptability and effectiveness of subsequent acoustic microm
he bin and detecting suitable parallel grasps in PGAs from point cloud data. To support our work, the training datasets are gene
em require supervised learning with datasets with annotations of objects. Since it is difficult to annotate all stacked instances in
are classified via bird's-eye view scene crops, where we project the detections of a 3D object detection model onto a context m
e rotation-equivariant detection and invariant description by design, many learning-based approaches learn to be robust only up
the VQA system by a sequential data stream from multiple resources is demanded in robotic surgery to address new tasks. In s
eration. Surgical robot is characterized by high precision and stability. However, navigating intricate intraocular environments an
 directly mapping image space errors to joint space actuation. We achieve convergence without a-priori knowledge of the targe
ve because they can adapt to various food shapes. However, if internal sensors such as tactile sensors and force sensors are u
ual servoing task as graph traversal. Our method not only extends the robustness of visual servoing, but also enables multitask
ay be unsatisfactory, and the velocity signal itself hinders the connection between the feature velocity model and the robot's dyn
le contour model allows adaptation to the evolving target's shape, while the proposed event-triggered scheme achieves improv
This makes our approach have full control over the deformation of elastic deformable objects of any general form (linear, thin-
trainable vision system. To address this, we propose a novel architecture called Darkness Clue-Prompted Tracking (DCPT) tha
grounded in foundational models for universal object detection and tracking, moving beyond specific training categories. Integra
drones. This decrease is attributed to the distinctive challenges in the MOT-on-drone scenario: (1) objects are generally small
ssumptions and rules to represent object permanence in multi-object, multi-agent interactive scenarios. We integrate these rule
odels have shown promising advances in detecting and segmenting objects and parts in 2D static images in the wild. This raise
point cloud data. Instead of performing correlation of the two branches at just one point in the network, in this paper, we presen
. These include methods trained with warp error (i.e., color constancy) combined with smoothness terms, and methods that enc
rs in ambiguous association decisions. Our method, Sliding Window Tracker (SWTrack), yields more accurate association and
t the incoming detections with no sense of their associated localization uncertainty. This lack of uncertainty awareness poses a
localization, a new fast dense 3D model-based tracking using wide-angle depth image is developed and used in conjunction wi
edge devices with computational constraints. In response to this, we introduce LiteTrack, an efficient transformer-based trackin
hods in this environment. In this method, we first propose a concept of quantitative windows to describe the robot’s trajector
r method calibrates and adapts the dynamics model online to improve the accuracy of forward prediction conditioned on future
g frame-to-frame motion, it tends to fail in case of significant intensity changes and motion blur. To overcome these limitations, w
NSS measurements within the square-root inverse sliding window filtering (SRI-SWF) framework in a tightly coupled fashion, wh
fisheye images and IMU data. Then, we implement a lightweight panoramic depth completion network based on multi-basis de
e consuming low power, making them ideal for mobile robotics. FPSPs, such as the SCAMP-5, use parallel processing and are
ments. Our approach utilizes line-based event features to provide additional structure and constraint information in human-made
ue to repeated perturbations in pose graphs. We address this overhead with a novel methodology, leveraging strategic graph pa
measurement unit (IMU). We introduce an improved IMU pre-integration formulation based on the exponential function of the au
on improving the speed and accuracy of perception, enhancing planning adaptability for the real world, and providing swift and
unding point tracks the forward shift of the wearer's center of gravity (CoG) during STS. Since the strategy keeps the distance b
n this letter, the neural delay dynamics of human stance are modelled and embedded in the control of a supernumerary robotic
where active mode is utilized for aiding in standing up (power generation), thus the energy flows from the exoskeleton to the use

ndividualized scheme is also important to ensure that the robot suits the specific conditions (e.g., movement habits) of a patient
ck exoskeletons are often bulky and require electric motors, making them challenging to use and consuming significant power.
d wearable robotics, necessitates the measurement of three-dimensional GRF vector information, traditionally achieved in indoo
ith abundant exoskeleton assistance experience have obvious advantages in speeding up the parameters exploration process
exoskeletons involves the misalignment of joints between the exoskeleton and the user, resulting in discomfort and potential in
ive disturbance rejection controller (AADRC) such that the orthosis provides the only required amount of assistance to complem
rying parameters of the L1 control law, ensuring precise tracking of the AAFO-wearer system by the state estimator. The projec

full advantage of the rapid switching capabilities of the leg-wheel design of the robot. The robot motion is designed based on a
ent unit (IMU) and encoders, allowing estimation without using the ground contact state as the a priori. In this paper, the initial e
 rolls without slip and maintains a single point of contact. Using an auxiliary coordinate system to define the rolling body and terr
y optimizer for generating trajectories over a large class of challenging terrains. By limiting the underlying modeling to the plana
ion of the design. Kinematic analysis is conducted based on a generalized four-bar mechanism driven by a coaxial hub arrange
wheel forces (WFs: longitudinal, lateral, and normal forces for each wheel) necessary to execute the planned trajectory withou
trol inputs by two steerable driving wheels are required to implement transformations between the two modes due to the differe
nomic chained-form system, the proposed method is designed based on the original tracking error equation, and therefore our a
ork, we propose an optimization-based reactive Landing Controller that uses only proprioceptive measures for torque-controlle
as object manipulation can introduce substantial disturbances that impact the robot's locomotion. This paper presents a novel fr
ling down (aka “self-righting”), an essential early milestone for babies learning to walk, and an existential capability for a
ocusing on link lengths and antigravity joints that contribute to motion, with constraints placed on the contact conditions between
y exploits the concept of three-dimensional divergent component of motion and passivity-based whole-body control. The latter f
ink with high accuracy and efficiency while ensuring the mathematical continuity and differentiability of SDFs. We further levera
activation matrix is proposed to continuously shape both null-space projectors and task-level control actions. Validation has bee
ses a reinforcement learning task specification that focuses on matching desired contact force levels. Integrating force control w
To mitigate this limitation, researchers have proposed shared-control methods incorporating robot decision-making to aid hum
m of domain-agnostic scene text spotting, i.e., training a model on multi-domain source data such that it can directly generalize
oach to improve point cloud domain adaptation by employing masked representation learning in a self-supervised manner. Spe
n re-identification module that uses continuous visual adaptation techniques. This module ensures that the robot can seamless
eir need for resource intensive point cloud registration between the query and each candidate. In this work, we propose an effic
owledge for visual and textual encoders. SG supervision generates fine-grained captions based on various graph augmentation
se a lightweight parameter-shared network (LPS-Net), which includes multiple bidirectional perception units (BPUs) to extract m
egularization analysis from the perspectives of response- and background-learning. Specifically, to address response degradat
mplate matching. This adaptation for pose estimation still requires expensive data rendering and training procedures. MegaPose
ly on a few frames from the original video to reduce computation, so temporal information is compromised by low frame rates. S
equirements in different exoskeleton-assisted locomotion modes, such as running, walking, sit-to-stand, and stand-to-sit. The V
energy consumption for a given pick-and-place task. Through simulations and experimental validation, we show that the optima
stiffness joints (VSA) has become crucial. Among the prevalent approaches for stiffness adjustment, lever mechanisms have b
is used to regulate the position of the pivot located on the Archimedean Spiral Relocation Mechanism (ASRM), so that the stiffn
ly, it allows for full link decoupling, while simultaneously being able to lock the spring in the drive train via a switch-and-hold me
mpares two prototypes on run-in behaviour and a number of quantitative metrics including transmission error, play, friction, and
ate modeling. The origami framework also improves the strength of bearing negative pressure, and thus can provide bidirection
ple for comanipulation. The design is presented. A prototype, developed in the medical context, is implemented and characteriz
getically conservative mechanical elements useful for building robots that require no energy when performing cyclic tasks. How
determine a synchronized collision-free trajectory for the three marsupial system agents: UAV, UGV, and tether. First, we prese
safety of UAVs. To address this difficulty, this study proposes a cooperative planning method for heterogeneous UAVs by ensu
otal energy consumption of a UAV and the perception quality of light detection and ranging (LiDAR) sensor mounted on the UA
d analogies between both domains in terms of floating nature, attitude dynamics, and thrust-wrench relation through the mixer m
h this fundamental problem, we propose, for the first time, on-device learning aboard nano-drones, where the first part of the in-
n of aerobatics trajectories that exploit the entire flight envelope, allowing agile maneuvering through the stall regime, sideways
s leveraged to explore the unknown model of uncertainties, which is subsequently utilized in an evolutionary model-based distu

full autonomy is reliable obstacle avoidance, which has proven to be challenging by itself in a generic indoor environment. Curr
etching through the integration of behavior cloning and reinforcement learning techniques. Our approach trains painting policies
hes the robotic hardware performance limits to play table tennis. In particular, we propose a pipeline to generate manipulator jo
itially, the system refines prior estimates of string positions, provided by kinesthetic teaching or visual estimation, through repea
ate and graduate experiences. Transformative advancements in robot navigation over the past decades have led to a significan
nt a novel learning-based approach for modelling the downwash forces that exploits the latent geometries (i.e. symmetries) pre
e observation that real-world disturbances fall into two categories: 1) those that can be directly monitored or controlled during tra
tem that incorporates a wide-angle camera, a high-speed search camera, and a galvano-mirror. In this system, the wide-angle
in a dynamic, narrow, and complex environment. We propose a complete software framework for autonomous robot perception
n. By generating PDs from language instruction and scene observation, we can drive symbolic planners in a language-guided fr
alman filter. This paper proposes a trust-region policy optimization method for training NeuroMHE. We achieve this by providing
ar viewpoint. Notably, we do not use defect-free samples. Instead, we use virtual results to construct a defect library. MCDEN d
ntly, many works have tried to utilize the strong in-context learning ability of Large Language Models (LLMs) to achieve general
ble interaction between human instructors and planning systems. However, task plans generated by LLMs often lack feasibility
his paper introduces a hybrid approach that exploits neuromorphic computing in combination with probabilistic inference to addr
spects by designing a variational auto-encoder framework that integrates graph-based representations and time-sequence mo
covering those features for scene flow tasks. Our paper introduces DeFlow which enables a transition from voxel-based feature
reinforcement learning. This method embeds the principles of equivariance as inherent patterns in the learning process under g
computational resources. Therefore, this paper proposes a partially observable ECAS (PO-ECAS), bringing this work one step
h that combines adversarial imitation learning with large language models (LLMs). This innovative method enables the agent to
tudy of learning expert reward functions is largely under-explored in legged locomotion. This paper brings state-of-the-art Invers
uarantees on the false positive rate --- i.e., when there is no distribution shift, our system is very unlikely (with probability < eps
d to introduce a belief-dependent constraint. Prompted by this demand, in this paper, we consider a recently introduced probab
switch between base policies with different characteristics, e.g. safety or speed, or directly augmenting a policy observation wit
This paper introduces a novel resampling-free particle filter designed to mitigate particle deprivation by forgoing the traditional r
bilities in models, while coverage-based testing aims to achieve comprehensive testing but overlooks application scenarios. Mo
s assistance services. The proposed system allows to dynamically generate and adapt in autonomy heterogeneous robotic actio
ollaborative robot or industrial robot arm to perform food-slicing tasks by adapting to varying material properties using complian
real-time, and user-controllable characteristics, seamlessly blending the consideration of commonsense knowledge with autom
t natural next direction is to investigate LLMs as multi-robot task planners. However, long-horizon, heterogeneous multi-robot p
explored. In this work, we introduce the Object-Centric Instruction Augmentation (OCI) framework to augment highly semantic a
es a versatile learning environment with the potential to yield cutting-edge research in various fields of artificial intelligence and
h Fast and Slow Thinking (RFST)}, a framework that mimics human cognitive architecture to classify tasks and makes decisions
mplicitly infer task information, constraints, and API usage from examples alone. Meanwhile, research outside robotics has suc
t training and testing of the corresponding subtask policies, while simultaneously providing guarantees on the overall behavior t
mic obstacles not present on the pre-built map, ``when to replan'' the reference path is critical for the success of safe and efficie
marks. The former approach struggles to distinguish visually similar but different objects, while the latter is both time-consuming
or child. By autonomously reconstructing a complete model of the object and inserting it into a physical simulator (i.e., the robo
s and environments. We propose Generation to Simulation (Gen2Sim), a method for scaling up robot skill learning in simulation
d plans being invalid or even leading to dangerous actions. In this paper, we investigate the faithfulness of LLMs for complex lon
modal foundational models, offering multi-modal visual and textual understanding. In this paper, we harness these multimodal fo
or factor the inference process by decomposing language into task sub-goals and relying on a motion planner to execute each s
ective policy that adapts pre-trained skills in out-of-distribution states. All components of ASC rely only on onboard visual and p
M is particularly challenging since knowledge about future data utilization is absent. Drawing inspiration from human forgetting th
collaboration between humans and robots through Mixed Reality (MR). This includes a novel bi-directional spatial awareness a
ent into the compilation process, THEMIS enables safe and regulation-compliant UAV flights by enforcing prohibited and restric
high nonlinearity, leading to a substantial increase in computational burden. This paper presents an efficient method to plan mot
models of sensors and quadrotor dynamics. This system is verified in extensive simulations in the scenario of a team of quadroto
ollowing. We propose a method to overcome this issue by introducing a dynamic step adaptation strategy, which provides addit
matic landing. A tethered lifting-wing multicopter has a structure and static stability similar to a kite. Inspired by kites, this paper
ations, not much work has been carried out to develop perception algorithm for fisheye camera. One of the main problems being
e position, but rather we propose to regress a neighbor density over distance. This allows for a more accurate distance estimati
nally, most of these methods struggle with efficient replanning and collision avoidance in dynamic settings. To overcome these
able to extract readings with a relative reading error of less than 2%.
clic graphs that combinatorially represent the underlying nonlinear dynamics) offer data-efficient RoA estimation without needin
Our approach is based on the alternating direction method of multipliers (ADMM) and leverages the structure of the MPC proble
used in the experiment. When oocytes move from bottom side of the channel to the tip side, they generate gaps between each
under a condition of interval excitation (IE) that is strictly weaker than PE. This paper puts forward a time-division multi-channel
al-time. The manipulated images are fed into autonomy, allowing the self-driving vehicle to react to such virtual events. The full
or modifying the trajectories of the existing agents. Our framework employs a reward function consisting of both risk and plausib
ne label generation tool. In addition, we account for the errors in the predicted control values, which can lead to a loss of tracki
by analyzing locations and speeds that commonly correspond to high-risk interactions within the dataset, and use it within train
e develop a framework that employs reinforcement learning from human feedback (RLHF) to enhance the realism of existing tra
any practical usage, it is as crucial to teach the robot the "don'ts": conveying explicit instructions about prohibited actions, asse
ut considering the interactions between agents or making differentiated use of the acquired sensing data. This approach can re
POP first employs SLL to help the model learn to reconstruct history representations, and then utilizes feature distillation as the
ufficient guidance for capturing future states of interacting agents, they frequently produce unrealistic trajectory overlaps. In this
he scarcity of rare case data that existing works struggle to deal with. Leveraging the controllability of this modeling approach, p
knowledge (behaviors) between agents. This paper presents a new knowledge representation framework and a transfer strateg
assignment phase. In the communication phase, agents communicate with their connected neighbors and exchange their loca
onal vision transformer model is used to realize the controller which is then trained to predict state and controls over multiple pr
ng sequential navigational decisions. Simultaneously, we perform generalized target object detection using a pre-trained Vision
sed work, the robot utilizes data about its surroundings from drone or satellite images. Alternatively, a volumetric representatio
nvironment, semantic maps are constructed by an object recognition algorithm. Then, we design a Parser module based on Bid
ween vertices (or hyperedges). Therefore, we propose DyHGDAT, a dynamic hypergraph dual attention network to capture the
scenarios, paired with high quality control commands collected with RL agent and question answer pairs generated by teacher
ions as they either focus solely on static, pairwise interactions, ignoring interactions of varied granularity, or fail to tackle agents
yzing uncertainties and system limits, sensor-specific outliers are detected and removed in the first filter. We introduce knowled
active emissions. Cameras operating in the visible spectrum represent a quasi-passive alternative, yet they see a substantial dr
uct a comprehensive study on two fundamental yet underexplored aspects of imitation-based planners: the essential features fo
such a map by human resources is time-consuming and error-prone. In this paper, we proposed a framework to automatically p
y prediction have developed in isolation, with the former based on simplified rasterized methods and the latter assuming full en
of maps with descriptors and complex optimization processes hinder system performance. To balance efficiency and accuracy,
tion). To this end, we present DiPA, an interactive predictor that addresses these challenging requirements. Previous interactiv
y enhancing the multipolicy-based pipelines from both behavior and motion planning aspects. Specifically, MARC realizes a criti
ajectory planning approach that overcomes this challenge. Specifically, the cost function recovery method based on maximum e
es. To improve wheeled mobility in those environments with vertically challenging terrain, we present two wheeled platforms wit
social situations, where no amount of tuning enables them to be both safe (people are too unpredictable) and efficient (the froz
ween the robot, the environment and the targets is analyzed, and a general measure named detectability is introduced to correla
well-performing pre-trained LVLMs. This enables a single system to cater to a variety of autonomous driving tasks encompassin
ion mechanism captures global context information and models the long-term temporal dependencies of waypoints for multiple
ion. This significantly drops the success rates and even induces collision. In this paper, we present a self-correcting visual navi
embedding, we ensure that the final generated scenario is fully conditioned on all available context in the existing scene. Our u
t transform 3D LiDAR point cloud into 2D depth map and upsample the LiDAR output by applying the image super-resolution m
l features from point clouds captured by LiDAR. CAD is designed based on the polar coordinate system and focuses on predic
ed architecture to predict human future trajectories in human-centric environments from input features including human position
 l applications. This work proposes an enhanced LfI approach for learning the optimal RL policy by leveraging multimodal huma
Specifically, we design a variational autoencoder (VAE) to learn skills from offline demonstrations. To mitigate the posterior colla
ntroduces a framework for real-time multi-directional traversability maps (MTraMap) generation tailored for unstructured environ
color-labeled plates and reflective balls. LiDAR presents a solution in terms of object arbitrariness, as it addresses the challeng
coupled approach that results in higher collision risks and lower navigation efficiency. Moreover, existing robot navigation techn
tasks. A Reinforcement Learning based policy was proposed for sub-goal selection on this topo-metric graph, which improves
olicies can be learned either offline or online and can then be sent back to the robot fleet. In our experiments we distribute 72 ro
er, the human-designed explicit interfaces between perception and motion forecasting typically pass only a subset of the seman
ntinual learning algorithm is proposed, which allows to train the model in a process-parallel manner without suffering from catas
ior studies. In this work, we present a focused study of the generalization capabilities of the pre-trained visual representations a
s paper, we develop an efficient framework (MT-ACT) for training universal agents capable of multi-task manipulation skills usin
he arrangement which best satisfies the user instruction is selected and recreated in the real world with pick-and-place. This en
nternet. However, reinforcement learning often requires significant human effort in the form of manual reward specification or e
been shown to improve generalization accuracy in many application domains. We first demonstrate significant performance ga
-specific assumptions to analyze the source code. However, these are ignorant of the fact that physical unit mismatches in robo
a take-off weight of 870g, and an auxiliary platform with initial pitch angle adjustment function. In order to explore the difference
flipper are introduced in detail. It can exhibit better adaptive deformation under the action of hydrodynamics without affecting th
steady turning maneuver based on the body roll angle. The relation between the velocity of the prototype and the turning radiu
his paper, we proposed a general kinematic model to illustrate the kinematics of all aforementioned swimming motions. The mo
n-monotonic dependence of the robot speed on mud water content. Robot speed was the largest on mud with intermediate leve
efits may arise have been proposed, it is still largely unknown which mechanisms fish employ to obtain hydrodynamic bene
parately on the head and tail parts of a double-joint robotic fish and presented a method for accurately and timely estimating th
ing for untethered swimming and online tunable stiffness, and conduct a series of tests. We first apply interchangeable caudal p
h-inspired self-adaptive approach is proposed for the collective escape of swarm robots. A bio-inspired neural network (BINN) is
mmands, while at the same time override these commands to avoid collisions with the environment. We investigate several ope
bustness in the presence of model uncertainties. We moreover show an example application where the concept of form closed
in the yz and xy planes such that the robot returns to an upright position by rolling around the longitudinal (x) axis, and then twe
offline-designed locomotion controllers cannot perfectly model the true dynamics of the system, such as the motor dynamics. In
etah’s tail might be enhanced by aerodynamic effects. In this paper, we introduce AeroDima: a follow-up to the original Dim
AY is able to trot and bound dynamically using its spine. Simulation results reveal the positive roles of spinal movement, such a
s biomimetic features. At the robot's ankle joint, we imitate the articulation of a radius-ulna joint to enhance the wheeled motion'
weight of 300 mg, size of 30 mm × 15 mm × 5 mm), composed of piezoelectric thin film, shape memory alloy, magnet-lock
is equipped with an IMU and two triple-axis magnetometers. Utilizing magnetic field strength as an analog for depth, a novel de
tery in short duration without stability control. By incorporating a high-efficiency direct-drive piezoelectric actuator, and an optim
leg to achieve power amplification. Moreover, we carry out a two-step optimization of the mechanism parameters to achieve hig
sensing, computation, and power, enabling autonomy. The microrobot was designed with a super-elastic shape memory alloy (
ed by these interfacial fliers. The robot is comprised of a flapping-wing vehicle that generates a thrust force parallel to the water
omposed of a carbon-fiber beam, which functions as a leaf spring, and shape-memory alloy (SMA) wires, which deform cyclical
ction from visual percepts during a learning flight in the vicinity of the nest. After learning, the robot will travel away from the nes
nspired flapping wing robot (DFWR) with inclinable stroke plane control degrees. The proposed platform employs a four-wing c
e design and development of an optimized ion-propelled flying microrobot that excels in low weight, high thrust-to-weight ratio, a
ys (MEAs). Therefore, here we propose a modular BNN (mBNN)-based neuro-robotic system via local chemical stimulation and
bility of woodpeckers to extend and bend their long tongues to catch insects in tree holes. The proposed mechanism can chang
the turtle's legs when it encounters obstacles, to enhance the traversal performance. In detail, we first describe the hardware d
nges for navigation in complex, rheologically diverse environments. Inspired by a notable control simplification via mechanical in
ms of inducing Moment of Inertia (MoI). In literature, only one type of inertial appendages has been used on terrestrial robots in
and decentralized development process of the human body from just 20 types of amino acids. In our previous work, we propose
DRMs continue to suffer from their limitations. Recently, a nematode-inspired cable routing method was proposed; however, stif
a deep deterministic policy gradient (DDPG) learning-based controller that requires full-state feedback of the SCP-driven robot
ccuracies. In this work, we introduce a new approach to hand-eye calibration called EasyHeC, which is markerless, white-box, a
eed additional training. We propose a novel LiDAR-camera calibration method by using the Segment Anything Model(SAM) with
graphy, which serves as the cornerstone for the development of our novel stereo camera online self-calibration algorithm, for ca
ws the calibration parameters drift in time. OCAMO is based on an adaptive online stochastic optimization with a memory of pas
s introduces additional complexities. Existing methods for extrinsic calibration either necessitate specific tools or highly depend
loud. But many of these approaches also face a significant challenge, which is their sensitivity to lidar resolution and Field of Vi
wed from a large initial error to a small error range in little time. The second stage can further calibrate the extrinsic parameters
work to use semantic information and geometric features jointly in calibration to promote accuracy and robustness. In the first s
ased cameras, though it requires extra work. By using frame reconstruction from events, a printed pattern can be detected. A bl
valuation data differs from what the network has been trained on, the resulting estimates may not meet expectations. Moreover
ue gearbox, the harmonic drive. To tackle the engineering problem, this work proposes a physics-informed neural network (PIN
an approach that exploits the structure of the reward function by modeling the reward for each observation separately and using
ages for calibration. However, these features do not sufficiently interact during feature matching, which limits the calibration ac
depend on the geometry of the end-effector and the sensor attachment. We propose Geometry-Based End-Effector Calibration
oints. In our method, a robot arm carries an unknown object as it moves through multiple discrete configurations. Measuremen
ual intervention, which has proven to be cumbersome and costly. Learning-based online calibration methods have been propos
eless, due to the distinctive representation of event cameras compared to traditional frame-based cameras, calibrating them wit
sensor. Thus, it is not surprising that little related work can be found. (2) Long baseline and large viewpoint-difference. Current
hile maintaining commanded travel speed and direction. Our approach first trains a controller in simulation using a heightmap ex
reinforcement learning (RL) based local navigation policy to guide our legged robot. Unlike previous works relying on heuristics
el environments. VLFM builds occupancy maps from depth observations to identify frontiers, and leverages RGB observations
we describe how we can train a single unified diffusion policy to handle both goal-directed navigation and goal-agnostic explorat
quately explored. Drawing inspiration from cooperative multi-agent reinforcement learning, we introduce novel network structure
o allow classification inference to be done exactly with a closed form expression. Gaussian processes provide a distribution as t
ns during planning. In this work, we introduce a joint prediction and planning framework, Partitioned Scenario Replay (PSR), tha
cts multiple plausible future positions. By repeating the MMP for each time offset in the future, multi-time-step multimodal motio
optimization problems of collision avoidance can be decomposed into several low-dimensional quadratic programmings (QPs)
high-dimensional raw observations demands massive neural networks and extended training periods. Furthermore, there are d
ve collision avoidance. However, the approaches directly outputting control inputs through neural networks have drawbacks in i
. Our approach provides an asymptotic, 2-approximation guarantee for MGPF. We also present extensive numerical results to i
ng the hazardous areas on uneven terrain. In particular, certain nodes of the tree are consistently kept to form a sparse graph
mazes, as well as in real-world deployment scenarios on Boston Dynamic's Spot and Unitree's Go1 robots. DiPPeR performs o
n be leveraged for further optimization. To find such domains, practitioners have harnessed a methodology known as Sum-Of-S
g the long-term execution. An alternative method is to train a DRL policy that outputs the navigation path directly. Then the robo
chieved by coarsely discretizing time, and by representing multimodal multi-agent interactions as distributions with infinite supp
ct, rigorous certificates of non-collision; it can never falsely claim that a motion plan containing collisions is collision-free. We de
and the side information. Furthermore, a multi-level cross view attention module is designed to capture interrelated details betw
nes by incrementally outputting subgoals given the current planning context. Then, we take into account the temporal informatio
nt. The differentiability of our model allows its usage within the MPC solver. It is computationally hard to solve problems with a v
rd invariance (i.e., keeps the robot away from obstacles) when combined with the probability formulation. Notably, our formulati
e demonstrate that MIMs' can help accurately differentiate obstacles that are safe to navigate through (e.g. beaded/string curta
ated representation of the robot's immediate environment. This map captures crucial environmental variations, including height,
oceptive sensors by actively colliding with its surroundings for more reliable localization. Firstly, we design a collision-based obs
raph size, limiting its applicability to 3D geometries. To overcome these challenges, we propose GDN-R, a novel 3D graph-base
propose an efficient LOS checker for maps of arbitrary dimensions with sparse obstacles. Our approach involves a two-step pro
nty-informed. For that reason, we propose conformal predictive safety filters that: 1) predict the other agents’ trajectories, 2
e high-dimensional multi-modal sensory inputs to low-dimensional feature vectors using encoder networks and represent them
ct adjacent nodes. Firstly, an analysis of angle constraints of the HRM is conducted, providing restrictions on node selection du
an unknown environment which leads to dealing with the sampled data set size, memory, and computational complexity. We ad
ial, fragility) of common objects, which restricts their usefulness for robotic manipulation tasks that involve interaction and physi
h to the next best view only. To address these limitations, we propose a non-iterative pipeline based on the Prediction of the Re
2) requiring a considerable number of less-gain views in terms of surface coverage. To address these limitations, we propose t
ely on LiDAR data however, and do not fully exploit image features. Besides, they introduce significant latency. To overcome th
d. To tackle this issue, we propose a self-supervised interaction perception method, referred to as SM3, which leverages multi-
, range images contain rich semantic guidance. Based on this intuition, we propose MF-MOS, a novel motion-focused model w
, which is usually inevitable, can lead to suboptimal reconstruction quality, particularly in some geometrically complex regions.
and masks for various real-world conditions. Instead of directly using the RGB images, we propose to use the estimates to train
protocols. In contrast to previous implicit cloth representations, our reconstruction mesh explicitly describes the positions and v
d occluded regions of the scene. To facilitate research on this new task, we extend the AmodalSynthDrive dataset to include pix

hreefold: (1) developing a novel progressive framework which integrates the cost volume module and scene flow estimator, en
al circum-views from the ego to excavate three-dimensional features of the surrounding environment. Circum-views provide a n
serve that the spatiotemporally continuous traces of events provide a natural search direction for seeking pixel correspondence
Transformer-based architecture. Our key idea is to enable a single robot to intelligently discern the relevance of the collaborator
ng stability. In this work, we present R2DM, a novel generative model for LiDAR data that can generate diverse and high-fidelity
e present a new real-time multi-task network adept at three vital autonomous driving tasks: monocular 3D object detection, sem
synergy to boost the performance of LiDAR perception tasks across multiple large-scale datasets and benchmarks. Our novel t
aptive grippers. The contributions arise from the integration of compact deflection sensors, that are particularly fabricated for lar
e incipient slip during the pivoting process, which introduces measurement errors and makes it hard to determine the boundary
tile sensing module, which incorporates inertial, magnetic, and pressure sensors. The module's flexibility enables us to estimate
mework for estimating contact models online based on torque and position measurements. To do this, compliant contact model
eformation and being sensitive to edge and surface following, improves the structure in that the markers are distributed within t
ynamics between a rigid mesh contacting a deformable mesh under external forces. Our approach represents both the soft bod
act, is necessary for nonholonomic robots to achieve stable pushing. We also show that the stable pushing constraint and the n
olling contact on an arbitrary smooth hand. Our formulation splits the problem into four subproblems: deriving the second-order
many factors. In this work, we conducted sliding experiments with robot fingers and a hand, and show that rubber friction varies
orizon tasks, thanks to their powerful state and action abstractions. But they assume predefined skill sets, which limits their real
environments. However, our investigation reveals that existing methods are vulnerable to dynamic distributional changes and of
ming (observes each training example only once), requires a single pass over the data, can learn in a class-incremental manner
new skills. However, existing replay-based methods store poor representative experience, and the experience utilization of old t
mise in handling environmental changes but lack efficiency in accommodation speed. To address this limitation, we propose an
we introduce LRLL, an LLM-based lifelong learning agent that continuously grows the robot skill library to tackle manipulation t
time-efficient manner for rigid object manipulation. Our system can re-use newly acquired skills for future tasks, demonstrating
ramework called the Probabilistic Spiking Neural Network with Variational Continual Learning (PSNN-VCL). In this framework, P
process using an open-vocabulary vision model, which extracts skills as recurring patterns presented in unstructured demonstra
ular bi-copters with an undocking mechanism. In combined configuration, a crossover docking approach is employed to compe
. In the face of the massively redundant actions, we propose to reshape the action space by considering the spatially local actio
cy. However, identifying an optimal module composition for a specific task remains an open problem, presenting a substantial h
hods are limited by their capability to move-over window frames and need more flexibility to access tight corners and curved su
o predict object motion and distance to obstacles. This distance is incorporated in a control barrier function for the resolved mo
onmental information is modeled using a 3D Scene Graph, a hierarchical map representation that describes both geometric and
propose a novel system named ColAG, to solve the problem of autonomous navigation for a group of blind UGVs by introducin
to charaterize ``robot-robot'' and ``robot-environment'' interactions. Specialized performance-related energies, e.g., motion smo
terventions. This has fostered the evolution of robotic systems designed to prevent such tremors. These robots are enhanced b
e, we introduce PanNote, an automatic annotation tool for people's positions in panoramic videos. Our tool is designed to be co
er unpreventable failures caused by the human participant. To address this deficit, we present the multimodal Handover Failure
ble datasets, our dataset provides both images with objects small and completely visible in the image, and images where objec
ce PointSSC, the first cooperative vehicle-infrastructure point cloud benchmark for semantic scene completion. These scenes e
both RGB and physically realistic depth images, surpassing previous dataset generation tools. We create an RGBD dataset tai
an-object interactions (HOI) and overlook object-object interactions (OOIs) despite the necessity of OOIs in effectively represe
ctories. With this objective in mind, we proposed the Robotic Total Stations Ground Truthing dataset (RTS-GT) dataset to supp
h data is generated by imitating the imaging principle of real-world sensors. A range-aware rendering strategy is further introduc
ailed, including organization, acquisition, and annotation methods. It comprises both real-world and synthetic data, collected with
ing conditional models is hard given a lack of paired human-robot interaction datasets. Can we instead leverage large-scale hu
e driving situations and to make safe and efficient decisions. To that end, we introduce a new large-scale dataset that consists o
peline, we designed a label correction web application that facilitates human verification of automated pedestrian tracking outco
uman, human with grasping tool). With this data, we show that models trained on all embodiments perform better than ones trai
enge. In this work, we introduce a novel approach based on Heterogeneous Graph Neural Networks (HGNNs) to transform regu
wever, due to limitations in the training dataset, the current focus of the community has mainly been on simple cases, such as p
avior, such that robots can navigate among humans in ways that don't disturb human behavior in visual navigation using only on
sk identification algorithms uses independent datasets, leading to difficulty in direct comparison and hindering collective progres
nd a stereo camera to record a novel obstacle detection data-set for inland waters. We annotated 1974 stereo images and lida
and sparse traffic scenarios, fall short of capturing the complexity of driving in such environments. To fill this gap, we present ID
a 3D point cloud are affected by the distribution of the points. The lack of a 3D domain adaptation benchmark leads to the comm
ollection. To overcome this difficulty, we propose a benchmark dataset, ROV6D, for 6D pose estimation of remotely operated v
 oor and Offroad Dataset (GOOSE), a comprehensive dataset specifically designed for unstructured outdoor environments. The
d discriminate target objects and outliers. This paper proposes a novel dataset generation pipeline for suction-grasping in bin-p
d from 13 distinct crowded scenes in New York City taken under varying lighting conditions with appearance changes. Each sce
cs dataset for precision agriculture and forestry addressing the counting and mapping of trees in forestry and orchards. TreeSc
letter, we propose a novel learning-based framework to remove the virtual points from LS3DPC. We first project 3D point cloud
se a novel source-free domain adaptation method for point cloud primitive segmentation, which follows the popular pseudo-lab
atures from cross-modality point clouds and achieves point selection by feature matching. We also introduce a cross-modality o
curately estimate joint parameters and states for various articulated objects with high precision and robustness. The paper also
lier detection task and use an energy-based framework to detect outliers in point clouds. More specifically, our method learns t
nment loss combined with LocalPCA compression to learn compact 32-dimensional descriptors. To enable efficient storage or c
odeling capabilities and quadratic computational complexity. To address these limitations, we present the Tree-based Transfor
eam robotics tasks, especially given the increasing prevalence of headset devices. This study expands EgoPAT3D, the sole da
ion of CTTA in the real world. Recently, existing CTTA methods mainly focus on utilizing a majority of parameters to fit target d
p judgements and the ignorance of interframe linkage can also lead to slip detection failure. In this paper, a new spatio-tempora
, the distance between the right hand and an object should also be predicted to be small,'' explicit to the model. Our model uses
elds to allow a robot to self-model its kinematics as a neural-implicit query model learned only from 2D images annotated with c
orld environments. In this paper, we propose a new method for language-conditioned affordance-pose joint learning in 3D point
nces plane segmentation by generating subplanes in Deep Plane Clustering (DPC) module, which are then merged with the fin
ior methods require accurate camera calibration and robot kinematic models in order to label training images for the keypoint lo
ere UAV-based domain require distinctive solutions to image synthesis for data aug- mentation. In this work, we leverage recen
able for mobile robotics applications. This work proposes an adaptive multi-pair contrastive learning strategy for visual represen
sity and color variations in images – existing techniques such as Background-oriented schlieren (BOS) allow detecting them o
which limits the complexity of the onboard algorithms. A traditional solution to overcome these limitations is represented by lightw
end-to-end framework based on the mean teacher paradigm, named PCTeacher. Our PCTeacher generates both point-level an
noisy input, our system is able to output scene geometry, photo-realistic renders and 3D consistent panoptic representations w
two novel components: gradient-aware mask and semantic constraints. The gradient-aware mask enables adaptive and robust
These approaches often falter in terms of performance or efficiency and grapple with the cross-camera gap problem. This pap
solve this problem, we present LiSV-3DLane, a large-scale 3D lane dataset that comprises 20k frames of surround-view LiDAR
l efficient Transformer-based network to predict the future LiDAR point clouds exploiting the past point cloud sequences. We al
oposed to decompose mobile manipulation tasks into a simplified motion generator for the end-effector in task space and a trai
ipeline is entirely self-supervised, thus enabling lifelong learning for real robots. The network is composed of a 3D back-end tha
ten occluded or lie outside the field of view of the camera, rendering the camera observation ambiguous with regard to their loc
n predicting trajectories conditioned on each goal. However, a single 2D goal location serves as a weak inductive bias for predic
such as joint states and current measurements. However, insufficient attention has been given to friction hysteresis approxima
we take a preliminary step in this direction through merging, or averaging, subsets of Decision Transformers in parameter space
ork with a novel asymptotically stable dynamical system integrated into a Transformer-based model. We use an asymptotically
ior and state of unobserved agents. To address this issue, we propose a Partially Observable Altruistic Q-learning (POAQL) me
mework in which large language models are prompted to maintain an estimate of the world state, which are often unobservable
nce MGTR’s capabilities, we leverage LiDAR point cloud data by incorporating LiDAR semantic features from an off-the-she
ation studies and human experiments, we investigate which components are crucial for quick and reliable texture recognition. A
or complex robotic control, can be achieved by using chemical synapses, within a sparse, bio-inspired architecture, called Neur
 trade-off between speed and accuracy. As a result, we obtained a model that maintains real-time performance while delivering
s SLAM and SFM require global optimization, which means that time and space resources increase dramatically with the growth
m the YCB Model Set. These simulated images are then translated into the real domain using the Diffusion Model and automa
p a transformer-based architecture called ECHO, which operates in the aforementioned shared space to predict the future moti
le portion of the data loses its semantic information during transmission. Consequently, how can we ensure both the perceptua
candidate solution trajectories on the GPU. Precise, valid solutions are then found through classical approaches such as differ
stness challenges. We propose SafeDPA, a novel RL and control framework that simultaneously tackles the problems of policy
tinuum robot by a physics-informed neural network (PINN). The underlying physics is modelled with the Cosserat rod theory an
ions have become alternatives to classical approaches, but they still lack comprehensive handling of complex constraints, such
mework for multi-agent object rearrangement planning, addressing the challenges of task sequencing and path planning in com
ontrol environments, encompassing multiple action constraint types. Our evaluation provides the first in-depth perspective of the
e learning algorithm with an effective exploration strategy can significantly improve its adaptation performance to unseen enviro
hat combines semi-supervised learning and reinforcement learning for online grasp learning. By treating pixels with reward feed
work, we present an end- to-end pose estimation approach that achieves camera-to-robot calibration using monocular images
tructed through this manifold, the input images from successful task executions will remain largely unchanged, while the image
risky behaviors. To overcome these generalization limitations, we contribute a novel system for task-oriented grasping called t
t motion planning, and robot contact optimization. We present the Part Assembly Sequence Transformer (PAST) -- a sequence
that can be attached to hard-to-use objects, thus improving "robot ergonomics". Our learning-based framework formulates the
llenges. Addressing this, we employ a general framework in visuomotor policy learning that utilizes visual pretraining models as
hat involve the transportation of heavy objects via ropes, e.g., on factory floors, at disaster sites, and in forestry. To address this
ur proposed method involves utilizing a technique known as knowledge distillation, in which a ``teacher'' policy pre-trained with m
n of symmetric policy learning to visual-force problems. We present Symmetric Visual Force Learning (SVFL), a novel method f
point clouds and an object discovery and tracking engine. Our approaches can perform multiple challenging tasks including rea
tion planning. In this paper, we propose an autonomous, effective and general approach for picking up tangled-prone objects fo
ng alternative is to learn fabric manipulation directly from watching humans perform the task. In this work, we explore how dem
p that considers the complete point cloud of the workspace. It integrates grasp pose estimation, end-effector pose evaluation, a
ce pre-grasping redundancy. This framework incorporates two distinct Deep Q-learning critics: Critic I selects the best course o
he goal shape at that specific moment, both of which are impractical in various robotic applications. In this paper, we solve this
stacle than others, but existing work sheds little light on precisely how much each factor contributes to the generalization gap. T
novel framework to reconstruct the scene on the production line depending only on RGB input, based on multiview stereo. Com
and more general learning from demonstration. Our approach utilizes Track-Any-Point (TAP) models to isolate the relevant mot
pose. Concretely, we employ hierarchical reinforcement learning to combine a sequence of learned parameterized manipulation
m with an RGB-D camera in eye-in-hand configuration, and an elongated, volumetric, elastoplastic object. We find similar objec
h suitable grasp poses to facilitate trajectory planning. Although these approaches have succeeded in certain types of articulate
can only predict limited grasp orientations which leads to a reduced number of feasible grasps in bin picking with restricted reac
ed as a contribution to learning. However, HER uniformly chooses failed trajectories, without taking into account which ones m
eo network for the purpose of transparent object reconstruction, enabling material-agnostic object grasping in cluttered environ
ge,and the correction stage determines a correction to the tentative action based on the cropped image according to the tentat
en suffer from issues like sparse sampling and noisy output due to the limits of the emission-reception imaging principle. On th
learn to manipulate articulated objects without demonstrations. We combine the strengths of 2D segmentation and 3D RL to im
his, we introduce a proxy hardware model, which is always differentiable and enables efficient co-optimization alongside a long-
nd their embedding in an iterative online/offline scheme ("collect and infer") can drastically improve data-efficiency by using all t
ent the agent's objective and accelerate the affordance discovery process. We provide a theoretical justification of our approac
on one type of gripper to another to work around the tradeoff is inefficient and not scalable, as it would require tremendous effor
6-Dof grasp generation. Therefore, we propose an effective local grasp generator combined with grasp heatmaps as guidance
w-level controller, where the hyper net learns to generate the parameters of the low-level controller and the controller uses the
ntation tools. Our method consists of an off-policy, maximum-entropy, Actor-Critic algorithm that learns a policy from a binary r
In this work, we present ATPPNet, a novel architecture that predicts future point cloud sequences given a sequence of previous
ing both global and local information from the whole image. In this paper, we propose a novel Semi-supervised Learning (SSL)
esigned for well-illuminated crack images and hence their performance drops dramatically in low-light scenes. In addition, conv
mputational resource requirements of neural network models for point cloud processing tasks, compared to full-precision point c
int cloud super-resolution approach for 3D mmWave radar data, named Radar-diffusion. Our approach employs the diffusion m
ages by nouns and modifiers and introduce a new embedded binding space where the linguistic and visual cues can effectively
n approach involving precise calibration and expert supervision. In this paper we introduce Cycle-Correspondence Loss (CCL) f
eural implicit encoding. Our proposed system successfully achieves the goal through improved multi-MLP decoders (MLM) and
ribution shift between sim and real. In this work, we introduce Composable Object Volume NeRF (COV-NeRF), an object-comp
ual Incentives (TAVI), a new framework that enhances tactile-based dexterity by optimizing dexterous policies using vision-base
eal-time performances, and the infinite degrees of freedom of the deformable linear objects. In this paper, we propose a real-tim
s high resilience in handling complicated scenes, reaching an average of 98% task completion in simulation and real-world scen
eature deviations in tactile space. Compared with the manually designed PID-based controller, the proposed controller depends
t each time step, predicting the optimal push configuration to isolate the target object, determining the appropriate grasp config
o the required hand actuator controls. However, acquiring such inverse Jacobians without complex hand-object system models
model for robotic manipulation using both human manipulation visual and tactile data. We develop a system for collecting visual
trate that it is possible to infer the stability of object placement based on tactile readings during contact formation between the o
on. This paper investigates screwdriving by a robotic arm/hand pair, dealing with the chain of contacts connecting the substrate
observation signals and high noise levels. We have successfully applied this method to a biomedical task of a simulated crania
ehavior tree-based primitive switching mechanism that leverages high-frequency tactile data for the estimation of contact state.
terous dynamic manipulation policies for deformable linear objects with a fixed-base dexterous hand in an end-to-end way. By
ificing compactness. A robotic hand that uses two Everyday fingers demonstrated an 80% success rate in picking up and placin
s and task-space reachability constraints. As a result, unacceptable errors may occur, and some tasks cannot be accomplished
is widely adopted in computer vision and simulation, its usage in robotic deformation control is still an open topic. We develop
RL addresses (i) and (ii) by sampling and averaging contact modes. In contrast, model-based methods smooth contact dynam
s based on Monte-Carlo Tree Search and has three levels: 1) planning object motions and environment contact modes; 2) plan
om 5 mm to 25 mm. Our integrated strategy incorporates the robot arm, gripper, and sensor to manipulate and orient small obj
rative decision-making for autonomous navigation in complex environments. However, the navigation problem of the R-UAV ha
S offloads conventional on-board state estimators and feedback controllers to Cloud and Edge compute hardware without modi
d use tether angle and strain, in conjunction with the physical constraints of the tethers, to adjust their position with respect to th
agrangian layer for local constraints and DDP as the local optimizer. The second one, Merged Distributed DDP (MD-DDP), is a
e resources. Since each robot can choose among many tasks, it remains a challenging problem to coordinate jointly these robo
es while guaranteeing safety. This approach leverages the active sensing capabilities of multi-robot systems to supplement GIS
e o(1) with respect to the network size. We establish two new theoretical results for LC-ADMM: (1) exponential convergence w
mpared under pre-allocation and dynamic allocation scenarios. The synchronous and asynchronous communication manners ar
 gile and efficient movements. In detail, we first design an eight vector-distributed thruster layout for fully vectored propulsion of
olve the attitude control problem of a morphing quadrotor. Prior to controller design, a custom-built morphing quadrotor is introd
sidestepping the need to ex- plicitly explore the entire state-space. However, the performance of this sampling-based approach
ertainty in STL. To this end, we formulate two relevant control objectives to enhance the temporal robustness of the synthesized
etworks (DeLaN) [1], it is possible to control and extract interpretable physical information of a robot. This paper introduces an a
strain modes. This choice imposes a predictor-corrector structure on the approach, and requires computing both the residual ve
constraints and the kinematic tree depth respectively. We provide an expository derivation for the original PV solver and extend
rmance, especially for real-time applications. Reduced interface models (RIMs) can be used to represent the dynamic behaviou
traversability. To address these challenges, we propose a novel approach called Reinforcement-Based Inferred Dynamics via
simulation-learned policies to real-world robots. In this paper, we propose a novel framework for robustly learning manipulation
posed PiH strategy is completely passive: After the peg is trapped in the hole during the hole-searching process using the sphe
ntify the objects within the image pairs. Subsequently, the object image patch of interest is extracted and fed into an encoder-d
s physically stable with a limited number of parts being held and a support surface. We apply efficient tree search algorithms to
o the specific characteristics of a given intralogistic problem, i.e., the plant layout, fleet size, statistical description of tasks, dispa
metric features (e.g., points, lines, and planes), which tends to fail in environments with ambiguous geometric features; or depen
n scenarios. In this paper, we propose a first method to deal with adversarial patrolling using a data driven approach. We cast t
rst involves coordinating all robots to follow a shared patrolling strategy covering the entire graph, while the second approach p
obal system state information, but it can be computationally expensive. In this work, we propose a novel framework called Long
near quadratic regulator (LQR). To ensure hover capability, we introduce a position controller based on a cascaded-PID. The n
. The classification is based on a desired post-impact motion created via an idealized rigid robot-object-environment model. To
nusual events. We introduce a dataset that allows training and benchmarking of anomaly detection methods for robotic applicat
nipulation. Therefore, execution of a manipulation action may result in an undesirable outcome involving accidents or damages
ort polygon. This pipeline allows the robot to dynamically and autonomously adapt to both partial and complete motor failures. U
a machine learning model that by observing one trajectory infers the corrupted printer dynamics to improve its accuracy. Our m
y defined for quadrotors, or indeed, many other autonomous systems. Moreover, any existing definitions are often incomplete a
such semantic data to be available or develop bespoke methods for capturing it. Instead, this work introduces a framework tha
e Models (LMMs). Our approach uses LLMs to update the required tasks in hierarchical task networks (HTNs) to handle conting
o generate Segmented Exploration Regions (SERs) and generate frontiers from a global-scale perspective. In particular, this pa
withstand external forces such as wind disturbances. Once engaged and applying the desired gripping force, the perching mec
he generative property enables perception in the small by providing high-resolution reconstruction capability. These properties a
odes of driving and traversing. By using a generalized coordinate description, the configuration space dimension is reduced, wh
lizability-aware constrained ICP optimization module and couples both in a unified manner. The proposed localizability detectio
wn promise, but demonstrated intervention tasks are limited to controlled environments or docked. To address these limitations
Markov Random Field (GMRF). This allows to improve source localization hypotheses by building on the GMRF's concentratio
s, a model is trained using Gaussian process regression (GPR) to achieve radius prediction. The feasibility and effectiveness o
pection viewpoint determination and do not consider dynamic obstacles. To maximally increase the level of autonomy, this pape
face towards sun, but these also come with their own drawbacks such as increased cost and weight. This paper presents SOL
eep learning to extract ball joint components. In order to generate rich training datasets, we develop a novel data augmentation
dered in the evaluation of energy efficiency. Lately, the paradigm in industrial robotics shifted towards lightweight robot manipul
oduce a mobile hub concept. This hub facilitates continuous mission cycling for unmodified off-the-shelf UAVs. Our approach ce
al requirement on map granularity, we propose a low-to-high resolution path planner that first guides a single robots to quickly a
een these cycles, Autonomous Surface Vehicles (ASVs) can be deployed to maintain an up-to-date model of the phenomenon.
d optimizes the transformation between the scanner origins and the robot pose by minimizing 3D point omnivariances within the
ldfires or volcanic eruptions. To solve this problem, we present an autonomous measurement system involving an unmanned a
sequence of RGB and depth images fused with pose estimations. Our method, termed WayFASTER (Waypoints-Free Autonom
es of objects during contact, which can be known as the see-through-skin mechanism. In the meantime, the biomimetic tips em
based on multi-walled carbon nanotubes (MWCNTs) and polyurethane (PU) sponge is fabricated for pressure sensing. The se
, within recent literature on tactile texture classification there is an emphasis on accuracy at the expense of real-time processing
y be particularly important for tasks such as object insertion. However, tactile torque estimation has received relatively little atte
shing using tactile sensing without visual input, namely tactile pushing. We present a goal-conditioned formulation that allows b
eration framework benefits from a learning-based contact detection algorithm which triggers switching from unilateral teleoperat
OUCH sensor comprises: (i) an elastomer structure resembling the classic GelSight sensor design for fine-grained contact geom
underactuated, linkage-driven robotic finger with low-cost, high-resolution vision-based tactile sensing and proprioceptive sensi
ressable RGB LEDs illuminated in a novel rainbow spectrum pattern. In addition to being able to scale the illumination scheme
ignificant progress has been made in imbuing robots with this tactile intelligence, challenges persist in effectively utilizing tactile
nd 3D structure designed for robotic in-hand manipulation tasks. AllSight is mostly 3D printed including a novel and simplified fa
onceptually, 9DTact is designed to be highly compact, robust, and adaptable to various robotic platforms. Moreover, it is low-co
arious complex surfaces. The GelFinger sensor is equipped with an embedded miniature motor that allows for the adaptation o
Tac, a novel vision-based sensor that combines tactile sensing with 3D vision. The proposed sensor relies on stereoscopic visio
d quickly, are affordable, and occupy a very small footprint. Our prototype sensor uses only a sparse array (8-9 mm spacing) o
primarily using the contact area spread rate (CASR) over their fingertips. CASR can be defined as the integral of tactile flow, wh
2.0 setup that integrates two affordable industrial-level robot arms with low-cost high-resolution tactile sensors (TacTips). We p
od driven by reinforcement learning for object reconstruction at scales that automatically explores the object surfaces in a limited
rceiving the intrinsic properties of various terrains. The intuitive approach entails directly analyzing data recorded by the robot's
at can track previously unseen objects held in hand. From consecutive tactile signals, TEG-Track optimizes object velocities fro
rectangular textiles by means of edge tracing. The system achieves >85% coverage on multiple textiles of different sizes and t
nse contact forces from the bubble sensor's internal RGBD camera and air pressure sensor. We present a finite element mode
delicately handling fragile objects, the absence of sensors has hindered their full potential to execute precise and secure grasp
n on temperature, pressure, and shear. SATac utilizes the thermoluminescence of strontium aluminate to sense a wide range o
ensing problem as an estimation problem of resistive sensory arrays. Utilizing a Regularized Least Squares objective functionâ€
monitoring capacitive changes due to skin deformation. CushSense is cost-effective and easy to fabricate. We detail the sensor
ces while 3D round sensors are essential for dexterous manipulation. In this paper, we propose a bi-directional Generative Adv
posing challenges in their application on curved surfaces and requiring the maintenance of stable lighting conditions throughou
generating marker motions caused by elastomer deformation. This disallows access to tactile information deducible from marke
te pose estimator that tracks the most likely sequence of object poses in a coarsely discretized grid, and a continuous pose est
nts and discontinuities in traditional shape reconstruction models. Three-dimensional experiments were conducted within eight
r at clock-driven intervals, we observe a decrease in the final distance to the target location. Furthermore, we combine the prim
ce Network (QTIN) model is introduced to reduce crosstalk, which has specific compatibility with the impedance-separating me
e propose a dynamic manipulation strategy: tracing the garment edge until the hidden corner is revealed. We introduce a mode
eless, much of the earlier work in merging these two sensory modalities has relied on supervised methods utilizing datasets lab
ent safely. The results showed that our system dramatically reduced the overall collision chance compared with several baselin
eedback input available in tight environment: an arbitrarily restricted motion space and a single feature point with unknown posi
duced by degenerate motions. However, those studies mostly stay in the specific sensor domain without aiding to extend the un
lanning and control framework using GTac, a hetero-Geneous Tactile sensor tailored for interaction with deformable objects be
ones, as well as holes through the side of a cuboid, is proposed to simplify the manufacturing of force sensors. To ensure the sa
-fast IFS method that uses multiple planes to establish surface geometry descriptions. The method can reduce high-order conta
Additionally, our estimator is a fully proprioceptive-based estimator, eliminating the need for any exteroceptive devices or observ
ue is employed to learn the mean and covariance of the residual dynamics. On this basis, a virtual measurement equation is es
tric tactile skill definition that directly links identified object constraints of the task to the definition of constraint-based unified forc
feedback to the user and suggest alternative modifications. We demonstrate our task modification process using a Hello Robo
hesize a winning strategy -- a reactive (robust) strategy that guarantees the satisfaction of the goals under all possible moves o
with variable update rates. We present an exact computational method based on polyhedral manipulations for the first class and
oreover, it is often nontrivial to define the robustness of complicated predicates precisely. To address these issues, we propose
ometric constraints effectively requires the incorporation of tools from differential geometry into the formulation of machine learn
n specification and their costs. For synthesis, we introduce a Mixed Integer Linear Programming (MILP) formulation that combi
er the durations of each edge's traversal, we want to find paths for all agents such that (a) the specification robustness is maxim
used in the cloud to compute the control commands for the present robot state. Such communication delays could potentially le
k specification language that allows us to capture explicit time requirements. Our objective is to find a strategy that maximizes th
-Transformer- guided Temporal Logic framework (T2TL) that exploits the structural feature of Transformer twice, i.e., first encod
ed as a deterministic, adversarial agent; as well as probabilistic synthesis, where the human/environment is modeled via a Mark
s goal, this work develops a real-time decision-making and motion planning framework. It allows the robot to follow a global tas
mula. The Task Batch Planning Decision Tree (TB-PDT) is then developed, which is a variant of conventional decision tree spec
ork on reinforcement learning with LTL specifications treats every new task independently, thus requiring large amounts of train
he film thickness distribution using impact angle and spray distance, which were previously modeled together. For higher accur
primitives and probabilistic verification, we construct a natural-language behavior abstractor that generates behaviors by synth
multiple pieces of evidence, and what they imply. A central idea in the approach is to maintain a separation of concerns so that

2K together with a Gaussian-Mask annotation method. Next, based on the TaTa gripper, we propose a grasping network name
e task, e.g. become stiffer or more compliant depending on specific requirements. Many robotic devices have been proposed in
stricted to a specific object category or a parametric shape. Our approach, however, can transfer grasps associated with implici
ce time and number of interactions. We propose an encoder/decoder action-feasibility model to efficiently adapt to new objects
roblem in three items: 1) graspable point segmentation; 2) hand grasp generation and 3) grasp optimization. Specifically, our m
To ease the transfer process, we propose Neural Surface Grasping Fields (NSGF), an effective neural representation defined
e. In this paper we propose two kinds of attacks on Grasp Quality Networks, one assuming physical access to the workspace (t
cks the ability for large-scale, object-centric manipulation by mapping object segments to one of the previously seen objects from
ion of features and the generation of redundant or missing relationships in complex scenes, such as multi-object stacking and p
ession and classification methods for estimating the normal force and its contact location. Our sensor can be affixed to any robo
But most of them are limited to simulations. The present paper investigates how automatically generated grasps can be exploite
ver, the abstract synergy representation complicates the association of relevant high-level grasp parameters, as for example th
a learned grasp-quality estimator, the uncertainty in the estimate and the distance from the seed proposal to promote temporal c
unpredictable object poses and geometries. We build on the foundational principles of the Fractal Vise to a broader class of grip
sed grasp, the robot needs to reason about the interactions with other objects in the scene. Finally, the robot must compute a c
g an object in a fixed location and controllable 1-D orientation. It also collects data and swaps between multiple objects enabling
 teaming, enabling the robots to operate safely and to continually improve their performances through deployment data. Noneth
better leverage the representation capabilities of pre-trained models for robotic grasping, we propose InterRep, a novel interacti
gnals significantly exacerbates the challenge of exploration. The applicability of such systems is thus typically restricted to simul
se selection. GPDF accurately models the shape of objects, which is essential for precise grasp planning. SE(3) equivariance e
hat generates accurate grasp quality scores for different multi-finger grasp poses using only hundreds to thousands of training
we propose a novel dexterous grasp generation scheme called textbf{GrainGrasp} that provides fine-grained contact guidance f
while effective for object categorization, face challenges with identifying precise grasp points on PCB surfaces. This paper studie
ble capacity to interact with them. Hand postures commonly demonstrate the influence of specific regions on objects that need t
fferent experiments (e.g. due to sensor noise, or variations in object placement). In order to address this issue, we classify the o

s effective for grasping in extremely confined spaces, such as the inside of a box in a shelf, to avoid interference between the g
ngers cannot be easily replaced to meet the requirements of objects with specific geometries. To overcome the complexity and
nnected by flexure hinges. The prototype was produced by means of a metamaterial approach, which utilizes 3D-printed infill str
gripper, tolerance to errors is provided. The key parameters such as gripper compliance and maximum allowed initial impact vel
high accelerations, as well as supply normal forces at least several times that required for static equilibrium. We also show tha
g angle and actuator stiffness. Our innovative approach utilizes a bionic finger design that incorporates a memory alloy skeleton
ation elements into the fingers, the gripper is equipped with the capability to open-close under an external magnetic field, facilita
like silicon die. The design is centered around a custom printed circuit board integrates functionality for vision-based error com
estructive actions are only included as predefined case-specific actions such as milling away a screw head. By contrast, this pa
geometry, governing kinematics, and quasi-statics mechanics of rigid and compliant versions of the hand are developed. Thes
. Its two-fingered design exploits a specialized structure to cope with constrained spaces (e.g., containers in narrow shelves). T
constraints encode regions of the search space containing reasonable co-design solutions. Furthermore, this underlying represe
stiffness and contact conditions. We find that stiffer fingers with soft foam pads enlarge the landing sufficiency region. Specifica
employs chopstick-like appendages for precise and dexterous manipulation. This gripper is capable of in-hand manipulation by
solving these challenges while satisfying the large-scale data collection demands of deep robot learning paradigms. Studies on
sed imitation learning achieves human-level motion speeds with environmental adaptability, only requiring human demonstratio
acked tightly, which limits access. A single suction cup is compact enough to reach and grasp the smallest items (down to 5 cm
iming actions. In this study, we propose a novel framework that combines ideas from task-parameterized Gaussian mixture mo
s, which is verified by four public datasets, Ev-Objects, Ev-Containers, Augment8000 and BioTac-Dos. Comparative studies sh
over other solutions. However, due to the non-anthropomorphic linkage, this hand exoskeleton also encounters difficulties in m
well as pressure stimuli related to grasping force. Ten able-bodied participants and one subject with limb loss validated the sys
th only a few sparse actuators deployed on the surface of the handheld device remains a significant challenge. In this work, we
hat the trainee can interact with thanks to the embedded physics engine. This virtual simulated environment is augmented by th
non-intuitive control. In this work, we present Hapstick --- a soft, flexible haptic joystick that uses fiber jamming to modulate its
h a small, preliminary human subject study and requires an average input voltage of 68.8 V to elicit tactile sensation. The senso
limited sensor life and the near-field nature of tactile sensors preclude the practical collection of exhaustive data sets even for f
discrete elastic rods simulator, we can model multiple stable shapes that fit specific boundary conditions. We then generate ac
es users with the streaming incremental reconstructed mesh in data collection for monocular 3D reconstruction. We propose an
E to exhibit "transparency" — minimising its interaction forces with the wearer to feel as natural as possible. This paper introdu
mental conditions and often fail to provide regular periodic readings, this paper introduces additional pre-processing blocks. Th
s work, we propose a novel Human-Robot interface (HRi) based on self-supervised regression of sEMG signals, combining Non
 human partner and affecting perceived safety and social acceptance. This paper investigates whether transferring the psycholo
, existing datasets often contain errors due to automated annotations aimed at expediency, making the improvement of these d
ach to this problem, where a semi-autonomous agent asks for external assistance when it has low confidence in the eventual s
This paper presents a training methodology for novice pilots of sUAVs. We presented two studies: the Baseline study (21 partic
y repetitive co-manipulation tasks, while taking into account the uncertainty in the human postural reaction to the robot motion a
ue which synthesizes queries, in a semi-supervised learning perspective. To achieve this, we leverage latent variational autoen
ation can aid in refinement and generalization, but often require significant data collection efforts to generate enough demonstra
ulator, JaywalkerVR, to obtain vehicle-pedestrian interaction data to address these challenges. Our system enables efficient, af
erally ignore human preferences in vehicle selection and assume all nearby vehicles have an equal chance of being selected, w
cally, consisting of both offline and online components. In the offline part, a global time-optimal trajectory is generated as the e
n contact/load loss, a power-aware strategy is implemented to immediately eliminate the portion of external loading in the force
per to design an elastic gripper. Subsequently, we explore the lumped parameter model for the force-displacement relationship
an impedance controller, combined with virtual workspace restrictions. Hereby, the passivity of the robotic system is ensured. T
rtest path to the goal. Therefore, an application is required, that is easy to learn and to apply even for inexperienced users. In th
hermore, this research has potential applications in autonomous vehicles, pedestrian tracking, and human-robot collaboration. T
ndle uncommon cases appropriately, but the low frequency and high diversity of such situations present a significant challenge
data that aims to verify how these biases translate into a human navigational setting. More precisely, we conduct a novel study
abilistic as the robot only observes humans sparsely through the day. The mobility of the robot adds further complexity. Conven
and do not have the capacity to generate diverse pedestrian styles. In this paper, we propose multi-agent reinforcement learnin
diction and many models in the literature allow for ego-conditioned prediction. However, leveraging ego-conditioned prediction r
ss and high computing efficiency. The environment is analyzed utilizing computer graphics, where obstacles are extracted as n
ure, and how their location in the world should change so as to make the planning problem feasible. Concretely, we propose tw
ns and provide 6-DoF grasp poses without grasp pose annotations. Specifically, the system collects the human demonstration
based method for trust recognition during human-robot cooperation. A human-robot cooperation game scenario is used to stim
or classification tasks. Our data-versus-data approach automatically quantifies distinctive differences in distributions in a high-d
we formulate a set-wise preference learning problem, and validate a cost function that captures human group preferences for co
er, we propose Self-Confidence through Observed Novel Experiences (SCONE), a policy to learn self-confidence from experien
models, we can achieve an end-to-end system from textual instructions like “Can you pass through the curtains to deliver m
hieving excellent gesture prediction and myoelectric control of prosthetic systems and neurorobots. However, factors like small
ment Learning (IRL). A set of feature functions were designed to explicitly capture users’ preferences during the task. The
mind in a multiagent setting where a human collaborates with a robot against another human opponent. We show that by only
human time scale. In addition, our results show that human evaluative feedback plays an important role in allowing Haru to ob
or assisting users in doing sports using sensory augmentation technology, which couples haptic feedback with people’s de
echanism that minimizes the influence on the front wheel steering operation. A method for realizing cooperative control with the
challenges. This paper introduces a real-world robot-assisted peeling system utilizing a single robotic arm and an assistive cutti
ethod to construct condensed representations of the coupled system kinematics over a sequence of time windows and generat
etails of the path for each navigation session, which results in high explanation costs for users. To solve this problem, we propo
an-centered autonomous framework that infers geospatial mission context through dynamic feature sets, which then guides a p
bot interaction systems a promising technology. This study presents a novel gaze-based human-robot interaction system, desig
g software infrastructure. Herein, we introduce a framework that uses the Robotic Operating System 2 (ROS2) and Bluesky to
es for human-robot trust as well as the Unified theory of acceptance and use of technology questionnaires. We performed two p
s an under-investigated topic. For the first time, we investigate the relationship between the EoP and muscle activity of the forea
ns, the existing works do not account for language instructions and have mainly focused on surface sensing, i.e., perceiving the
ponent in the first approach, and a pre-defined torque control based on gait speed and phase in the second approach. These c
velops us. We discuss the motivation for such robot-rooms, present initial work aimed at their physical realization, and report on
To this end, we report on a data collection for two tasks during collaborative draping, the transport of large cut pieces and the
ance control. The outer-loop MPC ensures precise control of the robot's posture while tracking the target person's velocity and d
Guided Therapy robot. VR-based user experiments confirm the increase in performance of the automatically tuned MPC shared
bservable Markov decision process (POMDP) framework for modelling the interactions between humans and robots. To specify
s on the human's actions, which may lead to inefficient interactions. We instead formulate the problem of optimally choosing a c
planning, with applications to human-robot teamwork. However, this approach has yet to be applied in an online teaming assista
We demonstrated the effectiveness of our approach in a variety of experiments and comparisons with other HRC assembly ap
ntroduces a trio of diverse agents, each with its unique capability to become an active member in the mapping process: mobile
ment of generative adversarial networks (GANs). In this paper, we introduce a training methodology based on GANs that utilize
ngle demonstration and a combination of data-driven machine learning methods with logic-based declarative learning to learn a
path, which may negatively affect the operator perception of the robot. In this work an optimal approach to address together the
ping them in charge of complex control phases. In this letter, we propose a task segmentation of the bimanual peg transfer pro
t-to-speech (TTS-Sonemes) and synthesized musical indicators (Tone-Sonemes).To profile the system's capabilities with respe
sed on a time-invariant potential field, capable of encoding wrench profiles, complex and closed-loop trajectories, and additiona
However, policies trained in multi-agent RL are often sensitive to the policies of other agents. In such a case, a trained caregiv
n learning (tGAIL), which is employed to train the robot from the expert's demonstration and show competitive generalization ab
d to be able to respond to feedback that can take arbitrary corrections about high-level human preferences to low-level adjustm
rol translation and rotation independently. Our research explored and quantified the factors that influence operational intuition.
nomic forces” that guide the human toward more ergonomic postures, based on online monitoring of a kinematics-based ind
nverse reinforcement learning (IRL) to enable the robot to learn path planning from expert demonstrations with fast convergenc
assivity of FIC to adapt to diverse cooperative scenarios. Unlike traditional impedance controllers, our approach ensures stabilit
lisation capabilities in robotics decision-making through knowledge extraction from Natural Language (NL). However, the transl
ons for single-agent policies, but little has been explored in generating explanations for collaborative policies. In this work, we in
ensure that the robot performs exactly the planned path. To optimize robot performance while complying with safety regulation
egulations regarding the co-working human. Meanwhile, the human can safely interrupt the robot via physical contact to enter p
gulated during the collaboration. Based on these guidelines, we propose a new optimal trajectory planner which, by exploiting th
multiple tasks by single-round self-thinking. In this work, drawing inspiration from the expert consultation meeting, we introduce a
ological and cognitive architecture tailored for HRI based on the interplay between robotic personality and memory-based cogni
pairs). However, when simply attributing a single preference to a pair of trajectories that contain many agglomerated steps, ke
engaging with content already painted by a human collaborator. To improve text-image alignment, FRIDA's major weakness, o
dress the computational constraints of mobile platforms, we propose a lightweight yet accurate model MateViT to perform pixel
the user, traverse a collision-free set of waypoints in a coordinated manner, while avoiding static and dynamic obstacles throug
aper, we propose a set of seven fundamental operations to incrementally learn, improve, and re-organize MP libraries. To show
of statically stable poses called ``waypoints". The paths between the waypoints can be modeled with a simplified kinetic motion
st learn a model of the cloth's dynamics. Then, we optimize over this model to uncover a given target limb using information ab
middle, collectively known as the serial-position effect. For example, if a human's last observation of a robot involves a task failu
nvironmental conditions. The influence of such factors can be captured through Maps of Dynamics (MoDs), which encode spat
s these methods unsuitable for personalized home service robots. Although research in continual learning develops methods th
s. We here formulate this as a mask-free image inpainting problem, and propose a novel method to estimate the complete vers
ctionality. Although BCIs have been designed for robotic wheelchairs, more research and development is required before they c
annotate a new dataset containing 3510 pairs of RGB-IR images to support model training. Finally, we conduct extensive expe
matically extracted from a certain viewpoint and a neural network was employed to create simplified contours. Then, expressiv
a real-time multi-agent visual SLAM system that co-localizes all robots and mixed-reality devices to a common reference frame
obot in the formulation of energy-states, with their interaction characterized as energy-exchange to reduce dimensionality. To im
orrection. The enhanced force model captures fundamental kinetic results considering the robot foot shape, walking gait speed
ck from the reference gait due to the terrain changes. The controller is validated through the simulations on the mixed uneven t
al differences between arbitrary human demonstrators and humanoid robots. An adversarial critic component is integrated with
oidal Model Predictive Controller (NC-MPC). The EBG is responsible for optimizing the physically feasible whole-body template
engths between the sensory neurons and the interneurons in the first two layers are learned using the bio-plausible reward mod
n of various fault types and estimates the lead time for potential falls. Our contributions include the development of an algorithm
human-like skin to imitate body dimensions, shape details and surface curvatures of target bodies. Structurally, the inner skeleto
n control architectures for human-robot interaction. The proposed sensor fusion can also integrate distributed, non-collocated fo
n this paper, we propose the first Simulation-to-Reality transfer learning framework for multi-agent cooperative perception using
t conflicts stemming from heterogeneous sensor signals.Therefore, we propose a novel Eliminating Conflicts Fusion (ECFusion
caused by time asynchrony across cameras; $2)$ information loss in transmission process resulted from limited communication
sify the state of such systems. Furthermore, we propose the visual intention former, a two-step transformer-based architecture
self can provide crucial depth cues. In this work, we introduce a framework based on contrastive feature distillation(CFD). This s
propose a novel end-to-end method for multi-class road defect detection and segmentation. The proposed method comprises
(Vehicle-to-Vehicle) technology enables efficient information sharing between different agents which significantly promotes the
und-truth label generation, and elaborate network design, aiming to achieve superior performance. However, the inference spee
in real-world scenarios challenging. In this paper, we present the 1st framework for end-to-end 3D human pose estimation, nam
ng a dependency on supervision solely on the whole network's output, single-frame input, and the utilization of a small backbon
w number of false-negatives. This can negatively affect tracking by making data association and track lifecycle management mo
perception system. The data silos by the above Distribution Gap could result in a significant performance decline in multi-agent
ngs. Given the pivotal role of GPS signals in perception fusion and the potential for various interference, it becomes imperative
and tracking. This paradigm, however, ignores any errors from upstream modules. Meanwhile, an emerging end-to-end paradig
y explain the concept, we propose a cooperative perception framework, termed QUEST, which let query stream flow among ag
Our approach hinges on the estimation of fog attenuation coefficient, facilitated by road-based prior knowledge. Subsequently,
maps, which would inevitably lead to information loss. Besides, overlapping observations between the two sensors would lead
ooperative detection and leave cooperative tracking an underexplored research field. A few recent datasets, such as V2V4Rea
to closed-loop generalization abilities, i.e., deployment realizes a causal feedback loop that is not present during training. This
ent domain with a human prompt video poses a significant challenge due to the big embodiment and environment difference. To
and the greedy strategy of existing methods, we propose an approach of policy optimization by looking ahead (POLA) to mitiga
align its end-effector with the novel object to enable effective interaction. Through a series of real-world experiments on everyd
can serve to guide robot learning for new tasks in novel environments, informing both "what" to do and "how" to do it. A powerfu
rs of the reward function. This can potentially be wasteful as many parameters may result in the same reward, and many rewar
y, a general framework is developed for transferring human skills to robots in fabric unfolding manipulation. The developed fram
(retraining), and clustering, covering all the options available at the most common DL frameworks. TensorRT optimization for s
deformable objects. Previous methods focused on establishing provably correct algorithms to compute lower bounds or analytic
ship between the input to the cloth anchor point and the position of the lowest point of the ball is expressed by a linear approxim
age, a spherical four-bar mechanism, and a set of bevel gears. The spherical four-bar mechanism is designed to provide 2-deg
onducted through an empirically defined one-step curtail. Small feature reduction does not sufficiently remove the actor’s f
llers like Operational Space Control (OSC) or joint PD control, which induces straightline motion towards these action targets in
method, which consists of two parts. First, an in-gripper mechanics model that computes a naive motion cone assuming all para
ng controllers available via domain knowledge. To this end, we design simple sub-skill controllers and demonstrate improved sa
nspired by human tactile-visual synesthesia. This approach allows for the simultaneous and seamless integration of both senso
he object remains manipulable. These challenges are exacerbated once the hand gets disturbed or when the internal dynamics
whereas three other suction cups called the middle cups, are attached to the end-effector of the parallel mechanism. The climb
the non-convex cable model, with grasp quality autonomously labeled based on simulated grasping attempts. The CG-CNN is r
smission effects. The proposed approach combines several benefits, including the possibility of a continuous execution of trials
cation, an interactive voice controlled wizard system for the AI-assisted bootstrapping and parameterization of robot programs,
ources efficiently is lacking. Moreover, there is an absence of an advanced navigation algorithm for facilitating coordinated nav
nt importance for practical applications in the computer vision domain. To effectively address these challenges, we present a no
FoV. Furthermore, we model the camera extrinsics as part of the kinematic model and use camera measurements coupled wit
ng solvers. Additionally, we supplement a game-theoretic stochastic surveillance formulation in the literature with a novel greed
racting with unknown surface geometries and with a human teleoperator in the control loop. Keeping a teleoperator in the loop
erformance. In this paper, we present a sampling-based SNMPC approach that leverages recently derived sample complexity b
aps have been utilized to understand free-space. However, predicting a grid for the entire scene is wasteful since only certain s
, whose decision variables are their rate of change. This results in a quadratic programming problem which directly provides sa
Neural Predictive Control (NPC), which does not use any physical model for the vehicle. After training with over 20,000 km of hi
pendent safety margins for collision avoidance, the evolution of which is derived separately from the robustified problem. Additio
re, this paper proposes a framework for learning-aided warmstarts of Model Predictive Control algorithms. Our method leverage
significantly reduces the sample complexity in training, allowing us to train high-quality policies with much fewer samples comp
sfy output constraints and thereby, increases MPPI efficiency. Comparative simulations and experiments of dynamic autonomo
mes across multiple queries; ii) lazy approaches to collision checking and propagation of control sequences that decrease the c
to asymptotically provide the globally optimal path. Our method is compared against a provably asymptotically optimal RRT* m
and dangerous to collect. In order to achieve data efficiency, this paper presents an algorithm for actively selecting trajectories
en traversing such environments at high speeds. We address these problems by proposing an algorithm which plans trajectorie
tion and Planning (BeBOP) that combines two approaches for generating behavior trees: we build the structure using a reactiv
otion Memory, which allows different motion planners to accelerate future planning using past experiences. Treating existing mo
mphasis on the context of visual imitation learning. As for the vector-based model, we observe that the states of surrounding ve
hs generated by sampling based methods as expert supervision data and train networks via regression techniques. However, th
manipulation tasks due to local minima. In this paper, we present a diffusion-based method that guides an MPC method to accom
mulate the sampling bias problem as a non-stationary multi-armed bandit problem, where the arms of the bandit correspond to s
CAD model, however, this is error-prone and a cumbersome user experience. We propose a correspondence-free alternative:
ent stochasticity and can only be approximately determined a posteriori to system failure. This paper proposes a generic progno
nce remotely; we call it the digital robot judge, or “DR.J.” By detecting key events on the board through performance circ
nizing boundary conditions, obtaining semi-analytical solutions for the corresponding PDE model. Model transformation is assig
the plate. Subsequently, a specialized algorithm detects a narrow seam line within the segmented image and generates a stitc
e contact point on the finishing disk as an additional degree of freedom. Different strategies for covering the workpiece surface
nufacturing system that integrates three essential control aspects: robot assignment, maintenance scheduling, and product qua
verage massive computing power in robotics applications, safety and reliability remain critical issues. This paper presents a min
en Vehicles(HDVs). To tackle this issue, this paper devises a coordination approach based on Conflict-Based Search(CBS) the
ontrastive divergence loss to address the non-IID problem in autonomous driving by reducing the impact of divergence factors f
major limitations, i.e. combinatorial scenario generation inevitably leads to many redundant scenarios, and each ODD query ma
man supervisors and investigating how autonomous agents can cooperate to achieve safety. This article focuses on the safety-
arameters. This approach is grounded in multimodal trajectory prediction, using a deep neural network with LSTM-based socia
oal quErieS with temporal Transductive alIgNmEnt (DESTINE) method. Unlike prior approaches, our approach 1) dynamically p
nd propose a computationally-efficient parallel optimization framework to generate strategies for CAVs with the travel efficiency
we should be able to evaluate autonomous vehicles against reactive agents with different social characteristics in the simulatio
CDAS) that issues warnings tailored to the driver's perception of the driving environment and driving style. In this paper, we pre
ot consider the safety factor in the deep neural network (DNN) training process. Thus, state-of-the-art DNN penalize all misdete
riving scenario, since the LiDAR points follow an uneven or even long-tailed distribution across the space, which prevents the m
ing complex heuristics to predict the states. In this paper, we propose STT, a Stateful Tracking model built with Transformers, t
is important to recognize that not all CAVs contribute valuable data, and some CAV data even have detrimental effects on colla
We evaluate this algorithm on a standard set of benchmarks for transit network design, and find that it outperforms the learned
arameters and exogenous time-varying disturbances affecting both quadrotor and load. The main feature of the UDE is its abili
ing a feedback-loop mechanism to fine-tune envisioned autonomous systems. Thus, a way to automatically profile autonomous
g the challenges posed by radar noise and point sparsity in 4D radar data, we introduce RaTrack, an innovative solution tailored
reinforcement learning (RL). However, most existing studies use precise observations that require domain expertise and hand
g a `pure-language' strategy. In this research, we propose a hybrid End-to-End learning framework for autonomous driving by c
ng in a direct improvement of the final planning performance. For conditional prediction, we introduce a query-centric Transform
of agents in the scene, some works leverage Euclidean distance to prune far-away agents. However, distance-based metric alo
Specifically, we conduct an extensive labeling effort to identify causal agents, or agents whose presence influences human drive
human drivers on highways at high speeds. To this end, we used reachability analysis to find safe velocity bounds on occluded
policies that generalize to multiple traffic scenarios is still a challenge. To address this, we introduce Multi-residual Task Learnin
on keeps the number of outstanding requests uniformly bounded over time. Although, rollout-based approaches are well-suited
f AV models. However, limited work has been explored on the reuse of these extensive scenarios to iteratively improve AV mod
nalysis to build a causal graph that defines our framework and the loss function; then we propose a novel domain-agnostic learn
ness. Additionally, their heavy network architectures impede real-time applications. To tackle these problems, we propose Map
changes are modeled as discrete events. Next, we select the best driving corridor based on a cost function that penalizes lane
prediction of target vehicles by an Episodic Memory implanted Neural Decision Tree (abbrev. eMem-NDT). The structure of eM
arnings are mainly generated in a one-shot manner without modeling the ego driver's reactions and surrounding objects, which
drivers through this passive planning. In this paper, we propose a novel dual control framework based on Model Predictive Path
Learning, where a structured rulebook is integrated as a knowledge source. We learn trajectories and asses them with a situat
cerning the relative importance of each cost term in the embodied MPC are generated by a neural policy. Furthermore, effectiv
o trajectory prediction models for AVs. We introduce a novel ''adaptive visual sector'' mechanism that mimics the dynamic alloc
VAE employs a dual attention mechanism for observation encoding that accounts for the environmental context information and
ctive participation. Although various systems have been proposed to enhance upper limb training, few studies addressed the in
epresentation of the robot's anticipated trajectory based on its recent position, velocity and acceleration history. Taking arm dyn
 classifier and integrated with an assistive robot. Central to our approach is a rich feature representation composed of the tange
sion information from a single microexpression image. Then, We combine the idea of cyclegan and optical flow consistency to m
e state-of-the-art performance in predicting future driver actions. Compared to existing methods, our approach explicitly extract
simple imperative commands (e.g., "go to the fridge"), we examine implicit directives obtained through conversational interactio

ensive empirical studies of existing models on these scenarios to expose shortcomings and strengths of different approaches.
s hand trajectories and gaze data to foster robust, real-time HMIP with minimal to no training. A novel 3D vector field method is
d of time. We present a bioinspired passive dynamic foot in which the claws are actuated solely by the impact energy. Our gripp
joint design and introduce bulk to the system. To address this constraint, we present a bio-inspired antagonistic robotic shoulde
f the proposed Variable Stiffness Spiral Cam Actuator (VS-SCA) are described, focusing on the cantilever beams with adjustab
. Consistent force during scrubbing increases the rate of removal for a contaminant. An elastic robot which has rigid links and lo
the manipulator's profile. To address this issue, the kinetostatic model is presented and simulations based on both the interval
rch direction. In this paper, the influence law of the tendon transmission path on the tension and displacement transmission is o
ted and analyzed using bifurcation diagrams, complementing the current state of the art and adding new knowledge about robo
ntally and three different mast models are proposed.
omprising integrated actuator units, pulleys, and a novel pretension mechanism connected by Bowden cables, is utilized to con
izing a neural network, our approach simulates inverse kinematics, converting measurement data into joint-specific parameters
miniature water jumping robot with rowing driving legs. The hydrodynamic model between driving legs and water is established
fastest progress on the path when selecting the highest feasible target acceleration. Similarly, it is possible to calculate a traje
redundancy of its inverse kinematics solution is found, establishing a circle of tip locations. For three-segments, the redundancy
Dubins path are built into the constraints of our manipulator model. Furthermore, we show that the number of solutions is not co
, this index is the minimum of a discontinuous scalar function leading to difficulties when computing gradients thereby rendering
ess of a robotic arm to perform a specific task based on human demonstration. Fitness of a robot arm is a measure of its ability
motions that incorporates both translation and rotation, and additionally exhibits several invariant properties. This representatio
hat for the most part, the isochoric (locally volume-preserving) constraint dominates behavior, and this can be built into closed-fo
contact matter because they affect the motion of the object. The physical quantity called hitting flux depends on the robot's con
em. In this context, we present RASCAL, a novel ASRS robot for small payload items in structured environments, with a focus
n each side only generates 2-DOF translational motions in a vertical plane. By treating each of the middle link as a 2-DOF virtu
agging CDPRs are local extrema of the functional describing the robot potential energy. For assessing the stability, it is then ne
tform (MP). By modeling the cables as elastic springs, the proposed approach correctly predicts which cables become slack, es
esults in a conservative usage of the actuator capability and hence hinders the robot's actual feasibility. In this study, new appro
emblance to gibbons and the integration of a tail mechanism for energy injection contribute to its unique capabilities. The paper
e EIC-based control. We first identify these conditions and then propose an enhanced EIC-based control with a Gaussian proce
e regularization technique to improve the controller performance of safe controllers by reducing the conservativeness inside the
imitations of current CBFs: optimality, stability, and feasibility. In this paper, we proposed a novel and unified approach to addre
ly addressed in previous works. For that purpose, we reformulate skills as base and failure prevention skills, where base skills a
from past activities, this approach requires large amounts of data to achieve acceptable long-horizon predictions. More importa
y. This integration is achieved by efficiently gating multi-modality and inputting it to the high-level RNN. We verified the interpret
s exceptionally complex. However, there is currently no control or learning method that takes all of these effects into account. I
y meaningful environment features and their effects on action success. While the CBN structure is often intuitively transferable
ES)-based quality diversity algorithm, does not have these limitations and has achieved state-of-the-art performance on standa
s, and (2) difficulty in perceiving depth information when working with a monocular camera that displays its images on a monito
ays reach their goal, independently of their initial state. This can be achieved by modeling motions as dynamical systems and e
tions. However, current MPs can only learn from low-variance demonstrations. The low-variance demonstrations include varied
of difficulty. During training, the GNN-based policy learns the underlying rules of the manipulation task focusing on the structur
atures for motion segmentation. This method includes a feature extraction phase using a differentiation process and detects seg
ots. However, propagation among robots without direct human supervision can result in robotic skills mutating from the original
ons. Our proposed semi-supervised LfD approach decouples the learnt model into a haptic representation encoder and a motio
eural risk-bounded path planner to quickly find near-optimal solutions that have an acceptable collision probability in the comple
vised training mainly focuses on imitating previously observed behaviors, and the training demonstrations carry no notion of wh
s, which act as strong priors for predicting actions given the world state. Learnt rules act as propositional knowledge and enable
hen disengaging the clutch. We implemented passive gravity compensation to counteract the robot’s weight, but this mecha
a knowledge expansion technique. We demonstrate the effectiveness of WayEx, our waypoint exploration strategy, across six
ould lead to improved surgical practice, and importantly, patient outcomes. However, such studies would require a large numbe
gistic function as a delayed goal system. This formulation inherently has low initial jerk, and generates the bell-shaped velocity p
or teleoperation and demonstration collection. As collecting teleoperated data is expensive and time-consuming, we further leve
od that infers constraints based on preferences over demonstrations. The main advantages of our proposed approach are that
ands proficiency with additional hardware. This paper introduces an alternative paradigm for LfD called emph{Diagrammatic Te
for the same task can be fundamentally multimodal, like a mug hanging on any of n racks. We propose a method that retains th
fficient training of scalable neural policies while formally ensuring global stability. SNDS leverages a neural architecture that ena
extremely conservative when the traditional linear parameter-varying formulation is used. Thus, different from previous approa
h domain randomization, which aims to broaden the span of the training distribution to cover the test-time distribution. However
ory generated by the learned NODE. A correction term to the NODE model is computed online by solving a quadratic program t
rooted in neural network technology. This algorithm not only possesses the capability to extract critical insights from demonstra
analysis approach to derive parameter constraints to guarantee the unique-minimum property of the neural energy function, whi
nted reward and a task-agnostic constraint, enabling the generalization of learned skills. Our method TCL additively decompose
an skeleton information extracted from a single red, green, and blue (RGB) camera. The point of this framework is to only extrac
odel of the environment can alleviate this issue. In this paper, we make two key contributions: 1) we introduce an offline model-
aximization fails alone in producing useful and safe manipulation behaviors. Furthermore, tackling this by augmenting skill disco
corporating elements of interaction control, in which the robot can adapt key parameters, such as feed rate, depth of cut, and m
ded for pre-training a diverse set of skills. Our method uses two core ideas to automatically expand a base set of pre-training ta
t guarantee the effective information interaction between representation learning and RL. To learn more efficient states, we pro
learn highly dynamic behaviors. In this paper, we propose a learning-based approach to enable robots to learn highly dynamic
rovide an effective solution to this problem. This paper proposes TWIST (Teacher-Student World Model Distillation for Sim-to-R
es supervision for a partially-observable one. We find that the quality of the supervision signal for the partially-observable pursu
method. Despite its effectiveness, ODT still suffers from the sample inefficiency problem during the online fine-tuning phase. T
d potential in leveraging these traits to enhance robustness and generalizability at the control policy level. We seek to develop a
y barrier functions, offering simplicity and ease of implementation across various environments and tasks. To evaluate the effe
e. Moreover, manually choosing suitable augmentations for a variety of tasks is a tedious and time-consuming process. To mitig
method is capable of estimating weights for policies that are much smaller than commonly used networks yet can represent hig
e between flexible improvement potential and focused, efficient exploration. APRL enables a quadrupedal robot to efficiently le
that instead, using a torque-based action space enables task-and-robot agnostic learning with less parameter tuning and mitig
 l observations can impact global motors through the single coupled neural network policy. In contrast, animals and humans can
eptive sensors. We also release the tools we use for collecting, processing, and querying the data, including our metadata syste
nevitably introduces considerable computational overhead and latency. (2) The process of enhancing images prior to training ob
diction method for a largescale terrain. Initially, precise terrain models are constructed employing Gaussian sampling (GS), ther
high-fidelity scans of [real]-world environments (UNRealNet). This network can be deployed on-robot to predict these high-fidelit
ation between the LiDAR point clouds and satellite images. And then, we can easily calculate the horizontal offset and angular d
ers (DMVAEs) for estimating the traversability of different robots in complex off-road terrains. Our method utilizes DMVAEs to c
leviate the local optimum solution and better utilize distant terrain information. First, the entire planning framework is adapted to
acities, making them unable to offer realistic and diverse data needed by various application users. In this work, we study trajec
rring. Then we conduct a number of transferring experiments on various manipulation tasks. Experimental results demonstrate
ethods need two crucial requirements: (i) a state representation space to search valid subgoals, and (ii) a distance function to m
about object interactions based on a keypoint representation. Using visual observations, our model learns to associate objects
inforcement learning to address intricate long-horizon with sparse rewards robotic manipulation tasks. We evaluate our framew
al Network (RNN) can be used to learn latent representations for handling observation and action histories. However, this appro
propose a learning framework. We first design a unified Transformer-based architecture to understand multi-modal data and ou
arning and policy distillation. After training a teacher policy to master the motor control based on object pose information, TAPG
that learned skills can be used by a task-level planner. Using an ablation experiment in four simulated manipulation tasks, we
oothly interacting with and assisting a person. Existing works for dynamic catching usually perform target prediction followed by
or better task performance with MPC controllers. This weak correlation can result in data inefficiency as well as limitations to ov
a non-trivial problem. To tackle this issue, we propose a planning-guided reinforcement learning method. It leverages a task-evo
planning, along with a set of adaptive intra-skill goal-reaching policies. Such design aims to solve the robotic furniture assembly
iverse objects in simulation, it enables the policy to generalize to unseen objects. However, leveraging simulation introduces th
costs, and utilized reinforcement learning for its solution. Our approach involved representing rearrangement tasks using a task
ures that guarantee generalization across all possible object translations, 3D rotations, and scales by construction. EquivAct is
rbations. In this work, we study the vulnerability of state-of-the-art DRL subject to large perturbations and propose a novel adve
LOs poses significant challenges, particularly in achieving large deformations that require highly accurate global models. To ad
mechanism in the brain to keep the neuron active for previous events. Inspired by this discovery, BioSLAM designs a gated gen
high-dimensional sensor data, such as Lidar or RGB cameras. The use of subgoals based on a compact intermediate represen
on the quality-diversity trade-off and show that we can obtain diverse policies while imposing multiple constraints on the reward
itecture, wherein the high-level orchestrates global navigation strategies while the low-level uses curriculum learning for local n
lizing radar scans in a LiDAR map of the environment, by jointly learning to address both place recognition and metric localizati
ntally reconstruct maps of unknown scenes. Specifically, voxel-permutohedral encoding combines a sparse voxel feature grid c
2e) deep reinforcement learning (RL) can explore a rich class of policies, delivering surprising performance when facing comple
al representation and interpretability, constraining navigation performance improvement. To solve this problem, we propose two
uch works build task-specific networks using limited datasets, we show that the existing foundational vision networks can accom
ar simulation framework. Most importantly, human-robot gym is the first benchmark suite that includes a safety shield to provab
amic movements while learning to reach their goals, can lead to conflicts with human pedestrians due to their homogeneity. To
es visual attention maps to detect pedestrians in the driving scene and dynamically switch between prioritizing safety or efficien
policy to the neighborhood of the baseline policy to accelerate the optimization process, and the third step addresses the cons
backward, even though it is symmetrical about its torso. This issue becomes further pronounced in high-dimensional or compl
me and maximizing the dual-arm cooperation efficiency. One of the difficulties in the task assignment is the scalability problem.
 icy from offline data to guide online exploration is a promising solution to mitigate the conserveness issue. Large-capacity mod
on the current position within the labyrinth, providing valuable information about the labyrinth layout. The learning of a control p
esulting in inferior generalization. Topological maps are a promising alternative as they consist only of nodes and edges with a
e present the semantic pointcloud filter (SPF), a convolutional neural network (CNN) that learns to adjust LiDAR measurements
he legged robot and three for the arm), off-the-shelf reinforcement learning (RL) algorithms struggle to learn effective locomotio
is employed to discover simply represented control policies, eliminating the need for a predefined reference motion. Then, we
posed approach is applicable to complex systems (such as quadruped robots). In addition, the proposed approach also enables
in a single composer policy in an independent process that is not conditioned on the other expert policies. By reusing the same
ow level is a motion-conditioned control policy that allows the quadrupedal robot to track desired base and front limb movemen
perties. Deploying such a motion to a physical robot can cause even greater difficulties due to the sim2real gap. We develop LA
obot to walk with impaired joints. The proposed framework consists of three components: 1) a random joint masking strategy fo
s, as the controller may produce irrational output within a short period of time, further creating the risk of serious physical injurie
sively leveraging a library of previously learned control policies. Our framework combines multiple levels of pre-learned skills by
aximize task rewards at all times, even as perturbations are being applied. As an alternative, we introduce an explicit recovery
mic Programming and Raibert heuristic, which serves as an expert policy. Then we train a clone of the MPC using imitation lea
PC while improving the robustness and agility of blind legged locomotion. This integration leverages MPC's strength in predictiv
redetermined obstacle course in labs. In contrast, humans are able to learn parkour through practice without significantly chang
der safety explicitly. Instead of relying on a value expectation, we estimate the complete value distribution to account for uncert
e uncertain scenarios such as abrupt terrain changes or unexpected external forces. In this paper, we consider a novel risk-sen
haping terms to obtain satisfactory task performance. To save the efforts of tuning these weights, we adopt the Extrinsic-Intrins
lity through confined 3D spaces, such as narrow tunnels or irregular voids, which impose all-around constraints. The cyclic gait
recovery strategies often struggle with complex task logic reasoning and locomotion robustness evaluation, making them susce
the future states of the robot in response to the walking strategies. This formulated NMPC enables the seamless application of
the region of attraction for the discrete-time dynamics. We illustrate the use of this metric towards synthesizing nominal walking
several interesting design features including an ALIP-based step adjustment strategy, stance leg length control as an alternativ
planning running and jumping motions. We propose a highly efficient motion planner that generates stable center-of-mass (CoM
ation is then used as states for training a robust RL-based gait policy, eliminating the need for heuristic state selections or the u
ditionally, the use of kinematic measurement results in an increased output state frequency of about 1kHz. This ensures tempo
balance and steer the robot. The newly developed controller can manage the non-linearities introduced by the new leg design a
otor, and thus it is named as PogoX. We show that with a simple and lightweight spring-loaded leg, the robot is capable of hopp
bination of phase observations, simple phase-based rewards, and the local feedback dynamics induces policies that exhibit em

CoM) and support area influenced by spinal flexion. During trotting, the robot balance is significantly influenced by the distance
y and foot motion, as well as a contact sequence that closely tracks the key points; 3) track the reference trajectory online using
ruped robots to negotiate higher obstacles and adds safety to the locomotion by allowing the robot to negotiate obstacles under
-wheels, the system uses its three legs to reorient the body during flight in preparation for landing. We control the leg motion fo
obot sizes, while maintaining the same degrees of freedom (DoF) and morphology, requires either complex learning framework
piral theory, we examine the stability margins associated with different leg movement sequences. Notably, our findings confirm t
for manipulation. By training a reinforcement learning policy that tracks position targets for one foot, we enable a dedicated pe
e it to train a Variational Autoencoder (VAE)-based policy network through supervised learning to imitate the natural behavior of
ata from sensors to compensate for defects in the control model. Specifically, we employ the Autoregressive Moving Average V
del predictive control (MPC) algorithms are proposed to address the optimal control problem of the interconnected SRB dynamic
argets for the RL policy to imitate, leading to the development of robust control policies that can be learned with reliability. Furth
al value of control input to the reaction wheel for one step becomes zero, and the control system for achieving this is designed b
ntally validated using an exemplary single leg-wheel of the leg-wheel transformable robot. The limb trajectory optimization is dev
reference trajectory. The controller then iteratively learns the feedforward control signal in a repetition process, without requirin

behaviors. In this work, we develop a mapping algorithm that combines planar region measurements along with kinematic-inert
sing perturbations to be mapped arbitrarily far away from the target trajectory. These properties make it difficult for local feedba
ard. An optimization-based control pipeline is developed through careful derivation of the system's equations of motion, conside
ative analysis of the experience accumulation process and experimental validation of the kinodynamic feasibility of the generate
ates the legs and propeller, enabling the robot to achieve speed and trajectory tracking tasks in the presence of actuator failure
loying reinforcement learning to develop a versatile controller applicable to a wide range of robots. In contrast to the convention
nding how it positively influences locomotion is unclear, which is significant for quadruped robot to achieve efficient and stable w
LiDAR inputs, our proposed framework can still generate high-quality 3D annotations of static map elements. Specifically, the a
nd sensor frequency. It uses up-to-date devices and presents effective procedures to accurately calibrate the intrinsic and extrin
propose a novel spatial-aware dynamic lightweight monocular depth estimation method (SAD-Depth). Specifically, we propose a
s of images improves performance. In our recent work on measuring domain gaps between image datasets, we proposed a Vis

ticularly when training with binary pairwise labels, and complex re-ranking strategies are required. This work introduces a fresh
ges such as illumination variation, occlusions, and dynamic objects, contemporary approaches have primarily focused on mode
nition method by incorporating depth information as auxiliary data to enhance the robustness of the VPR algorithm. The pipeline
trategy, which suffers from the problems of misalignment and information loss. To tackle this, we propose a novel Multi-modal B
er (between different processors and memory) that consumes significantly more energy than computing itself. To push the state
and landmark locations. Within this framework, we utilize the TPV-Former to convert surround view cameras' images into 3D s
se between volumes as pose graph edges, the entire neural feature field becomes both locally rigid and globally elastic. Locally
g on nighttime conditions. Specifically, we use pretrained visual foundation models to extract generalised features across challe
We exploit the trajectory from a real-time lidar SLAM system to bootstrap a Structure-from-Motion (SfM) procedure to both sig
update system based on ground semantic features, utilizing low-cost cameras. We present a precise and lightweight parameter
hods usually run offline. That is, the map is reprocessed after it is constructed, which undoubtedly increases additional time co
se neural radiance fields to represent objects and couple them with a lightweight object SLAM based on multi-view geometry, to
and free states using ray-shooting. In this paper, we propose an octree-based PVM, called OctoMap-RT, using a hybrid of off-
y, a series of data augmentations are introduced to enhance the training of the fundamental network. A curvature suppression c
diction, leveraging a state-of-the-art vision foundation model for improved out-of-distribution performance. Our method employs
sability in an off-road environment: surface slope, semantic information, and robot platform. We present two strategies for estim
atterns is nontrivial. This letter presents the first self-supervised learning (SSL) method for learning to infer navigational patterns
rge networks, leading to significant consumption of computational resources. In this paper, we propose a high-performance tea
hermore, we devise a batched BnB algorithm to fully leverage GPU parallel processing. Through experiments in simulated and
p. It contains the Instance Occupancy Check module for initial dynamic instance proposal and the Instance Belief Update modu
d categories; 2) sample inef- ficient in training. Focusing on these two problems, we propose a brain-inspired plug-and-play met
of existing implicit scene representation based methods. Opposed to prior art that optimizes a neural field in a self-supervised
metric feature-based localization is hampered by fluctuating illumination and appearance changes. Our preference for SOD ov
 ational load caused by maintaining large-scale maps. Firstly, we introduce a method for generating block maps and the corresp
ce map. An efficient feature descriptor, the dominant energy curve (DEC), is designed to identify different locations in cluttered
truction due to their precise distance measurements, a capability fundamentally absent in visual-based approaches. Researche
e that produces filtered images with improved brightness consistency within the image as well as across different scenes. We e
ovides efficient particle sampling. For an efficient neighbor particle search, it uses locality sensitive hashing and iteratively upda
correction that results in robust tracking of the sensor pose under severe point cloud degeneration and defective regions in a m
pposing viewpoint VPR that relies exclusively on structure estimated through stereo visual odometry (VO). The method extends
-to-fine 3D matching. Our key discovery is that by introducing the reference image, the source image's fine points are screened
wide variety of conditions is essential. In this paper we introduce a novel distribution-based approach for online outlier rejection
required to maintain point-cloud-based global maps. Recent solutions have resorted into neural networks, using satellite image
h uncertainty introduced by the dominant photometric loss, which leads to unreliable depth estimation in difficult regions and det
2) includes a re-ranking stage that re-orders based on local feature matching the top-k candidates retrieved using a global repre
ion by solving constrained ESIKF updates in the front end and incorporates multisensor constraints, even when dealing with ou
tion. To address this problem, we propose a new visual token-guided VPR framework that contains a semantic-focused patch t
R-inertial navigation system (FF-LINS) for solid-state LiDARs. With the INS-centric LiDAR frame processing, the keyframe point
he reconstructed semantic map may not contain sufficient information for registration. To address this problem, VioLA makes u
methods. Our objective is to estimate a robot pose within a wayfinding map, utilizing RGB images from perspective cameras. W
bsolute pose regression model, TransAPR, to tackle this problem. Upon the traditional CNN backbone, we design Transformer
encoding holistic image information for the first stage and local features encoding the appearance around salient image points
or SCR. Despite NeRF's efficiency in rendering, many of the rendered data are polluted by artifacts or only contain minimal info
collecting isolated images, which can be done in an automated way with little supervision. As a mitigation to this problem, we pr
nd catastrophic forgetting. In this paper, we propose a continual contrastive learning method, named CCL, to tackle the catastr
alman filter is enhanced through a single-rigid body model that incorporates ground reaction force control outputs from convex M
ghbour cloud descriptors — generated by an existing encoder— utilising the concept of pose-graph SLAM. Leveraging intra- a
anaged with the recent transformers-based local feature matching algorithms. However, under long-term changes --such as dif
ization method that achieves full pose estimation by superimposing oscillating magnetic fields for localization onto actuation fie
ned matching and fine-grained verification. In the coarse stage, our approach utilizes an attention-guided network to generate a
rmed LHMap-loc is proposed, which achieves accurate and efficient monocular localization in LiDAR maps. Firstly, feature enco
prior approaches, we directly work on the sensor data and do not assume a perfect model of the environment or rely on norma
ally, we first introduce a trainable strategy to extract robust underground features in multi-weather conditions. By further using s
ing the vSLAM with a deep-learning based ground-to-satellite (G2S) image registration method. In this framework, a coarse (sp
ffLoc's hierarchical layout, memory-bound self-attention, and feed-forward layers boost memory efficiency and inter-channel co
ently extracted by a 3D convolution network, and these point-wise labels are deeply involved in every part of the registration, inc
optimal performance compared to state-of-the-art (SOTA) APR methods. This work introduces a novel APR-agnostic framewor
tilize the underlying scene geometry for pose regression. The challenge in monocular relocalization is the minimal availability o
sing the existing methods with the spatial discerning ability. The first module constructs a key text graph that represents the top
s geometric similarity to directly query the compressed Lidar map via the real-time captured image sequence. In the offline pha
eryable in milliseconds, making it well suited for deployment on compute- constrained robotic systems. Our proposed system, V
gorithm. In addition, we show that the abuse of overlap can harm accuracy in practice. In light of these findings, we propose an
velop a universal solution to VPR -- a technique that works across a broad range of structured and unstructured environments (
ent. The key innovations are an improved database feature extraction algorithm, a dimensionality reduction method based on lo
ertial navigation system (VINS). In particular, by effectively leveraging the NeRF’s potential to synthesize novel views, the p
on for localization. Hence we propose a nocturnal vision-aided localization system in streetlight maps with a novel data associa
e frame. The behind reason is this parameter is observable. This paper provides novel proof through nonlinear observability ana
ated with semantic information. Unlike sensor-based maps, floor plans are readily-available, and do not increase the cost and ti
12 microphones, and accurate six degrees of freedom (6DOF) pose ground truth of 0.5 mm. We synchronize these sensors to
h IMUs. Besides the derivation of dynamic model equations of the proposed system, we focus on the observability analysis, for
era to achieve comparable onboard localization performance by tracking deep-learning visual features on a LiDAR-enhanced vi
oriented necessitate modifying the filter. To overcome those problems, this paper presents a novel method of fusing sensor da
IO to estimate the robot trajectory and any systematic biases that affect the range measurements in cluttered environments. Th
h significant skilled tuning effort. In this paper, we use the data wealth of the ArduPilot community to identify and highlight the m
o direct line of sight. We achieve this by fusing range measurements with inertial measurements obtained from a low cost Inertia
– databases pairing locations with their corresponding Received Signal Strengths (RSS) – increasingly accessible. These r
ch. This approach becomes infeasible as the number of landmarks increases due to the exponential growth of correspondence
n such likelihood functions are used for particle weight updates, the absence of particles within the peaks leads to all particle w
rridors or buildings. This representation is very compact, can represent different floors of buildings through its hierarchy and pro
ach that combines a multi-camera system with factor graphs for real-time and asynchronous 3D tennis ball localization. Addition
ween the velocity measured at the sensor frame and the vehicle frame. Experiments carried out prove the viability of our single-
of high-resolution Doppler-capable radars comes the possibility of estimating odometry from single point clouds, foregoing the n
d. However, few have explicitly considered the relationship between candidate retrieval and correspondence generation in pose
ptimization-based VINS resents a significant challenge – a computational bottleneck due to the need to invert the high-dimens
t personally, and allowing for qualitative analysis of user preferences. Simulation of an interaction-constrained learning task with
applications. Conventional end-to-end reinforcement learning navigation paradigms have encountered multiple challenges in me
Subsequently, the hyperparameters of the Gaussian model are trained with a specially compound kernel, and the Gaussian mo
PointNet encoder and a conditional neural radiance field (NeRF), facilitating a thorough acquisition of object deformations and
ve Control. The key idea is to embed a differentiable MPC within an actor-critic RL framework. The proposed approach leverage
e closed-loop models without retraining. We consider recent trajectory prediction approaches which leverage learned anchor em
olicy to handle different deformable objects with only a single real-world demonstration. To achieve this, we augment the policy
d carrying skills for the bipedal robot Digit. Our reward functions are designed to produce the desired interactions with the box w
e-labeled data. This paper develops an approach for learning robot dynamics directly from point-cloud observations, removing t
del inaccuracies. However, despite its real-world successes, MPC often under-performs the optimal strategy. This is due to mo
st towards modeling errors. While these policies are robust across a range of dynamics, they are sub-optimal for any particular
y from the fundamental technical ideas, the particular implementation details of these algorithms are often just as important (if n
rucially, we do not assume this structure is known a priori, and instead learn it from data. We use contrastive learning to obtain
video data (such as Ego4D), which are the largest prior datasets available for robotics, since video offers observation-only expe
rovers to traverse. Moreover, the navigation trajectories of rovers are heavily limited by the terrain topology and future systems
ethod to construct a boundary that discriminates safe and unsafe states. The boundary we construct is equivalent to distinguish
g long-horizon tabletop rearrangement tasks in which multiple manipulation primitives are required. Denoting the problem rearra
that facilitate accurate placement at the target pose. To this end, we propose a reorientation planning method, ReorientDiff, tha
emonstrated that deep models outputting human-interpretable kinematic goals can enable dexterous manipulation by real robot
w space is utilized to provide multi-point contact sensing in object manipulation. A reinforcement learning (RL) agent with a hyb
acle lies within this loop, it may be impossible to reach the goal. However, prior work has only considered the topology of the D
significant challengeremains: while previous works use perfect point cloud generated in simulation, the models cannot directly a
ed E-M4M, a graph search-based solver for solving such Manipulation tasks Among Movable Objects (MAMO). In this paper w
ecute the planned trajectory. These four steps must be performed under the manipulator's dynamic limits and extreme time con
egrate SD maps into online map prediction and propose a Transformer-based encoder, SMERF, to leverage priors in SD maps
ommon in robotics use cases. To address these issues, we first introduce a sparse structure, tri-quadtrees, which represents th
Robot-centric Implicit Mapping (RIM) technique for large-scale incremental dense mapping. This method employs a hybrid repr
rocess. In this paper, we propose a two-staged pipeline that normalizes images with varying lighting and shadow conditions to i
-based framework, namely QuadricsNet, to parse quadrics in point clouds. The relationships between quadrics mathematical fo
Pseudo Occupancy for effective dynamic object removal. To begin, we introduce the Height Coding Descriptor, which combines
representation that leverages implicit multiresolution hash encoding aided by explicit octree SDF priors, describing the scene at
d objects. The core idea is to leverage a learnt generative model for a category of object shapes as priors and to formulate a p
e semantically meaningful and open-vocabulary queryable, conferring several advantages over previous works based on pixel-
behaviours, e.g., “follow the corridor” or “turn right”, while avoiding detailed, accurate spatial information at the m
roblem, this paper generalizes the particle-based map into continuous space and builds an efficient 3D egocentric local map. A
ects in the current 3D LiDAR scan and a local map of the environment. We use sparse 4D convolutions to extract spatio-tempor
roblem of estimating normals for sparse LiDAR data avoiding the typical issues of smoothing out the normals in high curvature
mapping results, mainly due to inaccurate camera poses. This paper presents OmniColor, a novel and efficient algorithm to colo
e uncertainty measure for the maps is required. In this paper, uncertain building models with abstract map surfaces using Gaus
n-tree update strategy on a tree-based map structure. Thirdly, we remove known grids from the map at each update by leverag
an refresh and back-end removal modules, which realizes real-time map construction and online dynamic object removal (DOR
efinement techniques employed for RGB-D sensors are not widely applied to LiDARs. This paper presents a novel BA photom
pendage-free, and largely seal-free. Preliminary tests highlight its motion capabilities and set new benchmarks for centimeter-sc
the robot can acquire two locomotion modes, namely monostable and bistable modes, and it can also swim at different energy
on heavy hydraulic rigid manipulators that deep-sea diving robots are commonly equipped with, we introduce a new concept fo
ons do not capture the full complexity of the underwater environment, and deploying a UVMS at sea for testing/validation is reso
rag caused by aerial structures. By having a compact size and successfully balanced aerial and underwater performance withou
gned Snapp, an integrated robotic fish capable of swimming in open water with high speeds and full 3-dimensional maneuverab
bility. First, the design of the robot, which is featured with four novel articulated-compliant robotic legs, is given. Second, its hyd
mpling system has been designed according to user requirements for microplastics (MPs) analysis from sampled sediment, whic
eraction forces and feed them into a controller for determining thrust inputs. The key idea lies in using supervised learning to ob
ime. Though, most solutions remain overly cautious and conservative. This paper proposes RUMP, a novel locally-optimal mot
Our model extends recent state-of-the-art approaches to monocular image based depth estimation, using an efficient encoder-d
ata is not clear and the sensors expensive. For this reason, the community has focused on extracting pose estimates from RGB
environment stems from the dynamic and frequently limited visibility conditions,restricting vision to a few meters of often feature
or estimating uncertainty are often computationally intensive and have largely centered around Bayesian neural networks or en
s a combination of light attenuation with distance and backscattered light. The proposed neural scene representation is based o
ws), obstacles (e.g., ground plain and overhead layers), scuba divers, and open areas for servoing. Through comprehensive be
h beamforming. These observations are merged in a Bayesian Occupancy Grid to produce a 2D map of the acoustic activity of a
he LiDAR points in a two-octree structure instead of using traditional single-octree coding, and then design the cross-attention
out with eight cables of varying stiffness, weight and buoyancy. One end of the cable is fixed, while the other end is moved by
el approach for calculating the buoyancy and righting moment of a partially or fully submerged UV in order to model interactions
 cterized and defined. The novel surfing algorithm consists of the gradient perceptron, time-limited momentum gradient search,
-inefficient and produce significant noise and disturbance. Underwater gliders, on the other hand, use changes in buoyancy and
ceans and rivers, which are constantly subject to significant external disturbances, remains an ongoing challenge for MPC. It m
d of propulsion device named propeller-leg to have the ability to run rapidly in the seaside and seabed, and swim with multi deg
ylic-air) interfaces, and offset effects of a non-centered camera with respect to the dome. In this paper, we present an N-Sphere
an autonomous underwater vehicle in West Antarctica. Using this dataset, we systematically benchmark the performance of 2 c
elines. Previous works offer simulation environments and deployment pipelines for underwater vehicles, but fall short of providi
vely small set of parameters. The resulting behavior - both collective actions and individual actions - can be understood intuitive

However, the power generated at the recharging stations may be constrained depending on the sea conditions. Therefore, a ro
he complex interactions. To address this problem, we propose a decentralized multi-ASV collision avoidance policy based on D
y. However, performing belief propagation is computationally costly, especially when operating in large-scale environments. Th
nt marine scenarios. Sea-U-Foil features three distinct locomotion modes, displacement mode, foilborne mode, and submarine
view adaptive surveying and reacquisition using side scan sonar imagery.Our framework addresses this challenge by using a g
t a novel electric underwater manipulator, named SEU-4. This four-degree-of-freedom manipulator weighs 8.91 kg and has a m
with low computational power on a Raspberry Pi4. The framework builds upon Synthetic Scan Formation (SSF) where single M
haser AUV with a dynamic target by modeling the scenario as a factor graph optimization-based Simultaneous Localization and
ne water clearance is critical for maneuvering when approaching the quay. In this work, we present a vision-based approach to
which is significantly affected by water quality. In this paper, we present a vision-based method that utilizes adaptive color LED
n camera images, and the segmented region is projected onto the water surface using planar homography. A line-detection alg
alibration formulation leveraging epipolar constraints to estimate the refractive index and subsequently correct for distortion. Th
een frames producing relative orientation constraints between consecutive frames. The soft and hard iron effects are calibrated
n error, IMU preintegration error, and raw LBL measurement error are jointly minimized within a sliding window factor graph fram
hrough images collected from imaging sonars. A representation of the occupied space is obtained by exploration of the free-spa
les. To address this challenge, we leverage the problem's intrinsic structure, i.e., the payload at the end-effector being enhance
hod, which involves drilling into wooden structures to identify internal voids. The research centers on two key enabling technolo
d representation of the UAV and its suspended load’s dynamics and kinematics is achieved, facilitating the realization of loa
capable of translational motion while maintaining a constant attitude. Our newly developed platform has a cubic exterior and ca
g force distribution, as well as recognizing local textures. In this paper, we pioneer the use of vision-based tactile sensors on ful
in and high energy efficiency compared with running and rotor-based take-off. Despite recent research, there is no FWAV capa
erching. To enable stable perching and unperching maneuvers on/from a vertical surface, a lightweight (approximately 1 kg), fu
wever, these systems are inherently complex. They are often underactuated and evolve in nonlinear manifold configuration spa
velop a realistic hardware-in-the-loop (HIL) simulation of the perching and take-off manoeuvre of the UAV in Gazebo. A control
ty of consumer UAVs is usually limited to approximately 3m/s. This limitation reduces the potential of UAVs to execute tasks in
d-loop equilibrium configurations in the presence of the aforementioned uncertainties, and then we study their stability. Hence,
actuated pod for data collection. The system consists of a quadrotor drone, a slewing ring mechanism allowing 360° tether ro
hilosophy of disturbance rejection, we propose a predictive optimization scheme that allows aerial manipulator to successfully ex
mmands, which makes our method generally applicable. The identification uses an offline gradient-based method on flight data c
gramming optimization. Through phi-theory, we illustrate a typical predicted longitudinal and lateral flight envelope of a tail-sittin
avoid those surfaces; taking advantage of the efficiency benefits from proximity effects may be a route towards longer-duration
m and the asymmetric sinistral/dextral Markov-Dubins problem. The proposed formulation can be used to model an Unmanned
 Value at Risk (CVaR) costs are modelled to provide a holistic monetary cost model for path cost minimisation. A number of futu
become significant when the object's surface is moving and inclined. To address this problem, we proposed a real-time trajector
al Actuated ROtorcraft (SICARO) is chosen to harness this approach fully. The SICARO is capable of flying with either side of t
or and the work surface while preserving contact, requires that all available contact points at the end-effector tip are in contact
s for various tasks. To control PADUAV's 5 degrees of freedom, we propose a versatile and straightforward 5-DOF geometric t
stimated state values in close proximity to the manifold, while requiring significantly fewer computational resources than the stan
essitating margins to counter potential disturbances. In contrast, our novel approach introduces high-speed quadcopter control
precise rolling and turning control as well as improved rolling efficiency. Experiments were conducted to comprehensively valid
ex. We propose a co-design optimisation method that assists the engineers by proposing a morphing drone’s conceptual d
nces, a novel control framework is presented to by integrating the advantages of active disturbance rejection control (ADRC) a
l input and attitude-error state of the UAV. Specifically, this method uses a control law with a term whose sign is dynamically sw

orce generation of underactuated aerial vehicles can present the potential challenge of near-saturation operations. The blind ut
poles' large moment of inertia by adopting a ring design that accommodates the pole in its center. To grasp the load, we develo
y adapted to various robotic arms when achieving multi-task aerial manipulation. In this paper, we present a learning-based con
e formulate the collaborative transportation of payloads by a swarm of drones as a novel, under-capacitated generalization of v
nd the controllable DOF of the aerial system. To take advantage of the mechanical simplicity of a standard quadrotor, we propo
is an extended Kalman filter (EKF) with intermittent measurements obtained at different frequencies. An optimal state feedback
me trajectory adaptation while concurrently optimizing safety, time, and energy. The planner employs a risk-aware cost function
smissions. This approach offers several advantages such as enhanced backdrivability, hardware safety, and transparency when
ons. Here we show how combining the Johnsen-Rahbek (JR) effect with the exponential tension scaling capstan effect can prod
ngs to fold, and embedded torsion springs to release energy to realize the function of wing unfolding. In the experiments of the
ocus on developing a new kind of magnetically-driven bristle-bots with a wireless control and power supply that can be steered
se. We propose a new type of VSM that improves on the pre-existing Jack Spring concept by making it more compact and robu
(SRBs) that constitute the TM and are used to maintain the shape once the extension force is relaxed. We demonstrate modul
Fundamental parameters to describe the dynamic behaviour of an actuator are discussed, and an experimental method to dete
arm switching function through its reconfigurability. Secondly, the motion space and singular configuration of the exoskeleton a
mnidirectional driving gear for supporting flexible displays discussing its design, motion range, repeatability, positional accuracy
M mechanism, the gripper implements the function of zeroed offset, which removes the gap between the finger and the suction
echanical and electrical connections without actuators. Using xacro for writing the file of the robot model description of each mo
gami robot were analyzed, explaining how elastic restoring force is generated in each module. A physical prototype was develop
eel mode, collinear four-wheel mode, or upright two-wheel mode; it could even smoothly switch between any two modes by re-
ovel leg and toe mechanism inspired by human biomechanics, featuring a one-DOF knee joint, two-DOF ankle joint, and one-D
rge mass. This paper proposes a new approach for multiple autonomous sailboats to tow floating objects. A system of linear ar
e translational wiping motion. Also, the theoretical analysis for the existence of an optimal rotational speed at maximum wiping p
gh-bandwidth joint torque regulation and the outer one producing various types of compliance responses for effective humanâ€
width dimension, and a kinematic model is established to simulate the deployment process for the pattern. The hinge design e
al axes of bending stiffness of the beam. Then, the vibrational phenomenon of the actuator is theoretically revealed. It is shown
mit navigation inevitably in constrained environments. To realize grasping in such scenes without extra devices, we propose a ne
nature and designed a morphing robot with multi-functional thruster-wheel appendages that allows the robot to change its mod
sign method for eSKTs has been proposed. By combining simulation analysis with experiments, the eSKTs with multiple strips h
proposes the use of two lower-limb exoskeletons to mediate physical interaction between therapists and patients during a StS r
orce servoing control scheme of the extracorporeal ultrasound-integrated tracking arm (EUTA) are derived based on the interac
robably inducing physical damage. This paper considers a non-contact formulation, where the robot spins to generate a repulsi
ess. The steerable section is a cable-driven hinged structure, and through an antagonistic mechanism regulating the cable tens
nt insertion through the trocar, are more challenging in robotic surgery than in conventional manual procedures since the robot c
stimation and biomechanic performance of one able-bodied individual walking on a robotic ankle prosthesis via a bypass adapte
s cognition is often governed by machine learning algorithms, such as convolutional neural networks (CNNs), which provide se
le camera, limiting their ability to accurately estimate the depth of haustral folds in colon, which is of great importance for the ac
rgery, the retired first-generation da Vinci robots have been redeployed for research use as "da Vinci Research Kits" (dVRKs),
f the probe with the patient's body, a critical aspect of ensuring safe and effective ultrasonography. We introduce a novel appro
ovations include the hybrid coaxial continuum unit (HCCU) for improved bending characteristics and structural stiffness, and a d
on protocols and the possibility of automated acquisition. Additionally, robotic ultrasound systems enable access to 3D data via
n inaccessible in most operating rooms. Targeting intraocular assessment within the current surgical setup, this paper presents
plications with limited acoustic windows. To address this challenge, we proposed a skeleton graph-based non-rigid registration t
l position of the US probe. Our proposed system is comprised of the following three processes. First, an exhaustive scan of the
entification module that leverage the Doppler effect to enhance the final segmentation result. Firstly, the DopUS-Net combines t
F surgery based on intraoperative freehand ultrasound (US) 3D reconstruction is proposed. Building upon this, a CMF osteotom
ng room, and status changes after installation can also reduce accuracy. At the same time, there is a strong vibration coupling
vely, making it challenging, even for experienced clinicians utilising dedicated ophthalmic surgical robots, to peel these delicate
ol algorithms is proposed to accomplish effective automatic cranial-milling tasks. First, a pure data-driven model-free adaptive c
ack function which is not available for the existing controllers for microsurgical robots. A challenge of designing such a controlle
obotic system tailored for assisted drilling in skull base surgery. Our central hypothesis posits that this collaborative system, en
ystem consisting of a mobile platform, an RGB-D camera, and a robotic arm with a spray disinfection device. To address the ab
elicate blood vessels (arteries or veins) near the lymph nodes. Therefore, we propose a framework integrating augmented realit
mploying both vision-driven and force-driven mechanisms for continuous and large-scale FOV adjustments, respectively. The s
matic model to become unreliable. Therefore, there is a need for a more accurate kinematic model in the absence of a complica
on information of the interventionists and generate force to assist the interventionists in performing surgeries on the leader side.
to the distal end. This feature offers important advantages, particularly in enhancing safety by minimizing physical contact risk
achine. A flexible hinge was incorporated into the amplifier design for simplified fabrication and reduced friction in the actuation
ic field, fixed magnetization configuration causes the magnets on the MCRs exposed to coupled magnetic forces and torques,
es the robot back and forth in human intestine. The internal embedded targeted delivery mechanism is fixed with a radial magn
image-based visual servoing with online robot Jacobian estimation has been implemented to enhance motion accuracy. Furthe
and subsequently reconstructing the shape using optimization algorithms. This optimization, however, is under-constrained and
, learning, workspace estimation, and path planning, among others. In this paper, we investigate a mapping to mitigate branchi
led by applying an external alternating electromagnetic field to the capsule’s inner magnet to generate vibrations and impac
bined with the rotational movement, generates a propulsive force. Visual feedback is suitable for in-vitro closed-loop control. Ho
ving endoscope camera. Automation may reduce the complexity of surgical operations and consequently reduce the cognitive lo
nations of the features were used to develop different lesion characterisation models. Supervised classification using Multi-Lay
to its well-known benefits of small spatial footprint, fast data acquisition, and produce higher tissue contrast images. However,
ace of the robot, the equation is derived between the structure parameter and workspace to adjust the parameters of the robot t
erefore, analyzing environmental motion and vibrations is crucial to ensuring the optimal performance and reliability of medical
 otherwise challenging and high-risk surgeries with high precision and safety. Steady-Hand Eye Robot (SHER) with cooperative
required extensive manual tuning of core parameters, which significantly depends on prior knowledge and may potentially lead
a flexible capsule endoscope (FCE) design concept, specifically conceived to navigate the inherent volume constraints of capsu
a wireless MAGS consisting of soft material and structure design. The camera can bend at the exit of the trocar and maintain s
erges as a promising prospect in this context. This study presents the development and implementation of a magnetically-guide
rasping model is constructed, and the factors affecting the force measurement are quantitatively analyzed. Secondly, the simula
arbitrary locations along the flexible robots, which is highly required when with large deflections in robotic surgery. In this paper
(ASTR) and conducts a feasibility study by performing autonomous midline partial glossectomy for pseudotumor with millimeter
sioning and tele-cutting of a U-DT. We have shown the system's ability to cooperate with a remote surgeon by leveraging an on
odular robotic system for microsurgery enabling the surgeon to operate using conventional instruments. Our system was imple
, ideally, considers the comprehensive factors of the eye anatomy and its positioning, the insertion point, and the initial pose an
cisely navigate minimally-invasive surgical guidewires. The system is designed based on consultations with surgeons, and ana
full-wave propagation. To this end, we proposed a closed-loop hybrid control scheme and a learning-based modulation model
is a Co-Axial Slender Tubular robot (CAST), and modelled it using novel approaches. Another contribution is tendon-mortise sh
as used to track the real-time position and orientation of the guidewire tip, and a state space representing the guidewire within th
marily their rigidity and the lack of active control over the bending tip of the catheter. This paper introduces a new soft micro-rob
geometrical design parameters. To thoroughly evaluate the performance of the proposed modeling framework, we have conduc
ct with soft tissues. In this paper, we present a novel mechanics-based 2D bevel-tip needle model that can account for the effec
but also clinical and research contexts. Utilizing the Unified Robotics Description Format (URDF) to implement a symbolic Pyth
y pick and place components to assemble microrobots with a two-fingered micromanipulator, and orient these components thro
erages bare NiTi film (∼ 1 μm thick) for actuation, enabling a high actuator bandwidth up to 50 Hz. A comprehensive analytic
systems. Inspired by air bubbles in nature, herein, we report a kind of novel magnetic air bubble microrobots. The air bubble-b
r microrobot design features a compliant spring-like structure for significant deflection, enabling a gripping motion under specific
nels using biofuel cell (BFC) and electroosmotic propulsion (EOP) and a magnetic rod using magnetic steering (MS). The BFC
microrobot has a cross-sectional area of 1 mm2, and its design is readily compatible with conventional microfabrication techniq
behavior of helical microrobots driven by a rotating magnetic field, researchers have proposed a mathematical model for multihe
bryos, a popular model organism for vertebrate development. Our work leverages both classical and deep learning-based imag
motion and control strategy. The extraction of implicit human attention during the task and integration of this with robot control co
aper, we propose a novel dual-arm nanorobotic system setup for automated knot tying performed on mini-incision under stereo
ulation tasks obtain accurate depth information by detecting the contact between an end-effector and an object (e.g., a cell). De
sing accuracy, restricted imaging field of view, and uneven illumination. To overcome these limitations, we have innovatively dev
network is proposed to detect and track the needle tip of micro-scale displacement end-effector within the limited field of view un
rical joint based on the PC-MEMS process, aiming for miniaturization and cost-effectiveness. The spherical joint allows rotation
abrication and folding assembly processes. Our previous work has proposed an untethered robot S2worm equipped with a nov
underlying physical principles, leaving their theoretical limits elusive. This work constructs a theoretical framework from governin
limitations in analysing small sperm objects under microscopic imaging. The emergence of convolutional neural networks (CNN
cale visual servoing capabilities. The auto-calibration technique addresses the challenges posed by the limited depth of field an
he three-dimensional position detection and calculated the homogeneous transformation matrix. We secondly employed hybrid
Iterative Closest Point (GPA-ICP) algorithm. To validate the effectiveness of the approach, an AFM image of a TGX1 calibration
te. Manual sperm immobilization suffers from inconsistent success rates, and current robotic systems developed for the task fa
ntation suffers from large context loss and feature distortion due to bounding box cropping and resizing during ROI Align. More
 trajectory design required for actual implementation. The use of real-time adaptive robotically controlled sampling with micro-sc
duced to manipulate cells more easily, but most of them are restricted by expensive devices, limited work area, and potential da
tiating the feasibility through experimental validation, showcasing the direct teleoperation of UMRs within a fluid-filled lumen. Th
million annotated suction grasp poses. The dataset generation process combines analytical models and dynamic physical sim
rtains to object- level attributes and where a robot has subroutines for seeking and accurately observing objects. This models s
uality, safe, agile and feasible trajectories for mobile manipulators in real time. First, we present a novel environment-adaptive p
arget objects. These new challenges require combining manipulation and navigation skills in unexplored environments. We pre
s a whole-body constrained model predictive controller (MPC) for a mobile manipulator that balances objects and avoids collisio
s. However, most of the existing AMPC approaches used in mobile manipulation require either expert tuning or extensive traini
hat enables a temporally consistent grasping observation. Specifically, the predicted grasping poses are online organized to elim
otic arm and the legged robot. On the legged robot side, we explicitly estimate the wrench exerted by the robotic arm on the ba
rk, we introduce an active perception pipeline for mobile manipulators to generate motions that are informative toward manipula
cs of nonlinear robotic systems. We combine offline learning from past experience and online learning from current robot intera
challenges using the task of robotic sculpting with a parallel gripper. We propose a system that uses point clouds as the state
iscipline of Autonomous Racing (AR). In this work, we explore the use of Learning-based Model Predictive Control (LMPC) to a
a method for learning terrain-aware kinodynamic model which is conditioned on both proprioceptive and exteroceptive informati
cluding imprecision in manufacturing, chemo-mechanical forces, and poorly understood contact mechanics. Drawing inspiration
egression problem. The proposed extension enables memory-efficient online learning in addition to the benefits of geometry-aw
es on single-arm manipulation of stiff DLOs (StDLOs) connected to form a DLO network (DLON), for which the measurements
densely represented in training data. In this work, we explore the use of physics-informed neural networks to learn an end-to-en
Using the mean estimates from GPR, we propose a data-driven dynamic motion model that is more accurate at predicting future
of such metrics and the absence of standardized criteria impede their usage in novel contexts. In this work, we introduce a unifie
oft tactors to produce touch motions mimicking human massage. A key design principle behind MiBOT is its silent actuation, wh
ate variation in history for the emergency event, named the Extro-Spective Prediction (ESP) problem. Based on the proposed da
ificantly minimize the reliance on real-world data. Leveraging modern 3D pose estimation techniques, SynthAct can be applied
postures, the relative distance of the exoskeleton claws, and tactile images. Using the demonstration data, we summarized thre
uate the suitability of a workstation over a large population of workers in a physics-based simulation. Activities are simulated th
muscles to generate reasonable human movements. To fill this gap, we build a musculoskeletal model with 90 body segments,
o address these challenges are relatively scarce. In this work, based on the problems studied above, we propose a keypoints-g
te them. Due to their high computational cost, forward dynamics-based optimization algorithms were not viable for real-time est
first solution, to the best of our knowledge, to close the gap. We propose an energy-based model that is integrated with the Lu
uation mechanism for propelling eversion; an approach especially useful for miniaturised systems. Simulations are experimenta
M) bending actuator, the innovative control strategy abates hysteresis effects by dynamically modulating gains within a tradition
matic representation of a soft robotic arm with pseudo universal and prismatic joints that are capable of capturing non-constant
y confined workspaces greatly complicates the search for feasible IK solutions. Without an accurate initialization or multiple re-s
rs in the learned dynamics model and the optimizer can be extremely tedious. In this work, we explore using data-driven MPC t
promise accurate shape estimation. We propose a robust method to estimate the shape of linear deformable objects in realtime
rol scheme is developed with RBF NN, which can not only provide stable and accurate tracking control for the soft robot, but al
his challenge, this letter proposes a purely feedforward iterative learning control algorithm that refines the torque action by lever
on for a FEA to the application of active suppression of hand tremor, which allows a wearable FEA actively exerting force on th
approaches result in changes in the system’s dynamics. Optical measurements are problematic, especially in confined spa
 learn the relationship between the actuation pattern and the target landing position, i.e., the direct model of the task. Then, a re
VR-Soft IK uses a simulation based on the Finite Element Method (FEM) and a control method based on sensitivity analysis. A
ing of these robots, which is very different, as it is based on the mechanics of deformable bodies. These models, whose dimen
l misalignment between the everting sheath and the embedded catheter. Second, a switching control architecture is proposed,
tendon-driven soft robot with six embedded microphones. We first introduce novel applications of 3D reconstruction methods to
mprising actuator and locomotor dynamics based on data points from stochastically perturbed, repeated behaviors. By deriving
ek to address the issues of modeling and controlling such structures, particularly when underactuated. We introduce several sim
mploying this setup, we successfully demonstrated a pincer grip using a cadaver hand and measured both muscle forces and g
for cost function weight tuning using the previously developed and validated Adaptive Reference IOC (AR-IOC) algorithm. Give
innovations in gripper design and force control. Fully passive grippers and wrist joints allow secure grasping while reducing mas
tion tape (MST) and a detachment mechanism has been proposed. This design was implemented in the construction and expe
s). While thrusters can in theory be used to regulate the vehicle pose, it is often insufficient to establish a stable pose for precise
on method that overcomes the curse of dimensionality. We also identify a unified geometric representation of locomotion by ge
es of both wheeled and legged locomotion, the effective hybrid planning of the mobility actions that transparently combine both
onlinear dynamics. Reinforcement learning (RL) methods can learn the policy to satisfy the STL specifications via hand-crafted
method to allow for computational reuse in larger multi-query problems, however, requires additional thought. This paper shows
s can also be used to guide sampling-based motion planning and thus increase the effectiveness of planning in environments w
ith the multimodality of the optimal distributions. This is due to the less representative nature of the Gaussian. To overcome thi
e of the grid, and requires at most ten arithmetic operations per grid cell. In the proposed approach, we use a linear first-order h
ese challenges, we introduce two innovative techniques: 1) an efficient and precise clothoid approximation using the Gauss-Leg
e-of-the-art (SOTA) methods for multi-robot planning in discrete domains to the motion planning of multiple Ackerman (car-like)
t consider the robot's energy constraints, limiting how long a robot can explore. In fact, if the robots are battery-powered, it is ph
y sampling the entire state space and incrementally rewiring the random trees raises drastically before a feasible solution is fou
rward kinematics not well-defined, and characterizing the set of feasible robot configurations is non-trivial. In this paper, we prop
iency. The first method can significantly reduce planning times compared to other multi-robot planners. The second method he
etermined by the current path cost. Learning-based alternatives model the topology of the free space and infer the states close
cialized hardware. Our key insight is how to exploit fine-grained parallelism within planning, providing generality-preserving algo
r a receding horizon. We propose two algorithms capable of solving these problems: a dense sampling approach and an impro
ectives and stochastic motion dynamics. However, effectively addressing such reactive trajectory optimization challenges for ro
g angle constraints are satisfied. Conditions are provided under which the proposed algorithm is guaranteed to find a winding a
arse to be practically useful. Experiments were conducted to show that this differentiable representation can reproduce scores
nstant-time motion planners (CTMP) was introduced. These planners employ a preprocessing phase to compute data structure
al cost map, and processed proprioception data as state inputs, and learns the physical and geometric properties of the surrou
pecially designed cost functions for a given scene, the overall solutions tend to have low success rates within a certain time limi
spaces with a small number of polytopes. The approach constructs a visibility graph using sampling, and generates a clique cov
tion strategy for the pursuers. To explore the space of possible pursuer motion strategies, the algorithm utilizes a trio of hierarc
n space due to the absence of obstacles. This article puts forward a general and systematic post-processing algorithm for A* an
on space. In this paper, we extend the construction of infeasibility proofs to higher dimensions by adapting Coxeter triangulation
structure. We first demonstrate the relevance of this approach for standard path-planning tasks. We also propose a frontiers-ba
efined targets, each of which could be visited from multiple potential locations. To increase autonomy in these missions, we pro
esulting to a sum of safe landing points. Due to the low processing time achieved by state of the art object detectors and the ef
otential of the gates is not fully utilized. To address this issue, we present a time-optimal planner that can faithfully model gate c
rent choices for each stage, the eventual performance depends critically not only on these choices, but also on the environmen
, we present a real-time planning method that generates smooth trajectories considering the time-varying shape and non-linear
by integrating forward the current position and velocity estimates using an extended kalman filter, and implementing cubic inter
an Filter (UKF). Additionally, since the NMPC relies on the future behavior of the system, the trajectory of the ground target thr
to collect the rewards. The model characterizes the dynamics of the water surface and enables the planner to predict future sy
work with the human operator to transport the payload collaboratively in outdoor settings. A system comprised of a rigid object li
al. The flexible multi-DoF aerial 3D printing is a newly developed framework and has the potential to strategically distribute the e
ngle-motor joints (SMJs), which achieve a more comprehensive performance. For DMJs, which are over-actuated systems, we

rs’ fault is modeled as a lumped disturbance acting on the virtual control of the quadcopter system. The disturbance estima
ly feasible joint space commands. The crucial element underlying this approach consists on mimicking the end-effector's motion
ng scale vehicle platforms do not support four-wheel steering and cannot isolate the performance of motion control algorithms f
uity set is firstly constructed from movement data by leveraging a Dirichlet process mixture model and is updated online using r
actions. This is all the more true with human-robot collaboration where human capacities may fluctuate over time and have to be
a control strategy that modifies the control input over iterations based on previous results. ILC has been successfully applied to
This paper proposes an FFC approach for industrial manipulators using a six-axis force/torque sensor (F/T sensor), implemente
ynamic constraints in the environment. This planner leverages high-order CBFs based on the forward reachable set to ensure s
ys with a roller arm. To address this problem, our framework encompasses the following key elements: 1) Trajectory generation
rod-like end-effector may contain errors. This paper proposes an acceleration-level remote center of cyclic motion (ARC$^{2}$M
able under safety constraints. In this work, we present a novel data-driven approach, RAGoOSe, for safe controller tuning in th
caterpillars with three-dimensional movement commands. We first introduce the kinematics model and real-world prototype of
ework, an adaptive-gain controller and a super-twisting sliding mode controller are designed to speed up the convergence of tra
chniques used. Due to their appealing features, Piezocermamic-based fine positioning stages were successfully utilized in man
in the control structure. The proposed method is called constrained NRIC (C-NRIC), because the optimization allows us to imp
ng the surgical procedure. Three controllers are designed initially, including a master hand position (MP) controller and a maste
d enables explicit force tracking in a hybrid manner. The unified framework features compliant behavior in the free (motion) task
cess. In this article, we propose a novel control strategy, based on the synergistic usage of state-of-the-art passivity-based cont
ce and precision motion control. This control method allows for precise manipulation of forces and torques at the end-effector,
ks, but it lacks physical interpretability and flexibility in re-purposing the policy for a different set of tasks. For instance, we can i
l control-affine system. Further, we analyze the propagation and bounds of state prediction errors caused by the truncation in d
ection expensive. This paper proposes a cutting-surface-aware Model-Based Reinforcement Learning (MBRL) method for robo
-time adaptive multi-contact model predictive control framework, which enables online adaption of the hybrid multi-contact mod
ground vehicles (UGVs) input limits and gathering empirical model training data. We also propose a novel learned slip approac
s in a hierarchy and decouple lower priority tasks from higher priority tasks on velocity or torque level using projectors. We deve
red state space, we propose less-actions RRT (LA-RRT), a planner which optimizes for a low number of actions to solve a puz
roach is to make the most of compute when expanding the search tree by adding the last viable configurations found when con
es at desired faces. The global planner generates a building instruction graph for a target structure that we traverse in a depth-
d an equivalent planning angle model are established, enabling efficient and analytical determination of joint positions, collision
a breadth-first search. For each gap, quadratic programming finds an optimal speed profile, synchronizing the time and space p
ates from the reference trajectory, repositions itself in configuration space, and then resumes task execution. The occurrence o
mentally verified on a quadruped platform to demonstrate safe, collision-free exploration over long distances.
al seconds to find a path and require a computationally expansive path dataset for imitation learning. Recently, physics-informe
UAV. From a small set of experiment data and assumptions on modeling errors, a Lyapunov function is synthesized by which an
h attacks to motion planners: "planner failure" and "blindspot" attacks. Planner failure attacks make changes to the physical env
lack the ability to construct the target information spatial distribution for environments that vary stochastically across space and
e regions, follows semi-free paths in narrow passages, and smoothly transitions between the two types. We specifically focus o
mni-directional tether retracting mechanism, no trajectory planner can handle the unavoidable movement errors that cause robo
er RRT. Our method combines the strength of CBF for real-time collision-avoidance control and RRT for long-horizon motion p
ator motion planning problem under the end-effector trajectory continuity constraint in which the end-effector is required to trave
ce for constraint planning without retraining or finetuning the model. We also propose to update the neural network output to mo
hole-body configurations. Subsequently, the optimization refines the solution with a short-term planning horizon to satisfy all dyn
nmanned Aerial Vehicles (UAV) to guide the UGV in unknown off-road environments. Firstly, a fast terrain risk mapping method
y and then their location. We implement our method on a robot in real time, demonstrating the ability for robots to navigate popu
of the benefit of features representing the person’s gaze in this context. Extensive experiments on a novel dataset show th
stures) and test them in a series of experiments using the paradigm of the "Desert Survival Challenge". The goal of the robot is
adapting to human preferences, or (c) only focusing on single-arm handover with small objects due to the sensor occlusions ca
s does not quite possess the intuitive control and user-centric interactivity that one would desire. In this work, we propose an L
e sensing and programmable haptic feedback, underpinned by a layered structure comprised of flexible magnetic films, soft silic

mantic information from multi-view images and are not applicable in scenarios involving severe occlusion during dual-hand inter
and employ Temporal Convolutional Networks (TCNs) to extract spatiotemporal features. Despite lacking 3D information, our a
eractions using more advanced tactile sensors or by analysing joint torques. Instead, we propose to augment the intention reco
aper, we propose TiO, an end-to-end system for interactive visual grounding in human-robot interaction. Benefiting from a unifie
hot Interactive Personalized Object Navigation (ZIPON), where robots need to navigate to personalized goal objects while enga
OG, and the corresponding dataset, Intention-oriented Multi-modal Dialogue (IM-Dial). In our proposed task scenario, an intentio
int EIT-based force and shape sensing. The hybrid approach combines physics-inspired voltage preprocessing with an attentio
uncertain, we further distinguish it between ambiguous or infeasible commands leveraging LLMs with situational aware contex
eak to users and answer questions in a free way; a dialogue balance module to encourage participation of users in the discussi
action is required to safely open basic exoskeleton domains to the user and provide an adaptable setup system. This article pre
es. To track the controller's motion accurately and robustly, we resort to various forms of visual measurements, including 6 DoF
ility of integrated aircraft positioning and navigation systems is a challenging yet highly significant problem. In this paper, we int
ever, previous works lack robustness and are sensitive to the overlap rate of maps, resulting in unpredictable performance in re
in swarm robotics. Our system supports lidar, stereo, and RGB-D sensing, and it includes a novel inter-robot loop closure prio
System (GNSS) signal, most multi-UAV SLAM systems can't work, making it hard to construct a global consistent map. In this p
need for anchor deployment. The system optimally distributes computationally intensive tasks to a potent central server, thereb
ncy of excitation condition and two estimation algorithms to estimate the relative locations in a distributed manner. Moreover, ou
oximating transformation among sensor frames. Unlike existing methods, we elaborate the inter-sensor transformation using co
toMerge, a novel framework for merging large-scale maps that surpasses these limitations, which (i) provides robust place reco
ciated with the cable-towed load. Moreover, we design observers for the hybrid interaction between the robots and the payload
er subteams or reduce the overall completion time. To enable both abilities, we introduce a new graph learning approach that fo
minimize the summed uncertainties over all targets by all robots, or optimize over the jointed discrete actions while neglecting th
 plying such strategies, any multi-robot system breaks above a certain critical system size (i.e., number of robots) as too many r
alled MESA (Manifold, Edge-based, Separable ADMM). MESA is fully distributed to tolerate failures of individual robots, asynch
diversity limits the performance of these methods. We introduce the Phase Diversity Optimization (PDO) algorithm, a populatio
to generate waypoints for the next control step and a low-level controller to guide the drone to reach subsequent waypoints. Fo
ate trajectory planning. We also provide feedback from the environment, such as collision checking, and prompt the LLM agents
creases the number of possible interactions exponentially, thereby increasing the difficulty of training RL policies. In this work, w
t damaged even crash in complex environments, and have to be dynamically substituted. It seems not robust enough for most
ent is with respect to a state value estimate derived from a potentially poor base policy. Monte Carlo tree search (MCTS) prova
ndividual robot to acquire knowledge from the competition, fully stimulating its dynamics potential in the race. Specifically, the c
ents in visual range using light, unlike radio wave like communication which is mostly assumed in existing research on multi-age
tion among robots in communication-denied scenarios, where agents can only rely on visual information from a camera with a l
here robots are equipped with different capabilities to serve discrete events in an environment. We utilize a heterogeneous cove
her, and is flexible for any type of robot, motion or sensor. We define a simple and efficient communication protocol which can
shows promise, it can lead to a lack of consensus among distant flock members and the formation of small clusters, conseque
between robots simultaneously. Under a constant velocity assumption in a short time, we fuse time offset estimation with our p
multi-robot collaboration that enables large-scale exploration in the absence of continuous communications. MOCHA is based o
g station. The UAV and UGV need to occasionally rendezvous for recharging. We propose a heuristic method to address this N
bic dodecahedron. This geometry offers an appealing alternative to cubes as it greatly simplifies rotational motion of one cell ab
d optimal control protocols derived from historical data collected from quadrotor systems without requirement for exact knowled
ations infeasible, requiring a dynamic, capability-aware allocation of tasks to team members. To this end, we propose and impl
nd shared objectives, in which the latter encourages the emergence of coordination among robots. Next, we consider the multi-
multiple agents. This extension is facilitated by leveraging nonlinear model predictive control (NMPC) to compute trajectories th
nning space to consider manipulators alone, objects, and manipulators holding objects. We represent this decomposition with a
liable a-priori knowledge or cooperation from the target. The main contribution of this work is the demonstration of a method for
peat observations of the same target. To address these limitations, we implement a more general Dynamic Targeting framewo
ly reconstructs the range image, restoring the original point cloud. Additionally, RecNet utilizes the latent vector extracted from
one or more constraints on the robots’ motion behaviour in that area. The advantages of this approach are: (i) an MPC-ba
environments are complex, therefore the use of MRS intuitively can facilitate an inspection by fully taking advantage of sensors
presentation, we extend the kinodynamic A* and the corridor-constrained trajectory optimization algorithms to efficiently tackle s
ots. Each intent message is an adaptive policy that conditions on identified points of contention that conflict with the intentions o
theoretic belief space variant of an iterative Linear Quadratic Gaussian (iLQG). However, the scalability of this method suffers d
limits applicability when there are fewer agents than targets, since agents are unable to continuously follow the targets in their
the potential outcomes of both a local robot’s and its neighbors’ next-step actions. To evaluate the effectiveness of ou
s; 2) tasks that can be decomposed into multiple fragments; and 3) multi-robot tasks that need to be executed by a coalition syn
. However, deriving suitable heuristics can be tedious, risky and in some cases impractical if problems are too complex. This ra
nds in mining. It assumes the reader has no prior knowledge of mining and builds context gradually through focused discussion
rent areas of the environment; 2) the Middle Layer allows the path planner to compute a traffic sensitive path for each vehicle;
etween robots and tasks to determine the division of coalitions in such problems and evaluate the scalability and adaptability of
y exponentially stable under a Persistency of Excitation (PE) condition. We then propose a distributed control scheme that allow
r a given formation framework, however it remains difficult to determine the geometric conditions under which otherwise rigid fo
for approximate solutions without any performance guarantee. In this paper, we bridge the gap by proposing an efficient iterativ
 ehavior of tornado schooling fish. To achieve this objective safely, we employ a combination of a scalable overtaking rule, a gu
g algorithm and deploy a swarm of 3-cm sized wheeled robots to inspect a randomized black and white tiled surface section of
ls. The order of block placement is important; for example, a block that sits atop other blocks must be placed after the blocks b
p synthesis (FFCGS) method for the anthropomorphic hand to efficiently grasp unknown objects. The FFCGS is implemented b
delicate and forceful tasks. Antagonistic tendon-driven elbows and inline actuator transmissions, reduce the system's inertial ma
ich can be produced through digital fabrication. We made the robot hand using commercial servo motors as well as componen
poral-parametric grasp prior (TPGP) optimization method to simplify the difficulty of grasping trajectory optimization for the dext
ormation cannot be directly mapped to the robotic hand due to discrepancies in their degrees of freedom or structures. Further
ers and palms. In this paper, we propose a human-sized five-fingered robotic hand named WARABI Hand. It is covered with m
oviding soft contact surfaces. In this work, we present a method that equips the skeleton of a tendon-driven humanoid hand wit
ound that vanilla behavioural cloning (BC) can learn a very proficient policy with minimal human intervention when trained on e
aped parts with the proposed hand, and experimentally verify the feasibility of desired operations with the proposed method as
us landing features of different UAVs and USVs are extracted to establish a dynamic UAV-USV cooperative landing ability mapp
oller while ensuring constraint satisfaction. By improving over some limitations in the current state-of-the-art, we are able to app
r stable formation control. These constraints are summarized as follows: 1) In a team of first (second) order holonomic mobile r
on of distributed information acquisition as a robot team explores a large space to show that we can formulate the whole proble
robotics scenarios. This article considers the operation of time-critical applications for 5G-enabled unmanned aerial vehicles (U
cused on continuous operation and features a 10 Gbit/s network connectivity allowing for high throughput remote-controlled ex
ry hardware configuration, FogROS2-Config quickly samples tests a small set of edge-case servers. We evaluate FogROS2-Co
uristics for network performance such as algebraic-connectivity or network flow optimizations and instead positions communica
We reduce conservatism of the TDPA using the model knowledge on both sides of the communication system to identify passi
the NeRF model. From these sampled points, we cast rays and deduce the color for each ray through pixel-level view synthesi
establish accurate object detection and velocity estimation with a single-frame point cloud of Livox LiDAR. Based on the non-rep
ap tool based on kernel density estimation (KDE) to make predictions in bird’s eye view (BEV). The denoised occupancy m
methods which directly predict depth distributions by using a supervised estimation model, we propose a cascade framework c
previous methods do not explicitly address weather corruption during adaptation, the domain gap between clear and adverse we
e RGB-D frames from multiple perspectives in a deep multi-directional fusion network and predicts predefined keypoints for all o
nformal inference (CI) with information theoretic measures to perform lightweight, Monte Carlo-free uncertainty estimation withi
d size estimation to mitigate the influence of imperfect scales on rigid transformations. Specifically, we leverage a pre-trained m
ge in an end-to-end manner, i.e., Feature Auxiliary Perception Network (FAP-Net). To address inaccurate pose estimation caus
nomaly-detection approach has attracted increased research attention. Anomaly-detection techniques, such as uncertainty estim
ction. Our key idea is to replace the PointNet pooling operation with an attention module, leading to a better point-to-voxel aggr
can focus on regions where the knowledge distillation works best to shrink the detection performance gap between these two
ors perform worse on smaller objects and affects their reliability in indoor environments where small objects are the majority. Th
o the current frame and apply weighted Non-Maximum-Suppression on dense bounding boxes to obtain a fused frame with me
espite their advantages, pillar-based methods often underperform compared to voxel-based and point-based, largely due to the
orm, and thus we design the network structure, WTDecomNet, approximating the decomposition of high-resolution image into i
information for accurate 3D object detection, especially in the complex driving scenarios. To address this, we introduce an intrig
d instance benefit from each other, we design a novel Gated Encoding (GE) module, incorporating complementary cues betwe
oposed to deal with the geometry discrepancy for better instance matching between LiDAR and radar. The feature alignment st
objects will occur, termed class distribution shift. Motivated by this, we propose a framework for explicitly addressing class dist
 hile the latter only fetches sparse image features to the BEV representation. To overcome their shortcomings, an optimized vie
To address this, we advocate for supervising a model to predict objects' poses given past observations, thus explicitly guiding to
improved Yolov5 model: HIC-Yolov5 is proposed to address the aforementioned problems. Firstly, by adding an additional pred
n segmentation, which is a prompt-based approach, to provide more in-depth information for robot perception. To generate high
ed objects, contend with strong distortion, handle severe occlusion, operate in uncontrolled environments, and execute real-tim
images. In this work, we seek to improve monocular 3D detectors by leveraging unlabeled historical LiDAR data. Specifically,
spatial context to characterize the segmentation model’s uncertainty and detect pixel-wise OOD data in real-time. Addition
sult in inaccurate segmentations. We build upon these methods and introduce a novel approach to correct inaccurate segment
ncertainty-aware multiagent trajectory planner and (2) an image segmentation-based frame alignment pipeline. The uncertainty
n 3D point clouds, leveraging knowledge distillation and text-point correlation. Our approach employs pre-trained 3D models th
te pseudo labels for unlabeled data. Nevertheless, these generated pseudo labels contain noise and are not as accurate as tho
es in differentiable physics and rendering. Our approach employs rendering priors to align with input images and segmentation
apping. In this paper, we propose a radar-centric dynamic occupancy grid mapping algorithm with adaptations to the state comp
ud classification. Some schemes try to reduce the gap by fusing multi-scan information through SLAM or completing single-sca
nsing, state estimation, and mapping limitations during online robot operation. The spatio-temporally conservative free space es
ts: VIP500, 4772 VI-SLAM point clouds covering 500 different object and environment configurations, and VIP500-D, an accom
d semantic segmentation method (LaCRange)}. Specifically, a distortion-compensating knowledge distillation (DCKD) strategy
arning-based radar tracker incorporating temporal offset predictions to enable direct center-based association and enhance seg
hapes to form a BEV representation of the whole road scene.Therefore, a BEV often has multiple levels of freedom on motion,

al applications, bias in the training data can lead to errors; therefore it is crucial to understand a network’s limitations at run
thod, query features extracted with convolution networks are used to represent things/stuff. In this way, different vision tasks su
. In addition, the feature extraction process prior to fusion is not sufficient, as it lacks exploration of the local semantics and con
need for meshing or normal estimation, which is beneficial for robotics applications when dealing with raw point cloud observat
s) which can be difficult to operate in the limited resources. In response to it, we introduce a new algorithm, called FocoTrack, w
underexplored yet practical task that only requires few labeling effort on the target domain. To improve the DA performance in a
uencies of the sensors. Additionally, some parts of the LiDAR data may be occluded and parts of the data may be missing due
e detects lane semantics and topology from camera images, enabling mapless path planning without the reliance on highly deta
pts to fuse both modalities at feature and instance levels. For feature-level fusion, we propose a Radar Guided BEV Encoder wh
emantic segmentation. However, existing active learning methods often perform well in the majority classes but struggle with th
nuScenes to demonstrate that LiRaFusion leverages the complementary information of LiDAR and radar effectively and achiev
eline, a practical framework must consider common labeling techniques such as sequential labeling that drastically improve ann
segmentation datasets. Second, to bootstrap the active learning loop when there is no labeled data available, they train their in
er (EKF) is employed to mitigate noise and estimate the absolute states of the target accurately. To enhance robustness, we p
curve. Symmetric negative steering (SNS) with two wheels on each side sharing the same steering angle can turn with a small
tion. One significant challenge in jerk-constrained TOTP is a non-convex formulation arising from the inclusion of third-order co
cation of a high-quality preconditioner that reduces the spread and increase the clustering of the eigenvalues of the target matri
for parallel hardware acceleration. In this work, we introduce MPCGPU, a GPU-accelerated, real-time NMPC solver that levera
rors and external disturbances can lead to deviations from the desired trajectory during tracking in real time. This paper propos
ypes of invariant descriptors are presented depending on whether the trajectories consist of screw or vector coordinates. Metho
the path as fast as possible), it is not always the most desirable objective, particularly when we seek to balance multiple objecti
 sed to define a set of behaviours that the robot should reflect in strict hierarchical order. Their main issue however, is related to
over the entire planning horizon. Then, we propose a sample-based approximation that unlocks an efficient and general-purpos
ng elastic robots. A possible remedy is the introduction of locking elements (i.e. clutches and brakes) in the drive train. This give
pure feed-forward control to perform explosive tasks. Further, underactuated compliant robots are known to be difficult to contro
on the online estimation of the discrepancy between the force predicted by the dynamics model and force measurements, comb
howcase the capabilities of geometric algebra when applied to robot manipulation tasks. In particular the modelling of cost func
a controlled forward invariant interval tube around the reference trajectory. By taking observations of the system, we then propo
Fs). However, such methods face difficulties when scaling to higher dimensional systems under input constraints. In this work, w
a convex optimization control policy, which can be efficiently evaluated in an online fashion. The developed filter minimally mod
ynthesize both under- and over- approximations of the value function for continuous-time, control-affine systems, use these app
ontact information. Then the interaction locations and forces are accommodated in the MPC framework via a spring contact mo
N) simulator has been proposed to learn the underlying dynamics. In this paper, we perform optimal control of a rigid body-drive
cal minima when the search is not initialized close to a “good” minimum. On the other hand, when dealing with continuo
h a local, linear, data-driven learning of the error dynamics. We conduct experiments in simulation, 1/10th scale hardware, and
ynamic regret, i.e., bounded suboptimality against an optimal clairvoyant controller that knows the realization of the noise a prio
d robot is achieved if its zero moment point (ZMP) always stays in the supporting region at each step. Based on this norm-base
wn. These challenges lead the collected online data to be imperfect. In this scenario, traditional data-driven control techniques,
novel co-design algorithm, namely Robust Trajectory Control with Design optimization (RTC-D). An inner optimization layer (R
ng in space. We introduce an approach that decomposes the whole-body of a robotic manipulator into multiple kinematically co
he deployment of ReLU-QP on GPUs using standard machine-learning toolboxes. We evaluate the performance of ReLU-QP a
s for which the internals of the navigation stack are presumed unmodifiable. Our system, which we call CODAK, achieves the c
g game-theoretic approach based on the Koopman operator to address the challenge. We first formulate the guided trajectory p
st. The convoy may traverse an impeded edge for an additional cost (associated with repairing the edge) than if it were unimped
led ergodicity. In this paper, we build on this insight to construct a reinforcement learning formulation of the problem of allocatin
s paper, we bridge this gap by presenting a new multi-robot social navigation environment crafted using Dec-POSMDP and mul
kinematics. In this work, we hybridize insights and algorithms from MAPF with the structure and heuristics of optimizing the cro
olution and communication cost. Our approach is tractable in close-to-real scenarios and does not necessitate prior environmen
ous intersection management for grid-based formation planning by including a new type of reservation called non-exclusive rese
s. Our approach efficiently scales to both spatially and spatio-temporally correlated environments. Moreover, our approach can
esence impacts different coordination strategies in both simulated environments and the real world. With the goal of studying a
everage learning-based approaches to enhance computational efficiency. By training a conflict area prediction (CAP) network, w
vailable free space for quadrotors to coordinate. This paper presents an Alternating Minimization-based approach that does not
as the low-level planner. We show that tracking high-level multi-robot plans with a vanilla MPC controller is insufficient, and res
perates in three levels. Initially, we compute trajectories for individual robots using a graph search that allows bounded disconti
iew (FOV) around its position, as a natural means to fix the input dimensions of its policy network. However, this often makes p
with many robots. As a remedy, we propose an algorithm that computes the paths for a subset of robots that have traversed pr
n planning methods have shown improved performance in these constricted environments. In this work, we extend an existing t
mming (MIP) model to optimally solve the RMMTC problem, resulting in an MCPP solution with a coverage time that is provably
gebraic symbolic equation and included in the N-MPC, explicitly constraining predicted positions to be collision-free throughout t
that persistently identifies the cable and estimates its complete state – required for any controller with feedback of the cable's
he magnetic field inside a split-core current transformer to provide sufficient holding force as well as battery recharging. The sy
 esented to predict the action-values of motion primitives from the current depth image and the state estimation of the quadrotor
which is usually not the case. To address these limitations, a novel exploration method is proposed that combines frontier-base
ecifically designed for slant-angled aerial imagery. Sim2Real-based transfer learning is adapted to compensate for the limited av
hannel point cloud tensor and RGB image are fused into a 6-channel tensor using an early fusion strategy for object detection b
o prevent tagging many non-corner points as a corner in FAST detection, improving SLAM accuracy. (ii) A novel Pyramidal Cull
nstead of completely avoiding collisions, this article proposes Air Bumper, a collision detection and reaction framework, for fully
sity function, Phi, which leverages CBF constraints to map the collision risk of all local coordinates. Then, we convolve Phi with
MM-based mapping works have developed methodologies to determine the number of mixture components using information-t
and Deep Reinforcement Learning (DRL) for exploration, designed to run on modern SWaP constraints neural processors. The
pre-processing.We decompose the scene into several simple subspaces by askeleton-based space decomposition (SSD). Add
g this information among robots to coordinate their behaviors in a distributed manner. To this end, we propose (1) a geometric
sions. However, autonomous navigation and exploration of previously unknown large spaces still constitute an open challenge,
a low-dimensional latent space encoding collision information while accounting for the robot size. This compressed encoding is
g for two reasons: a precise and robust whole-body planning and control policy is required for variable-tilted and narrow gaps, a
lely on onboard IMU measurements, stereo infrared images, and depth images and autonomously corrects teleoperated inputs
es applied simultaneously from different directions. By densely packing an array of flow sensors (or whiskers), we alter the path
ta processing make those methods not applicable to small robots, such as vision-based UAVs with small onboard computers. T
ccount feature matchability during trajectory planning. We seek to generate a perception-aware trajectory that reduces the erro
make sure every plant receives optimal care and ideal growing conditions. By structuring the work cell environment around a s
oar as well. In contrast to previous works, our approach does not require separate algorithms for the detection of individual upd
oling (SPP) module to enhance both the accuracy and speed. By analyzing the structure characteristic of the ResNet and RepV
ndence between large-scale teams of heterogeneous robots. Our method works like a QR code, using keypoint-based alignme
ed upon static techniques, which are susceptible to environment change during manipulation process limits their efficacy in real-
hile enhancing object manipulation capabilities. Despite the evident advantages of conformal e-skins, current fabrication method
of a cognizant awareness of the forces stemming from the object's motion and grip, can result in suboptimal and inefficient robo
We then propose a global-to-local alignment pipeline with individual global domain classifiers for scene features of RGB and de
scenes with transparent objects, and these depth maps can be used to grasp transparent objects with high accuracy. NeRF-ba
er, we employ straight edges constructed from points as underlying primitives of the representation. This is to address the hars
and monitor the liquid response using a force sensor to identify liquid properties. Based on the liquid properties, we predict the
e to constrain the possible relative configuration of the hand and object geometry. We design a generalizable implicit function, H
s of a deep learning-based vision system to first identify the individual trusses in the crate and then determine a suitable graspin

ent of detection accuracy. In order to address these, we propose a visual enhancement guided grasp detection model (VERGN
axial tactile sensor degradation model and propose a tactile pattern reconstruction framework based on the Kalman filter. This f
by observing the object in motion, requiring the robot to already be interacting with the object. In this work, we propose to use t
n about friction. Notably touch can detect the precursor of the object slipping away from the grasp. To find this information, tact
e each point may have multiple valid correspondences. Our method, SCAlign, consists of a Set Correspondence Network (SCN
novel pixel-wise weighted dense fusion architecture. This method allows for direct learning from unannotated RGB-D data facili
e induced constraints as unilateral, bilateral and fully constrained, exploiting the redundancy of force sensors. A semantic appro
cifically, we develop a dense supervision strategy with real perception and analytic labels in the spatial-temporal domain. Addit
fruits are accomplished through a multi-task YOLOv5 model coupled with a detection-based adaptive DBScan clustering algor
ction, grasping performance, and alignment with their high level task-specific goals. Given a pair of stereo images, we perform u
pportunity for robotics to excel in liquid identification and manipulation, given its potential role in chemical handling in laboratorie
tion, which causes operational burden and limits the scope of data collection and deployments. In this work, we leverage interac
e available due to occlusions. In this paper, we explore whether kinesthetic haptics can provide in-direct information regarding t
e surfaces. We present a fully automatic object pose estimation system based on a robot arm with a single wrist-mounted came
order to establish sufficient associations for accurate estimation, the system should observe an object from a viewpoint similar t
e (which occurs due to variation in camera velocity and scene texture) and avoids frame-jumps and motion-blur which are probl
le of solving a smoothing task for such objects by means of a single robotic arm. To train our network, we leverage an oracle po
against, while being compatible with existing tools and libraries. Since ground truth data generally consists of pose data from m
offline reinforcement learning research. We provide 19 datasets, including real-world human driver's datasets, and seven popu
ss the different aspects underpinning grasp execution, including the mechatronic design of grippers, planning, perception, and
ble to researchers and practitioners. We also provide our experimental results and analyzes for model-based and model-free 6
well they can handle different scenarios and how admissible and diverse their outputs are. There exist a number of complemen
respect to depth estimation errors. Specifically, our novel longitudinal error tolerant metrics, LET-3D-AP and LET-3D-APL, allo
uman-aware robot navigation systems in simulation. In addition to a general human-navigation model, HuNavSim includes, as a
g upper layers and replacing them with a test application, and grey-box testing, an application-specific measure that observes i
o primary objectives: (1) to provide a framework suitable for robotics applications to characterise cloth objects, and (2) to study
orkspace is shared with human collaborators, additional limits have to be imposed to the power the system can exert, in order t
y with humans in dynamic environments such as sports, relates to modeling human skills and designing appropriate interaction
an-Robot Interaction (pHRI). The primary objective is to regulate targeted muscle activation through the use of Robo-Coach (ro
, which allows to consider several different future paths concurrently. The second contribution is the encoding of a general noti
important for rendering a broad range of virtual environments. To address these issues, this study designs subsystem-based ad
control method for garment opening insertion during robot-assisted dressing. Specifically, we propose a model predictive contro
wer-limb performance. We designed a virtual reality (VR) based exergame with dynamic difficulty adjustment (DDA) conditions i
etween humans and the robot, i.e., physical human-robot interaction (pHRI). Studies reported varied lower limb responses for a
high granularity in the scale could overwhelm participants, leading to unclear and biased responses, while a low granularity may
ame kinematic data and the output of the Transducer -- the human motions and the contact forces over a time window in the fut
difficult to scale to arbitrary objects and human grasping motions. In this paper, we introduce a framework that can generate pla
changes is also the ability to dynamically assign a role to each participant, either taking the initiative or deferring to the partner's
n modeling and robotics. In recent years, as robots have become safer and smarter, they actively participate in our lives and he
ith 21 volunteers predicting in both cases the force that the human will exert during the next 1 s. A modification in the architectu
suspension systems and uneven terrain on human interaction experience and task performance. To fullfil this current research
urrently exist for estimating power loss at physical human-robot interfaces, especially for upper-body exoskeletons. This letter p
er introduces WANDER, a novel Walking Assistive omNi-Directional Exo-Robot. It consists of an omnidirectional platform and a
g due to the false-positive detection of the overlap and is thus not robust. Meanwhile, some state-of-the-art place recognition me
stical tool for uncertainty quantification, that requires availability of offline trajectory data - a reasonable assumption in many app
utionally Robust Safe Corridor Constraints (DRSCCs) to consider the uncertainty of the safe corridor. Then, we integrate DRSC
al disturbances exist, preventing its use in safety-critical applications. Moreover, the multi-threaded simulations used by MPPI re
ons incur collision risks and lack formal safety guarantees. In this paper, we present a computationally efficient safety filtering s
burden due to an increasing number of hypotheses with the planning horizon, which can grow exponentially. As part of this wor
hm that maintains multiple data association hypotheses, represented as a belief mixture, where each component corresponds
ees which are imperative for real-world safety-critical tasks (e.g., autonomous driving). In this work, we consider safety requirem
o contribute to low-cost plans under various realizations of the environmental uncertainty. Determining the optimal set of edges
that directly models sensor-specific aleatoric uncertainty to find safe motions for high-dimensional systems in complex environm
ify planning by using tools for low-level control given high-level action primitives (options). We introduce Constrained Options B
vironments. To address this, we propose a novel approach to estimate a 3D Uncertainty Field based on the learned incomplete
online to correct its planning strategy and adapt to the actual system response. We propose a sampling-based motion planning
namic obstacles but also include them in the plan itself, to deliberately exploit the dynamic environment in the agent's favor. The
ulation uses a Conditional Variational Autoencoder (CVAE) generative model that is enhanced with traversability constraints and
ory graph. Then, an informed cost-to-go heuristic for the original problem is computed based on the nominal trajectory graph. F
models imprecise. In this work, we address this challenge and introduce a distributionally robust chance-constrained model pre
k of future motion planning failure. When this risk exceeds a certain threshold, a recovery behavior is triggered that leverages th
ication. This makes the planning problem tractable, but the cost of following the planned path in the real world may be different
e of occluders, variations in lighting conditions, and redundancy of information due to similarity in appearance between nearby
VMDP) with partial observability and introduces Time-Varying Partially Observable Markov Decision Processes (TV-POMDP).
nments. For robotic systems that may experience a wide variety of sensing conditions, this difficulty in generalization presents a
on-Gaussian distributions utilizing particle representation of the belief state, while simultaneously maintaining high computation
currence of contacts (or the lack thereof) as observations to reduce uncertainty and reliably complete the task. We present a pr
e motion planning is limited by the inherent miscalibration of neural networks and the challenge of obtaining adequate datasets.
mpling-based planner. We also develop three different methods for robot-robot probabilistic collision checking, which trade off c
y presents an uncertainty-aware trajectory planning algorithm for the rover on rough and loose terrains. The planning algorithm
ue, this paper presents the Robust Optimal Time Allocation (ROTA) framework. This framework is designed to optimize the tim
State Graph (JSG) to implicitly incorporate the support actions, those methods do not scale well to large graphs and teams. To a
ike prior research focusing on static scenarios or rule-based heuristics, we introduce a real-time, reinforcement learning-based
h as â„Ž-signature and winding number. However, these algorithms are inefficient due to their separate treatment of topology an
set over a connected graph is a set of vertices such that there is a path on the graph connecting any pair of vertices in the set w
to start them. For example, construction robots may need to collaboratively air-lift heavy objects at different locations (e.g., pre
by causing delays. This paper addresses the problem of Multi-Robot Task Allocation under Uncertainty by proposing a hierarchi
ential congestion and increase efficiency. In this paper, we propose a hierarchical framework, which includes a traffic flow predi
s to robots by taking into account travel time and ensuring efficient execution of collaborative tasks. We extended the MILP for
legs address the aforementioned problems, they suffer from lower energy efficiency, higher volume and weight. In this paper,
paration time for patients. To tackle this issue, ultrasound-based bone tracking could offer an alternative. In this study, we prop
address this gap by proposing a novel approach that leverages Convolutional Neural Networks and Long-Short-Term Memory
s on ten healthy subjects for stair ascent and descent. The network’s performance is evaluated by F1-score, recall (true po
crete transition logic to switch to kinematic control during swing, necessitating two separate models for the different parts of the
can facilitate ankle plantarflexion- dorsiflexion and keep human body move forward steadily. A compliant crank slider mechanis
he effects of generic low-amplitude force fields applied by an exoskeleton on motor strategies in asymptomatic users. The force
nd in real life. Herein, we propose and substantiate a human-machine interface for motor control that introduces neural informat
BSE with a differential series elastic actuator (D-SEA) for lifting assistance. The unique differential working principle can accomm
oint offers two degrees of freedom (DOFs) and can support high loads while also providing adequate ROM. In addition, we desi
lows for 3 DoF ankle motion using a symmetric parallel frame design principle resulting in a strong frame while weighing 1.8 kg
factors like actuating power, balance in the neutral posture, and user's hand workspace. Our optimization analysis recommend
ng the processing-time performance. By establishing activity-specific threshold values through trained machine learning models
hen translated into joint trajectories using inverse kinematics. These trajectories are executed via an impedance controller, prom
ed hip-knee-ankle torque profiles to expert users (N=3; 1.25 m/s; 0 deg incline). The participants walked with the profile optimiz
ystem-level characteristics such as the center of mass (CoM) that play an important role in balance. Controlling the prosthesis b
the exosuit and human motion, as well as effectively controlling interactions with the surroundings, presents ongoing challenge
ates and reduces the computational cost. We demonstrate the use of the geometry-informed candidates in a set of model varia
dels offer the possibility of converting them into another model. This paper contributes to filling this gap in, to allow an algorithm
vironments, and other robotics tasks. While high-resolution sensing using radar data has been previously reported, existing me
. This abstracts away important structural information from the scenes, which is crucial for traditional registration approaches. T
n using local geometric models based on point-to-distribution matching. To obtain an accurate description of local geometric mo
oth fast nearest neighbor search and real-time dynamic updates, such as point insertion, deletion, and on-tree down-sampling.
t traditional filters, as well as the deep learning paradigm, in order to suppress the aforementioned depth sensors error while pr
acle prediction. In this article, we propose our system GroundGrid which relies on 2D elevation maps to solve the terrain estima
age both the general Morphable Model on common objects and a multi-view synthesis-based shape-from-silhouette model to re
ormulated through Riemannian Motion Policies (RMP). Leveraging multi-resolution obstacle avoidance policies, the proposed na
hods to decline. In order to evaluate the efficacy of tactile sensing in this context, three algorithms were developed and tested w
mpling noises. To quickly find paths to avoid large and/or newly detected obstacles, the sampling noises should be set to large
citly accounts for rotational forces for seamless maneuvers based on the tool’s geometry and featured points. Furthermore,
ence within a predefined finite time that is specified by the user. To further compute the optimal convergent time and the local g
in robotics, leading to a high learning cost for the operator. In this work, we introduce a novel framework using mixed reality tec
constraints and control policies from visual inputs alone remains a challenging problem. In this paper, we propose the keypoin
robotic soccer games, which is able to operate omnidirectional robots safely and efficiently despite high measurement delays a
he robot in terms of keeping the robot's roll and pitch angles bounded within a specified range, is used to select a safety-aware
mization, with an underactuated hand exoskeleton (U-HEx) used as a case study. We propose improving the performance and
we design a portable rigid ankle exoskeleton to use as a research platform for investigating the effect of assistance on post-stro
context of shoulder motion in upper extremity exoskeletons for rehabilitation. Our primary goal is maintaining interaction forces
s limited mobility or is potentially injured, unintentional forces may occur during the robot's trajectory that could be incompatible
nts are enforced. In the swing phase, an optimization problem is solved to achieve minimum jerk planning while respecting a se
nto real-world settings. One dominant challenge is the sample efficiency of RL. This study presents a practical RL setup to cont
, prevent muscle fatigue and encourage the patient’s active participation in order to accelerate recovery. In this paper, we p
r to use in homes and clinics. We present Stretch with Stretch (SWS), a novel robotic system for leading stretching exercise ga
ize goal attainment without considering safety, while others excessively focus on safety at the expense of training efficiency. In
pace of a modified Markov Decision Process (MDP). The RL agent produces a sequence of actions that are transformed into sa
have addressed this issue, they frequently give rise to new complications, including computational burdens and diminished stoc
ning for multi-agent systems with CBF. We show how emergent behaviour can be profoundly influenced by the CBF configurati
cept of a Contrastive Initial State Buffer, which strategically selects states from past experiences and uses them to initialize the
y, simultaneously. The optimality is achieved by the environment reward function that is subsequently shaped using a safety cr
uge training datasets and large experience replay buffers. This poses a challenge for the next generation of field robots that will
n this work, we introduce our new architecture, Projected Task-Specific Layers (PTSL), that leverages a common policy with de
uce an end-to-end framework named Bi²Lane which models temporal dependency in a continuous sequence. It recurrently ut
ety, automotive Augmented Reality displays, e.g.,head-up displays, have garnered attention and are gradually being deployed.
 such cases, most samples taken from the ambient action space generate low-value trajectories that hardly contribute to policy
r, the issue of privacy leakage in embodied AI tasks, particularly concerning reinforcement learning algorithms, has not received
c platform for multi-agent exploration that integrates a broad range of state-of-the-art MARL algorithms and representative scen
m the inaccurate simulator cannot be directly used in offline RL due to the well-known exploration-exploitation dilemma and the
FORMA can perform a real-time online adaptation to these disturbances and generate appropriate velocity actions as counterm
trols of robot tasks. ReactHD employs hypervectors to encode sensory inputs and learn the suitable high-dimensional pattern fo
me-consuming data collection for the state estimation. In this paper, we propose a two-stage manipulation approach consisting
ng possibilities on LTL instructions, confusing action decisions. Moreover, the queries to the uncertain event detector, necessar
Attention and Interactive Comprehension), a novel framework designed to facilitate the learning of unified multi-sensory object p
h as having identical, simple geometry objects in the environment. This paper presents a neural object retrieval framework that
representation for novel view synthesis can only handle simple forward-facing unbounded scenes (e.g., the scenes in the LLFF
this, we present a novel neural implicit swept volume model to continuously represent arbitrary motions parameterized by their
prior methods have investigated registering multiple neural fields by directly utilising these continuous implicit representations.
However, the previous auto-encoders merely focus on the global shapes and do not distinguish the local and global geometric
computer vision, we use segmentations from a pre-trained model to stably locate across timesteps, various entities in the scene
across different objects and supporting items, and (2) identifying a robust action sequence to accommodate the shape variation
s inducing challenges regarding the interpretability and transferability of the learned models. In this paper, we propose a transfo
mages and robot interactions. Our SlotTransport architecture is based on slot attention for unsupervised object discovery and u
arrive at three insights: 1) the performance trends of PVRs in the simulation are generally indicative of their trends in the real w
ures that best correspond to robot motion as estimated by an external signal. Specifically, we utilize the robot’s transformat
xperiments on both model-free and model-based reinforcement learners, we introduce a novel Point Cloud World Model (PCWM
based on either neural networks or voxel grids. Neural networks are capable of representing the SDF continuously. However, th
From Transformers, abbreviated as SOFT, a wrapper around pre-trained vision transformer (PVT) models that bridges this gap

fficiency or loss of resolution. To address these issues, we propose a multi-resolution planar region extraction strategy in this p
ucturally breaking the problem into two steps: rendering novel views via inpainting and 2D to 3D scene lifting. Specifically, we le
but also explicitly models the geometric similarities between observations and the model prior. Our key insight is that the corre
ork includes a deformable propagation module for generating a semi-dense hint map and a confidence map by propagating sp
ints during matching, improving correspondence accuracy. Additionally, we introduce a novel pose block that combines a GRU
by focusing on aerial detection, UAV-type classification, and trajectory estimation—a perspective often overlooked but laden w
nce segmentation (ZSIS) from RGB-D data to identify unseen objects in a semantic category-agnostic manner. We introduce a
network training framework. In this work, we propose the Mean Shift Mask Transformer (MSMFormer), a new transformer arch
ioned pipeline fueled by Depth Anything. To associate the 2D image features with the 3D scene volume, we introduce Gaussia
t present a progressive curriculum learning scheme with three simple-to-complex curricula to gradually adapt the model from cl
-making method that efficiently and effectively integrates collaborators' representations to control the ego vehicle in accident-pr
d 3D registration where one wants to reconstruct a single pose, e.g., the motion of the sensor picturing a static scene. Moreove
am tasks like tracking and manipulation, this paper proposes a 3D DLO perception pipeline to first segment a DLO in 2D image
distance estimates, which are commonly used in several robotics applications. Our methods utilize the transient histogram dire
rained visual concepts that can be transferred to various downstream tasks. However, effectively integrating VLM into the dom
tors, which do not scale well in larger environments, nor do they contain semantic spatial relationships between entities in the e
e tracking failure: tip occlusion, mid-section occlusion, and self-occlusion. This is achieved through the application of Motion Co
ong-tail distribution of classes. Common methods to reduce class imbalance during training include data re-sampling, loss re-we
movements and joint radius. This study introduces a solution to these problems through a torsion-based compliant joint (T-C join
panel, which makes it a compelling strategy for improving solar systems. Unfortunately, while solar trackers are commonly used
policy gradients for designing freeform robots with arbitrary external and internal structure. This is achieved through actions tha
ieve coiling around and holding the objects with-out energy consumption. Two air pouches were designed as the actuator to co
the determinant of the robot's Jacobian matrix, and provide an in-depth understanding of the robot's singularities, which is esse
of actuation amplitude, frequency, and phase shift were determined experimentally to determine the optimal parameters that m
e to lower drag and improve agility. It has a maximum recorded speed of 2.1 BL/s (body lengths per second) and a minimum co
e fully avoided by utilizing kinematic redundancy. A gripper mechanism is developed to increase the orientational workspace by
oblems. Therefore, this research letter proposes a technical solution for the aforementioned problems. The main contribution o
ng methods train a separate model for every application, every robot, and even every environment. Can we instead train ``gene
exity (including massive billion-parameter vision-language-action models or VLAs), into their efficient linear-attention counterpa
s, fingernail grasping leveraging rolling/sliding contact between the object and fingernail, and fingertip grasping with two soft fin
pretraining becomes increasingly crucial in the low-data regimes common in robotics applications. In this paper, we address thi
n leverage large-scale human videos to learn how a human would accomplish a desired task (a human `plan'), followed by `tran
sents a challenge to sampling-based motion planners. We leverage an analytic solution to the inverse kinematics problem to pa
etworks, leading to the notion of neural CBFs (NCBFs). Current NCBFs, however, are trained and deployed in benign environme
ble risk-aware control synthesis, leveraging all information provided by state estimators. This allows robots to stay in predefined
on. In this paper we develop an algorithm to obtain a control-invariant set that allows a given system to move from any state in t
compute the Jacobian of the scaling factor. In this paper, we propose a framework that uses this scaling factor, with an offset,
h them. Many recent works omit explicit modeling in favor of learning-based methods that may yield control policies directly. Ho
eal-time that the system avoids dangerous regions of the state space. Specifically, we employ an optimization-based moving ho
standardizable with hand-pushing, foot-kicking, rope-dragging, stick-poking, and ball-swinging. This paper aims to bridge the g
es and compromise system safety. In this work, we introduce a run-time anomaly monitor to detect and mitigate such closed-lo
g a state-conditioned generative model to represent the distribution of errors between the nominal dynamics and the actual sys
set that contains the ground-truth value that is being estimated by the perception module, with high probability. We apply the pro
mics of each robot to ensure collision avoidance with previously unknown obstacles and (ii) ensures that the robots reach their e
ning policies) safe, is the so-called control-invariant set (a.k.a. safe set). Unfortunately, for nonlinear systems, such a set canno
em is extremely challenging in general. Several numerical algorithms for approximating it have been proposed in the literature, b
ent a closed-loop perception-action pipeline that bridges this gap. Our system encodes an online-constructed dense map, along
al to improve the safety of collaborative robots without reducing their force and speed capabilities. The main contribution of this
s state. In this work, we define the belief state and belief dynamics for continuous-discrete PFs and construct safe sets in the un
a safe backup policy at run-time. The decisions produced by our algorithms are safe in the sense that they come with provable s
fety and input constraints, textit{backup control barrier functions} (BCBFs) construct implicitly defined control invariant sets via a
N, a novel equivariant graph neural network for 9DoF point cloud registration. HEGN utilizes equivariance to rotation, translatio
entation, a classification network predicts a categorical distribution over semantic labels, while a normalizing flow estimates the
mbedded systems when imposing limited word length. To overcome these challenges, we develop an ultrafast and numerically-
o-Sequence tracking paradigm and a tracker named SeqTrack3D to capture target motion across continuous frames. Unlike pre
dule and maintains a global consistent map by representing the scene using multiple neural implicit fields and performing a quic
utationally intensive, and cannot readily be applied to new tasks. To address these shortcomings, we propose Universal Visual
on. To evaluate the gripper, we developed a physical proxy that approximates a realistic apple-stem-branch dynamic system. W
 ruits of different weights after being harvested. Estimating strawberry weight through visual technology can save time and labor
onsuming and error-prone. This paper presents a machine vision-based customization of an automated system for grain appea
ork starts by presenting an accurately registered image time series, captured up to twice a week, by an unmanned aerial vehicl
d light which produces results better than state-of-the-art techniques for vertical shoot position (VSP) vines and demonstrate ini
labor. To automate this process an important step is the precise distinction between separate leaves, which is the problem we
ps under contact interaction. Our proposed method involves encoding the state of a spring-damper modeled tree crop as a grap
tions: harvesting and transportation. An experimental validation is conducted in agriculture-like settings by using multiple simult
earning to interpret these measurements, but require a substantial amount of annotated data to work well. Obtaining such labe
mining SWP. In response to these challenges, our proposed robotic system aims to automate time-consuming and difficult to pe
ator. To solve the problem, the dual-camera setup is designed and implemented on a self-built robot. The perception of the tom
mentation remains a challenging task since large amounts of labeled data are required to handle variations in fruit size, shape,
ion of different plant organs over a series of spatio-temporal data. The approach is based on two stages. In the first stage we d
n-field plant scanning. We create digital twins of the plants through 3D reconstruction. This allows the estimation of phenotypic
atic structure. This task incurs a significant jump in the interaction force on the end-effector so that the analysis using the conce
Unmanned Ground Vehicle (UGV) guides the UAV using the novel template of visibility-aware cooperative navigation (VACNA)
igation using a visual locking control barrier function (VCBF) and a parametrizable switching descending CBF (DCBF) respectiv
ons have shown great promise in indoor well-structured environments, their use in outdoor structured environments such as ag
ore, accessing specific information based on a combination of spatial, temporal and semantic information is generally not suppo
nly includes end-effector design, visual perception, and grasping pose control. The end-effector adopts electrical control, mainl
, the path of transplanter is generated by using 3D point cloud data, and the transplanting part follows it and plant seedlings of
ue of the soil in real-time, using a live video stream from a webcam with an average of 7 FPS. The method is suitable to be imp
complex, highly unstructured environments. In this article, we introduce a system for pruning modern, planar orchard architectu
domain-specific data (normalized difference vegetation index [NDVI] acquired via a red-green-near-infrared [RGN] camera) and
ntically labeled regions of interest, along with an attention-guided information gain mechanism to more strategically select viewp
obot has to plan an efficient set of camera viewpoints to overcome occlusion and improve the quality of perception. Our formula
transplantation of saplings and the harvesting of grown plants. Our method uses a single demonstration from a farmer to extra
njected Noise Discriminator (INoD) which exploits principles of feature replacement and dataset discrimination for self-supervise
procedure is labor and cost-intensive. Therefore, we propose an approach to improve the performance of machine vision syste
the efficiencies of timber-harvesting processes. To this end, we demonstrate the full autonomy pipeline for the log loading ope
od are adopted for the occupancy mapping and feature mapping. To bound out the available working region, we define the inte
cally producing a forest inventory by exploiting LiDAR data collected by a mobile platform. To construct an inventory, we first ex
g the ability to leverage this knowledge for improved planning efficiency. To this end, this work combines transfer learning and a
e robots to encircle the herd by tracking the outmost animals and maintaining a safe distance with other neighboring robots. On
d hydraulic circuit, then present preliminary experiments on pressure-based torque control and docking control using proximal m
the flatness of the finished wall. The proposed simulator approximates the real-world plaster-trowel interaction adequately while
tors with a functional voxel, which routes electrical power and signals. Robotic modules then assemble into reconfigurable robo
onths. Our system combines multi-mission LiDAR SLAM, volumetric differencing, object instance description, and corresponde
problem, we propose a layer-based clustering approach that completes the manipulation tasks one cluster at a time. We employ
n-annotated data. We investigated the feasibility of this approach using a full-size 24-tonnes wheel loader in the real world and d
ffer, the optimal plans for efficiently excavating the target area will significantly differ. In existing meta-learning methods, the do
tructure's condition. In addition, manual NDE data collection can be labor-intensive for on-site engineers. This paper presents a
a large language model, which infers action language from natural language instruction and environment state. The robot then z
ely, foundation models possess an extensive repository of real-world knowledge, including objects we encounter in our daily live
age Models (LLMs), but they do so at the level of high-level tasks and/or require many training examples. Our framework levera
can be easily compared to Meta-World results, allowing for a point of comparison between recent methods using Large Langua
ng information from the environment, updating state estimates, and generating actions based on the updated state estimates. In
convert a natural language task into a linear temporal logic (LTL) automaton. Our main contribution is to enable optimal hierarch
planning. CAPE improves the quality of generated plans through few-shot reasoning on action preconditions. Our approach en
ng the generalization to novel concepts out of the pre-defined sets. To address this issue, we propose GraspGPT, a large langu
ully automated fashion. With this CI/CD architecture, this paper mainly introduces three essential components that uniquely diff
rvoing error. Unlike other sliding mode methods, the proposed approach diminishes the complexity of the system due to the red
ated along the UAV’s vertical body axis instead of the center of mass. This strategy allows: (i) to avoid the two-subsystems
law to actively identify the parameters of the plant. By means of the Lyapunov analysis, the asymptotic convergence of tracking
d systems when engaging in complex tasks. The intuition to address this issue in this paper is that by predicting the successor s
udy demonstrating 6-DOF trajectory tracking. The NS-ATTC algorithm does not require acceleration instrumentation and provid
e to factors such as tire wear and tear. Current approaches to this problem typically either propose offline model identification m
ive Extended Kalman Filter that can adapt its estimation to the control objective, hence allowing safe output-feedback Model Pr
nd the convergent time by GTSMC is qualitatively evaluated by the sliding surface. Then the integral/derivative function of the l
we highlight the non-robustness of the closed-loop system against non-modeled dynamics, such as those relative to joint-dynam
``safe-by-design'' digital twin that operates alongside the real twin robot in the loop, engaging real-time safety framework durin
to constrained regions even in the presence of perturbations or human intervention. Moreover, existing DS research often assu
pect to ISO standards. However, true optimality has not been achieved due to inaccurate distance computation resulting from co
ng to discomfort and safety concerns. This study aims to develop a model that can determine a pose for the robot that ensures
at clearly conveys their intent to reduce ambiguity in how they intend to avoid others. Toward building robots that can safely an
inference and planning-based human motion prediction. First, we employ Long Short Term Memory Networks (LSTM) based in
rate, detection quality, and latency. Furthermore, we introduce two novel perception metrics for Human-Robot Collaboration sa
such as kinematic limits, self-collisions, external collisions and singularities and is passive only when feasible. This is achieved b
ach to achieve this, but its effectiveness heavily relies on the quality of the training data, which can be expensive to acquire. In
significant technical hurdles to bridge the gap between analogies with terrestrial interaction strategies and those that are possib
using on non-contact inspection (visual, electro-magnetic, etc.), the new tasks at hand involved reaching adjacent conductors to
mobile base oscillation increases the risk of resonance. In this study, a disturbance rejection was realized for the Dragon Firefig
ranging from victim localization to hazardous area exploration. However, a critical limitation in the deployment of cellular-enab
-graph and ii) the global graph layer respectively. The local sub-graphs encode local volumetric gain locations as frontiers, bas
er the computational time in the cost constraints. Due to the submdularity, the greedy algorithm achieves $frac{1}{2}(1-frac{1}{e}
ng constraints. The experiment results show that the proposed approach outperforms state-of-the-art methods in multi-robot se
n routing. The proposed method decomposes the given area into Open space and Corridor space where specific deployment pa
ns. The exploration consists of a methodology that extends frontier- and sampling-based exploration with a path-following vecto
variables as well as perceptual inference tasks to solve. Intuitively, the system is closed under the injection, update and depen
egion mapping via embodied navigation in indoor environments, generating a high-level representation of the knowledge of the
unding pipeline. LLM-Grounder utilizes an LLM to decompose complex natural language queries into semantic constituents an
ity from images that utilizes only self-supervision and no manual labels, enabling it to easily learn traversability in new circumst
popular cylindrical and the neglected spherical voxelizations aim to address the problem of imbalanced point distribution by inc
ompletely unlabeled images. In this work, we propose to leverage such task-agnostic image features to enable few-shot panopt
and restores edge texture. To address the issue of reduced texture among objects, we propose a dual closed-loop bipartite ma
nt modal data. To overcome these challenges, we propose a Deformable Multimodal Representation Fusion (DefFusion) frame
c changes and selectively updating these regions of the environment, avoiding the need to exhaustively remap. Human users c
ke advantage of the Universal Scene Description (USD) format which is an emerging standard for the authoring, visualization an
erform latent feature re-projection from previous frames based on differentiable rendering. Fusing re-projected feature maps fro
s of an environment. The proposed exploration strategy leverages metric information of the frontier, such as distance and angle
vision-language foundation model (VLFM) for open-set semantic comprehension and employs the Truncated Signed Distance F
of sparse and irregular sequences of 3D point clouds into a single joint model. Our model directly predicts semantic instances a
r extends this information with temporarily consistent instance IDs to identify the different instances in the scans consistently ov
mory to achieve flexible navigation. However, due to the oblivion and underuse of the scene priors, these methods are plagued
ferred to as blind nodes) that are relevant to a robotic mission. We propose the method of Computation of Expectation based o
dom Field (GGDRF). Our method applies Dirichlet Forest priors from natural language processing (NLP) to the vision domain a
edictions of collision-free spaces. Such feature-fusion approaches become infeasible in scenarios, where the sensor for 3-D inf
essing these limitations, this study introduces a novel meta-learning-based few-shot segmentation approach to robustly segme
bjects. In this paper, we address the problem of moving instance segmentation in radar point clouds to enhance scene interpre
outputs. Several methods to improve uncertainty calibration at different stages in the fusion pipeline are presented and compare
arios and long-range detection tasks. In this paper, we explore a new perspective that boosts HD map construction through the
earn complementary and meaningful representations for each modality. This paper proposes 1) a complementary random mas
Graphs (CURB-SG) that enable higher-order reasoning and efficient querying for many functions of automated driving. CURB-S
rception in the pre-defined BEV range, while simultaneously improving the completion of LiDAR's sparsity in the entire BEV spa
pping-based and learning-based navigation methods. In this work, we present AGRNav, a novel framework designed to search
nts mmEMP, a supervised learning approach that enhances radar point clouds using a low-cost camera and an inertial measur
ning tricks, we devote attention to the impact of sensor configurations on the performance of learning-based methods. To achie
res that rely predominantly on the RGB modality, resulting in limited robustness for all-day detection. In this paper, we propose
hile 3D-to-2D fusion methods cannot explicitly make full use of the 2D information. Therefore, we propose a Bidirectional Fusion
ounter limitations when faced with scenarios involving missing modalities, especially during inference when only IR images are
y of the LiDAR point cloud. In this work, we propose to use an imagine-and-locate process, called UYI. The objective of this mo
ment to realistic autonomous driving applications. In this paper, we focus on the task of building the HD maps in both short ran
eme is established for asynchronous multi-sensor fusion and online calibration. We conduct a dedicated validation of the quate
of formulations, error models, satellite constellations, signal frequencies, and service types, which lead to different precision, ro
which may affect the structure integrity of the extracted moving object. We discover that the moving object has a complete colum
e AYDIV, a novel framework integrating a tri-phase alignment process specifically designed to enhance long-distance detection
st monocular depth prediction with the dense metric scale induced from sparse and noisy Radar data. We propose a Radar-Cam
ates the availability of appropriate data. Although extensive research has been conducted on datasets for outdoor environments
pose a novel deep neural network approach called FusionRAFT, which enables early-stage information exchange between sen
the projected intensities at each point are large in the presence of errors in the estimated motion or calibration parameters. Co
eature fusion. However, event streams exhibit spatial sparsity, leaving some areas unperceived, especially in regions with marg
iltering approaches, post-processing procedures, and safety protocols. In particular, herein an iterative extended Kalman filter a
bility and tumbles. These extreme motions can saturate sensor measurements, especially gyroscopes, which are the first sens
Transformer Tracking model (PVTrack) is presented for robots. It is the first multi-modal 3D human tracking work that incorpor
g event- and depth data for gesture recognition. The network is simulated using the open-source neuromorphic computing frame
ction techniques that expand conventional contact impedance control using proximity sensors have been examined. However,
rol and verified in simulation on quadrotors and in hardware on quadrotors and the TurtleBot3. This technique enables safe nav
pseudo-inverse for the interaction matrix involved. This further enables us to suggest a new control law producing a decoupled
ontinuously captured to explore the scene under varying illumination sufficiently. An attribute control method is designed to adju
nteraction contacts. Moreover, the sensor is produced using the multi-material additive manufacturing technique, which simplifie
uses on microphone pair training, while the second step is designed for array geometry-aware training. The first training step en
ech recognizer and a sound source localizer. To build distant speech recognizer, far-field speech data is collected at 1, 3, and 5
obots in the realm of cross-domain text replication. To address these challenges, we propose CalliRewrite: a coarse-to-fine app
f the fruit. This is typically done by template matching using traditional optimization algorithms, which are slow but accurate, or
localization algorithm that is run onboard the Perseverance Mars rover. Our approach matches rover images to orbital maps u
ent learning and 3D volumetric representations to enable robust and versatile locomotion in confined and unstructured environm
e by prompting each model. Then, by analyzing the performance of the developed system, we found three types of failure in m
ately recognize the states of the infant and the quilt over her, which involves addressing the interferences resulted from an infa
tness. Recent learning-based methods achieve moderate precision and large convergence basin on specific scenes but face is
mechanism and a control-oriented guidance. It bridges the disturbance adaption of classical VS methods and the large conver
of medical education, especially in rural areas and during pandemics. This work introduces ARIS, a multitasking medical service
t of inertia and high dexterity. This paper proposes a original cable-driven parallel continuum manipulator with a spherical paral
es to enhance pose estimation and developed new marker patterns and measurement algorithms to maintain recognition distan
of vision methods to solve individual perception tasks remains a challenge, as the combination depends on the requirements of
ould improve its productivity in terms of area coverage. This paper proposes a novel system that can be used for mass customi
ess interplay between multiple cognitive levels, encompassing reasoning, planning, and motion generation. At its core lies a nov
resent implementation details. With proven global convergence, effective warm-start, the ability to take large time steps, and sp
ssive navigation, or learning from human trajectory demonstrations, which can be challenging to generalize effectively. In this p
gh these simulators improve efficiency, they often lack the flexibility to generate diverse learning scenarios as often needed in m
as their interactions via constraints, contact or friction. Most state-of-the-art physics engines commonly used in robotics either c
tational Fluid Dynamics (CFD) to investigate the implications of using real fish kinematics in hydrodynamic interactions among s
rse Lunar Robotic-Sim (OmniLRS) that is a photorealistic Lunar simulator based on Nvidia's robotic simulator. This simulation p
trol, a framework for controllable traffic scene generation. To capture the complexity of real traffic, SceneControl learns an expr
pt the backtracking strategy to prevent intersection between bodies with complex geometry shapes. We derive the gradient calc
ystems like highly dynamic and holonomic mobile robots presents additional complexities in simulation. This paper addresses th
ully convolutional network with a probabilistic detection loss. In particular, this self-supervised detection loss fully adapts the ke
dometry systems, and datasets with a single camera often lack other sensors, making them unsuitable for sensor fusion approa
transformation between two sonar frames by minimizing the aggregated sonar intensity errors of points with high intensity gradi
L). Termed CURL-MAP (Mapping and Positioning), the proposed approach can not only reconstruct 3D maps with a continuou
-Inertial fusion for odometry is proposed, enabling the mitigation of the effects of LiDAR degeneracy by leveraging a complemen
to represent an arbitrary multi-camera system as a single imaging device. Additionally, it takes advantage of the overlapping fi
nals within a factor graph to achieve accurate and reliable localization both indoors and outdoors. To ensure successful initializa
HERO-SLAM, a Hybrid Enhanced Robust Optimization method for neural SLAM, which combines the benefits of neural implici
combines feature points and neural radiance fields. ONeK-SLAM uses the joint information at the object level to improve localiz
we propose a new two-step sparsification pipeline. Given a pose node to remove, the first step is performed in the Markov blan
 pproaches for loop closures using 3D LiDAR sensors rely on individual scans, we propose the usage of local maps generated f
sistance of kdtree or volumetric maps. To achieve fast normal estimation, we look into the structure of LiDAR scan and propose
nd one is advantageous for 2D radars with a low number of samples, and particularly useful for scenarios where the sensor is b
s on robust optimization aim to mitigate the presence of outlier measurements, robust solutions capable of handling large numb
es. The primary focus of this paper is to address the limitations of deep feature matching (DFM), a state-of-the-art (SoTA) plug
sults than feature-based visual SLAM systems, as the generic coarse object models, such as cuboids or ellipsoids, are less acc
g self-supervised SVO lacks an effective solution to address this obstacle, for either overlooking this issue or exploiting a too he
modate fixed data points and is able to circumvent the occurrence of zeros for rational quadratic interpolants. We demonstrate
ious direct pipeline Event-based Stereo Visual Odometry in terms of accuracy and efficiency. To speed up the mapping operat
mation between them. Specifically, our S-Graphs+ is a novel four-layered factor graph that includes: (1) a keyframes layer with
) It provides a systematic introduction for newcomers to the field, covering topics such as the formulation of the VPR problem,
ception in adverse conditions. Despite its potential, challenges remain for indoor settings where noisy radar data does not pres
semantic descriptors enrich 3D map points, enhancing loop closure detection by providing deeper insights into the underlying m
photometric factor and re-projection factor into the end-to-end differentiable structure through multi-factor data association mod
global point cloud, whose scattering is then iteratively reduced. This is achieved by dividing the global point cloud into uniform g
obust LiDAR odometry (LO), which utilizes continuous-time estimation to remove LiDAR distortion and builds the spot uncertain
ly estimating the trajectory of the sensor. To achieve real-time performance, this paper proposes a novel information-theoretic l
ffective, LiDAR-inertial odometry systems require proper parameter tuning to be deployed. In this paper, we propose LIO-EKF,
y system (Multi-LIO) that outperforms existing state-of-the-art solutions in accuracy and scalability. Utilizing a novel parallel stra
acy while providing additional estimation of object motions. State-of-the-art literature informs two main formulations for Dynamic
of coplanar features (kid planes), these kid planes' 3DOF estimation can be regarded as the measurements with covariance o
raw point to the respective geometry feature. A nice property of this formulation is that the geometry features can be analytica
ance. To address this challenge, we propose DORF (Dynamic Object Removal Framework), a novel coarse-to-fine offline frame
or pose online by registering LiDAR scans to maps, and dynamically grows the map. Then, our meshing module takes the regi
or maximize simple heuristics like field-of-view (FOV) coverage to decide where to place exteroceptive sensors. In this work, w
a single system. By integrating hand-crafted and learnable features, we seek to capitalize on their complementary attributes in
a shape alignment between the 2D segmentation of the object and the projection of the 3D object in the image plane. To avoid
e lost if measurements are removed without any principle. To remove measurements which is redundant, we propose a trade-o
onvergence of the non-linear system. We propose a gradient-based method to estimate well-conditioned covariance matrices by
presents ultra-wideband (UWB) radar as an alternative to the existing sensors. UWB is generally known to be used in anchor-t
We conduct experiments on both public datasets and our self-collected data, with different mobile platforms and sensor types.
due to the limitations of the RGB-D camera. We detect walls beyond the range of RGB-D cameras using four-point LiDAR; sub
extraction method, we employ an incremental BoW model that is updated regularly. This enables a coarse-to-fine candidate se
rge Language Models (LLMs) to create a unified approach to scene understanding and place recognition. The approach first bu
which primarily relies on camera-camera edges, our approach centers on the introduction of a dynamic element - any rigid obj
e scalability to large-scale environments. A novel and correspondence-free, inherently probabilistic, formulation of LiDAR residu
ad-mounted MR headset presents unique challenges due to viewpoint changes. This work investigates how active visual locali
the paper, we propose to 1) incorporate frontier-based exploration tasks for global coverage with implicit surface uncertainty-ba
s occupancy map quality. We calculate a probabilistic reward of getting a loop closure at any pose on a pose graph, which con

ed in visual odometry (VO) pipelines by approximating temporally close measurements as occurring at one common time. This
nd camera intrinsic and extrinsic calibration states. In contrast to state-of-the-art invariant extended Kalman filter (IEKF) approac
od that enables multi-frame tracking of the same lane line, thereby obtaining multi-frame data of a lane line. Then, we utilize mu
so challenging to identify potential vulnerabilities of a navigation policy, i.e., situations that the robot may cause a collision, in va
al world has been challenging. To reason in the real world, robots must go beyond passively querying LLMs and actively gathe
sted in a lab experiment with 31 participants. The experiment showed significant differences in trust ratings and decision-makin
ation attributes. Our visual and textual explanations of a robot's navigation are influenced by the robot's personality. Moreover,
understand the complex, dynamic, and sophisticated interactions of the crowd resulting in proactive and foresighted behaviors
ed with human participants interacting with a Pepper robot equipped with visual capabilities. The results affirm the viability of the
arrations delivered by either a human or a robot. Findings indicated that participants displayed longer response latency interac
tween groups of schoolchildren from different countries. First, we conducted a participatory design activity with an interdisciplin
large language models (LLMs), which possesses a significant corpus of information obtained from human society, and exhibits
pporting structure, featuring a flexible leaky coaxial cable as the backbone, which allows for continuous tunability and precise co
challenges, efficient motion planning methods such as Rapidly Exploring Random Trees (RRT) and its variant, RRT*, have been
igns which are only suitable for objects with certain speed ranges once the grippers are fabricated, the working range of object
ability and compliance. To improve the restorability of the actuator, we used cotton threads to connect the spherical acrylic bea
s to reduce the costs of deep-sea sediment sampling by introducing a novel actuation system for suction samplers, that capitalis
. Our implementation is a tip-clutching winch, into which vine robots can insert themselves and anchor to via powerful overlapp
p reinforcement learning from demonstrations to learn a memory-based agent that acts purely on haptic and proprioceptive sig
ce-position control. For this, a force estimation model is derived to substitute the force sensor at the tool-center-point. The force
grippers have been proposed for harvesting different produce types, however their employment in field testing has been extrem
r, turning, and transitioning motion capabilities. The soft robot employs three pneumatic bending actuators and two adaptive ele
ahedral robot (Tetraflex). Tetraflex is composed of six pneumatically actuated bellows joined at four points by rigid nodes. By e
ns. In this work, we propose an innovative approach to underwater soft manipulator design: 1) by constraining high- power optim
ents limit the design and implementation of a compatible sensing module, which restricts the further development of robotic func
nlocking device enters the gap between the satellite and the solar wing, re-unlocking the solar wing. This solution effectively lev
ovative pupal-mode actuator for NOTES robots. Through the manipulation of its internal air chambers, this actuator is capable o
ot consists of a pneumatic soft actuator capable of multiple modes of bending deformation as the body and four identical multi-s
autonomous robotic solution to harvest fully ripe blackberries with minimal damage. The present paper details the mechanical d
tly limited in its working range and activity mode. This paper presents a small untethered bionic manta ray. The robotic fish is dr
flexible links in an existing mechanism, and optimize their rest shape. We demonstrate the efficacy of our approach on a set of
on via antenna edges, but requires non-metal actuation, large antenna rotations, and high antenna angular accuracy for long p
r, this design, along with the development of a new low-cost, flexible illumination system, is able to incorporate a high-resolution

to achieve adaptive grasping on surface asperities with a smart shape-memory silicone sponge. Heating above 60â—¦C make
an origami design software to generate desired crease patterns based on analytical models and Evolution Strategies (ES). Add
. The objective of this work is to achieve more flexible manipulation capabilities of the robot. To this end, we utilize a redundant
lthough previous research has explored the use of DEAs in soft robot locomotion, achieving optimal behavior is challenging due
ng and elongation finite deformation taking into account the geometric orthotropy and nonlinear elasticity. We experimentally va
otation, and elongation). Furthermore, we develop kinematic and stiffness models based on the constant curvature assumption
has 15 pneumatic actuators based on a series of hollow chambers that are driven by an external pressure system. The thumb
wearable soft hand extension glove based on the X-pouch and strap system, addressing the above challenges. The glove enabl
 to unfold, resulting in the twist of the actuator to a helical structure. The SFTTA has great scalability. It can be fabricated throug
rdware at the distal part, compromising the inherent benefits of soft growing robots. Existing soft retraction mechanisms that m
at can navigate cluttered environments by extending from their tip. Yet, these vine robots are currently limited in their ability to
ss irrespective of its position. Both actuation and stiffening mechanisms are independently controlled by pressure. Design, exp
d experimental testing of a modular, tendon driven bead-style continuum manipulator with tunable stiffness. By embedding the
ased on MEMS barometers suitable for tactile localization on UAVs, featuring lightweight, low stiffness, high sensitivity, a broad
ttices that are sensorized by embedding air channels within the structure. The lattices are 3D printed from a single build materia
epresentation of the actuator genotype. This makes it possible to represent a wide variety of structures and to take into account
r to solve these problems, this paper refers to the structure of the human hand and designs an articulated rigid-soft hybrid robo
s encountered in nature and translate them to the proof-of concept design and development of a soft pneumatic actuator (SPA
re incorporates rigid retainer pieces within each layer, which allows for sliding within a hollow cuboid structure and enables mod
integrated within a soft robotic retractor that interacts with brain tissue to generate a surgical workspace upon actuation. We de
The robot is able to bend and sense its own shape up to a curvature of 44.7 m-1, corresponding to a bending angle of 102â—¦,
he use of prosthetic scenarios. In this study, a novel optical waveguide sensor was used to collect muscle deformation informat
multiple electrical connections onto the resistive layer of the sensor. This allows us to measure changes in resistance at multipl
multi-modal sensor (providing both force and contact location) and robotic collection of simple training data of different motion p
degrees of freedom (DOFs). First, a generalized, control-enabling dynamic model is developed for DEAs by taking both nonlin
uctures of the human body. However, there are still challenges in accurately estimating external forces applied to these system
ning the structure, variable stiffness by jamming, and fiber-optic shape sensing while sharing the same structure and materials,
date, vine-inspired robots have yet to show such localized embedded responsiveness. Here we present a vine-inspired robotic d
e outlines a geometry-mechanics-optimization integrated approach to automatically shaping a pneumatic soft actuator or robot
er milestone to such systems would be to increase the controllability and responsiveness of their underlying reactions in order to
hcare solutions. In this work, a pulsating fluidic soft sensor (PFS) capable of determining location, pressure and contact area of
are side, the use of EtherCAT (Ethernet for Control Automation Technology) allows connecting a freely selectable number of C
ecommend to employ similar microservice architectures for the deployment of small- to large-scale robotic applications in order
en the model and the actual code, using which it auto-generates all required integration code. This allows Plug'n Play behavior
ultitude of robots that can be assembled creates a host of additional parameters to optimize. We include tasks such as machine
ind conditions, and gas simulation fidelity. They also lack photo-realistic rendering for the integration of obstacle avoidance. In t
able plans. In this work, we use Linear Temporal Logic (LTL) as a formal language expressive enough to model the temporal na
ions, generalize examples to new tasks, and write code. In this paper, we suggest how these abilities can be harnessed and ap
on) for flexible/adaptive robotic tasks where the actions to be performed may vary in planning and execution time depending on
by tall grass and weeds that can be safely traversed by a robot, but that can be perceived as obstacles by LiDAR sensors, and
ularly in the poorly illuminated polar regions of the Moon -, leading to uncertainty. This work shows how a novel single-image Su
ive Control (MPC) in the zero-gravity laboratory (Zero-G Lab) at the University of Luxembourg. This approach leverages PPOâ
relative location of the debris with respect to the chaser spacecraft. This paper represents and solves this problem as a hierar
ring Rendezvous and Proximity Operations (RPO). The pseudo-labels are provided by an adaptive unscented Kalman filter whe
dology incorporates an off-policy reinforcement learning framework, employing the soft actor-critic technique to enable the gripp
, resulting in a performance drop due to the domain gap. State-of-the-art approaches employ Domain Adaptation techniques to
errestrial and aerial navigation contexts. The new technique specifically targets mapping from stereo scenarios with ultra short b
ing techniques are used to identify known landmarks in images captured by the spacecraft. These landmark locations are then
able sloshing and modal states. This is achieved by two identification experiments. In Exp.1 all unmeasurable states are elimina
 vel standard interface, docking control method, and path planning method for space multi-branch robots. Firstly, a novel standa

a novel iteration of such a robot with a pressurized pneumatic shell design. A soft robot of this type brings benefits of a passive
data transmission. The paper presents the further development of the reliable electro-mechanical interface (EMI) from the Tran
e applying the appropriate forces to activate their adhesion. However, in the case of a free-flying robot in microgravity, physical
e time, the single point contact with the ground cause the contact friction force with the ground is small, so it is hard to climb ob
errors. Therefore, controllers robust to parametric uncertainty and disturbances are needed. A robust controller attractive due to
ol to improve robotic and augmented reality-based surgery.
leading to a need for less computationally intensive methods. We propose a method for autonomous robotic manipulation with
es are estimated from existing actuator feedback, with a deep network for dynamics identification. Performance is further optimi
input modalities and their representations on gesture recognition and prediction performance. We perform an end-to-end asse
eterministic policy output. In this work, exploratory bottlenecks have been used by a human expert as convenient boundary poin
ock, a task that has complex dynamics interactions, small scale objects, and a narrow field of view. We train agents in simulatio
singly better performance at 6D pose estimation tasks; yet, their success depends on the availability of large amounts of annot
otic surgery more autonomous which has the potential to reduce variability of surgical outcomes and reduce complication rates
a sequence of sub-steps and understanding of inherent dynamics of goal-reaching tasks. In this paper, we propose a new learn
assisted bronchoscope with a spring-based extensible segment. By manipulating two driven rods, the segment can be lengthe
ractions between manipulators and the environment. In this work, we estimate the camera-to-base transform and joint angle me
o naturally capture the pixel-wise tissue deformations and adaptively correct the tracked template. Specifically, we first impleme
ery, such as membranes and soft tissues, have complex, anisotropic physical parameters that a simulation with simple initializa
ovide such geometric knowledge. We model the anatomy via a Bayesian Hilbert map-based probabilistic 3D occupancy map. U

y critical for safety monitoring and autonomy, as it is unaffected by common camera issues such as occlusions and lens contam
First, we present an interactive and physically realistic simulation based on the Finite Element Method (FEM) that mimics the t
da Vinci Research Kit (dVRK) and Smart Tissue Autonomous Robot (STAR) which represent common subtasks in surgical tra
pplications. We categorize centerline-based vascular 3D-2D image registration problems as an iterative Perspective-n-Point (P
or confined spaces. Tube interactions, unmodelled phenomena, and inaccurate tube parameter estimation make physical mod
aper, we develop a new, more general CPPR model accounting for any number of tubes, describing their variable-curvature 3D
thod using a fully distributed grating-based sensor. The proposed method uses simple trigonometric and geometric modeling of
es to both the patient and the surgeon. Aiming to reduce the duration of x-ray radiation in interventions using MSRCs, this letter
ValveTech platform prior to valve implantation. First, we present a force observer that transfers forces from the manipulator body
ive 2D ultrasound (US) and preoperative 3D computed tomography (CT) fusion is proposed, addressing the shortcomings of ex
rmation. Our approach uses spatial-predictive control (SPC), to optimize a cost function given strictly in terms of inter-agent dis
s of our approach, we conduct experiments to detect the maxima of the dynamic gradient, which was implemented with a set o
experimentation possibilities with Kilobots, also relieving researchers from demanding tracking and logging activities. However,
a robot platform for morphogenetic engineering in swarm robotics. Its direct current coreless motors provide physical support an
he sensory capabilities of FireAntV3 to enable the robot to sense forces, sense the direction to a light source, and to sense cont
orces on-demand, we can enable a collective to exhibit desired behaviors previously not possible. We use a general form of the
aper, we investigate the design of robot swarms that perform their mission by interacting with other robots that populate their en
n verifying that drone fleets operate correctly and can significantly reduce development time. The simulator interface that we de
ly have access to a container. We propose a modular topological graph-based transport policy (HTP) that explores efficiently w
equire an impractical amount of time to solve relatively small problems. We propose Constraints and Streams for Task and Moti
he modeling of various tasks. We show this formulation by first presenting the CDTS and then its extension that is based aroun
propose Dynamic Logic-Geometric Program (D-LGP), a novel approach integrating Dynamic Tree Search and global optimizat
s are then transformed into a hybrid map with the following functional components: object proposals to avoid missing trolleys du
mpute effort in such hierarchical solution processes. More specifically, we introduce infinite completion trees as a problem forma
mpling point to ensure accurate gas measurements. However, this approach not only prolongs the time required for a single me
stacle avoidance issues due to local minima. In this work, we extend LGP with a sampling-based reachability graph to enable s
problems, which can lead to the solution of the manipulation puzzle. Our method combines a heuristic- driven forward subproble
obot systems through a sequence of transformations; 3) the automated translation of the plan into an executable formalism. Th
approaches in EIF often neglects the knowledge from historical data and also do not effectively utilize information across the m
gration of NARS as a ROS node. A case study is provided demonstrating the system's proficiency to carry out a garbage collec
thm follows several design principles that yield the following features: (1) efficient solution finding due to sequential heuristics a
s, and we solve it as a Mixed Integer Linear Programming (MILP) problem. Our contributions include a general planning framew

RT) is presented. The proposed method is based on a unified Rapidly-exploring Random Tree (RRT) that operates on both the
d to collect measurements to estimate domain-relevant scalar parameters such as soil moisture or nitrates concentrations. In th
nipulation tasks for the mobile platform. By minimizing the expected energy usage in body key state planning while satisfying ta
been designed to be accessible and extensible. Parts are either 3D printed or otherwise constructed from materials that are rea
ise avoiding any edges and corners in contacts. Nevertheless, pointed or edged impactor geometries cannot be completely rule
oding introduced in earlier work. In this work, we contribute a practical approach to planning with strategies by introducing a nov
obot, as the concepts in a FOON are too abstract for execution. We thereby introduce the idea of exploiting object-level knowle
ree, capturing sequential and parallel dependencies among subtasks. The pipeline then mitigates the uncertainty and unreliable
r extremely high computational costs and may even fail. In this paper, the proposed method easily achieves the goals of multi-a
of frontier point detection and task assignment determine the overall efficiency of the system. Most of the existing methods impl
oncept, which relies on a virtual physical interaction interface between the human operator and the remote robot equivalent to a
k-aware confidence weighting strategy for remote learning from demonstration. Our approach extends the Hidden Semi-Marko
manipulators (LiCAS) in leader-follower configuration. The leader dual arm (LDA) captures the movements of the operator's arm
n-making for mode-switching is generally assumed to be done by the operator, which brings an extra DoF to be controlled by th
operators during teleoperation. Additionally, the incorporation of haptic rendering, including kinaesthetic and tactile rendering, c
ccurate force feedback and reduces the number of model updates. Furthermore, the OBM can be deployed with or without pre
The introduced novel method calculates the near-optimal Cartesian end-effector pose in terms of visual feedback quality for the
o ensure safety of workers. However, the conventional command interface is not intuitive for tool manipulation (i.e. pivoting, swe
acilitate their design, we propose a method to learn one of their core components -- active constraints -- from demonstrated en
f any robotic arm using a digital twin as the interface between users and the robotic system. Due to TELESIM's modular design
e achieve full-body synchronization. We assess our humanoid-to-human whole-body synchronized motion model on the HRP-4
object placements. Our proposed method combines simulation-driven physical stability verification via real-to-sim and the sema
ective to see objects obstructed by the robot is impractical (e.g., teleoperation in narrow environment). Diminished reality techni
on and drone piloting applications. Comparative analysis shows no significant difference between task performance with iRoCo
orating open-world shared control into avatar robots to combine the benefits of direct and shared control. This framework prese
at hierarchical levels i.e., low-level actions and high-level tasks, by incorporating multi-scale hierarchical information in neural n
ne Interface (HMI) and a bi-wheeled humanoid, SATYRR, is proposed. Secondly, the study introduces a dynamic locomotion ma
as helical motions, which are more challenging to the users. The proposed framework uses Artificial Intelligence (AI) to predict j
efficient fine-tuning approaches have significant progress in other domains, they generally perform worse for point clouds. To o
anced performance is mainly caused by: 1) self-training with imbalanced data and 2) the lack of pixel-wise 2D supervision signa
his paper, our primary focus lies in domains with different state and action spaces, which has broader practical implications, i.e
h learnable contexts to automate the prompt tuning process. Then, we lock the pre-trained backbone instead of adopting the fu
rios and make the first attempt to solve the Domain Adaption (DA) problem for multi-view 3D object detection. Since BEV perce
ch is the utilization of learned implicit representations of visual and geometric features derived from a pre-trained NeRF. Though
meframes. In this paper, we propose a novel multi-level framework for RL to accelerate convergence as the first attempt to add
become feasible for industrial robots to acquire complex motion skills within simulated environments. Nonetheless, the ”reali
“Yes We Dance!” initiative seeks to make inclusive dance accessible to all, irrespective of age, including individuals with
prings and is designed with a 2-Degrees-of-Freedom(DoF) structure. Furthermore, DARFT-1 is designed to prevent external ob
hrough folding and exhibit various bionic movements, such as snake-like movement and worm-like movement. Based on this m
ine blades, etc. The adsorption module of the climbing robot primarily consists of three active compliance suction units, each o
traightforwardly and needs a complicated motion to enter the groove, transferring with rotating. We can reduce the number of d
dexterous cooking actions. Even actions that require impact mitigation, such as chopping zucchini or tenderizing meat with a m
d one receiver. These multi-device technologies are specific to their relative device positions and cannot generalize well to a ne
ed by AI and democratized telepresence, robots deployed by autonomous systems continually monitor and combat environmen
he difficulty in interacting with heavy industrial equipment and components which often requires the involvement of multiple ind
tations. Depth-aware sensors detect only visible body parts, RGB-based solutions rely on complex pose estimation, and most r
the implementation of three Neural Networks designed to augment the cobot with visual, spatial, and auditory inputs to improv
ation with its multi-fingered DLR-Hands II: it can perform in-hand manipulation -- purely tactile (w/o cameras) and learned from s
e lack of sufficient training examples and computational resources. In addition, the learned models lack transparency, and it is d
ireless signal is often blocked due to the Non-Line-Of-Sight (NLOS) issue. At the same time, recent active localization methods
ground-based PC and the robot, one can control the robot's movement using real-time footage from the camera mounted on th
rwater light and turbidity. In this work, we introduce a novel method for 3D detection and tracking a moving object for marine an
s, especially on fire. In this research, the authors proposed a new method of variable stiffness mechanism using a string of bea
h, so any new forms should require minimal intervention. To provide for both requests, we propose a system that uses a social r
orter hospitalization time, which also provide an alternative solution for patients, who cannot suffer open chest surgery. Howev
obot consisting of a mobile base and 5-degree-of-freedom manipulator was developed. It was used to create a service robot app
framework offers a promising solution. However, it requires tuning of second-order dynamic system parameters. Our contributio
rd locomotion, jumping, turning, and traversing rough terrain. This study explores three distinct turning methodologies. The first
unpredictability of outer space means that pre-planned cooperation is only sometimes feasible. Therefore, these robots must p
the reflexive functionality of the biological human hand, is automating grip force adjustments independent of human input. In o
the Unity model updates its joint values according to the physical counterpart. The robot cell populates a pointcloud, which is f
. Wheeled climbers have low surface adaptability due to their small reachable workspace. Legged climbers tend to be heavy, s
ious application fields, including healthcare, daily activities, and communication. Additionally, we are developing sensitive hand
perelastic nitinol wires with soft materials for performing deep-sea underwater sampling tasks. A manned deep-sea submersibl
ts, six motor-cable pairs for locomotion, nine stretch sensors (six for motor control, three for structural integrity), IMUs, a microc
first device to propose this motion had a structure that generated only straight motion. In this video, two epoch-making devices
m alloy wire. Heat is transmitted by the thermal conductivity mechanism. causing a change in volume and thus altering buoyanc
rovide valuable insights for researchers in the field of future surgical robot advancements.
 conversion time between series and parallel configurations was completed within 20 s in the experimental test (see the video),
dance control, which determines the device’s compliance, can greatly affect overall performance and the physical load exp
nd perceive SRL poses a challenge. We apply a mixed reality device (Hololens2) in a wearable robotic arm system. Taking the
r either on their latent spaces or on the parameters for constructing the graph, we use the action information as well as the emb
sentation and predict brightness increment images with motion interpolation. Results of object tracking show the new methods s
od which boosts the accuracy of NRSfM using sparse surface normals. Surface normal information is available from many sour
is very easy to do, and marker-less state estimation of the robot from this camera's images is an ideal solution: both simple an
be integrated in different appearance-based multi-object tracking methods. We also propose an attention-based visual feature m
ern trackers rely on the appearance of objects rather than motion for association, which can be ineffective when most targets ar
approaches still have some limitations: 1) these approaches either require the UWB anchor position to be calibrated in advance
precise gyroscope bias estimation on rotation accuracy. This, in turn, affects trajectory accuracy due to the accumulation of tran
reconstruction systems which omit features that are critical to robotic path planning, such as computation of the Euclidean Sign
GX) where separate modules for exploration and exploitation come together in a novel and intuitive manner. We configure the
ovel T&R system based on a robust motion control technique for a skid-steering mobile robot using sliding-mode control that eff
plates for stability during training and zero-shot inference. We implement our method on Astro robot and deploy it in both simula
ng environment so that the estimation can be used as a realtime feedback signal for autonomous closed-loop control of mobile
ment modeling, but distinguishes itself by substituting LSTM modules with transformers to greatly elevate the performance our m
orm robot navigation in unseen environments using different coarse maps. To do so, CMN addresses two challenges: (1) nove
ed resolution. Furthermore, the growing adoption of event cameras in robotics captures rapid changes in pixel brightness but fa
as a noticeable gap with more robust methods. In this paper, we develop a novel method to efficiently estimate high-quality opt
ur preliminary findings, most instances of navigation failure in previous models were due to turning or stopping at the wrong pla
g. In this work, we present PlaceNav, subdividing the robot-independent part into navigation-specific and generic computer visio
representation of the scenes, which results in misalignment with visual images. To provide more accurate and coherent scene
o approximate the computationally-intensive POLo score. Our approach addresses critical limitations of both end-to-end reinforc
e knowledge from foundation models to navigation tasks. In this work, we propose a Pixel-guided Navigation skill (PixNav), whic
nvironments. To address this issue we propose GeoAdapt, which introduces a novel auxiliary classification head to generate ps
t terrain types and their affordances. Moreover, these methods fail to identify traversable geometric occurrences, such as stairs
ng and draws inspiration from the navigation strategies employed by human divers, who do not rely on localization. UIVNav con
sting solutions either require labor-intensive manual data re-collection and labeling or use hand-coded reward functions that ma
e image representation learning for joint low-light enhancement and deblurring. Specifically, we employ a Transformer backbon
or generalizable and transferable feature extraction, which is followed by training-free feature enhancement. The best-prototype
of 3D occupancy models. To address this, we present RenderOcc, a novel paradigm for training 3D occupancy models only usi
neural networks to further exploit the potential of neural networks in V2D. Specifically, our neural optimizer alternately updates t
ates the dense flow field encoding the displacement of each pixel in one image to its corresponding pixel in the other image. D
ged as novel backbones for various downstream tasks. We argue that leveraging the pre-training capabilities of MLLMs enable
-modal relationship between point cloud videos and image videos to acquire meaningful feature representations. Intra-modal an
FSNet. FSNet contains four important improvements over existing self-supervised models: (1) a multichannel output represent
Prior efforts to convert APS data into events often grapple with issues such as a considerable domain shift from real events, th
the dynamic nature of the environments. Moreover, joint object detection and pose estimation models are better suited to lever
een thermal and RGB images. In this paper, we present a self-supervised approach to enhance thermal-based depth estimation
ses unseen during training phases. To tackle this problem, we introduce an innovative Open-Set Surgical Activity Recognition (
 rom inefficiencies arising from non-end-to-end processing, reliance on separate models for different object categories, and slow
his paper advocates fusing the complementary information from image frames and event streams to achieve more robust keypo
elegant framework named TiV-ODE, formulating this task as modeling the dynamical system in a continuous space. Specifical
methods have employed traditional fusion techniques or fusion networks to generate fused images that combine thermal and v
image leads to non-trivial misalignment, while directly drawing a cuboid on the image cannot access the original 3D informatio
rmance 'without bells and whistles' using the standard components such as visual transformers, dense prediction transformers
ncurrently, diffusion based conditional image generation models have demonstrated remarkable outcomes, adept at producing
ring. To tackle these challenges, we present Lightning NeRF. It uses an efficient hybrid scene representation that effectively uti
nts to augment NeRF training. The proposed novel pipeline can directly benefit from those priors to reconstruct 3D scenes with
D maps but often come at a higher price and lower resolution. While active illumination sensors like LiDAR provide precise dept
new problem of faithful FOV extrapolation that utilizes a set of pre-captured images as prior knowledge of the scene. To addre
ficult to apply in deep learning frameworks. Second, they often require high computational resources due to complex deep netw
tial enhancements in such models when the denoising process is informed by a better visual representation. We study the scen
nsolved challenges. In this paper, we extend our previous work on keypoints-based visual imitation learning (K-VIL) to bimanua
y blending 3D hash grids and 2D plane grids representations, our approach provides a comprehensive scene representation. W
d data. Recent methods have used prior knowledge or partial reconstruction of the target to accomplish view planning for active
arker integrity, resulting in reduced detection reliability. To address this challenge, we propose a novel fiducial marker called Co
oise and potentially malicious attack, jeopardizing the precision of spatial-temporal alignment. Rather than relying on external h
degraded by water- drops. Our method exploits an attention network to predict the location of waterdrops and then train a Neu

uffer from local minima, making them impractical for real-time robot applications. We aim at solving this problem by marrying im
OP-GCN) to address this challenge using the optimal occluded body parts. The proposed model uses an occluded part detecto
we present Motion-Aware Loss, which leverages the temporal relation among consecutive input frames and a novel distillation
the fact that these methods often rely solely on deep learning techniques or classical approaches. In addition, they neglect ove
ally needs HD maps, 3D object bounding boxes, and camera extrinsic matrixes. In this paper, we present a novel semi-supervi
lti-modal 3D data. In this paper, we introduce OpenAnnotate3D, an open-source open-vocabulary auto-labeling system that can
addresses the dilemma of poor generalizability of previous segmentation models and the disadvantage in requiring to retrain th
tures with well-defined scale and apparent motion within each burst by jointly searching in a scale-slope space. We demonstrat
porally intensive spike streams. Therefore, we introduce open-source RGB data to support spike depth estimation, leveraging i
oding, 2) Occupied Pillar Prediction (OPP), and 3) Pillar to Point Generation (PPG). The input point clouds are encoded using a
e integration. Nevertheless, the camera position of the photogrammetry system is difficult to select to guarantee an optimal field
computational procedures, and exclusive reliance on a single type of morphological information. In this paper, we introduce the
rt VMMR datasets, namely bias, human error, privacy, and the challenge of providing systematic updates. On the other hand, th
ctive and texture-less, making it difficult to identify distinguishable point features to establish 2D-3D correspondences. To addre
bsequent acoustic micromanipulation tasks. Here, we proposed a miniaturized soft acoustic end-effector and demonstrated aco
training datasets are generated automatically in Pybullet simulation environment including 5000 depth images and above 30,00
e all stacked instances in bin picking dataset, learning without real-world datasets has become a major interest. In this paper, w
n model onto a context map. We learn map context information and aggregate temporal sequence information without requiring
learn to be robust only up to a certain degree. At the same time learning-based methods under moderate rotations often outpe
to address new tasks. In surgical scenarios, the privacy issue of patient data often restricts the availability of old data when upd
ntraocular environments and handling minuscule high-precision areas remain enormous challenges. These include issues of un
ori knowledge of the target object, a reference 2D image, or 3D data. By doing so, we can reach targets in unstructured environ
rs and force sensors are used in the underactuated hands, it may cause a problem with hygiene and require complicated calibr
but also enables multitask capability based on a few task-specific demonstrations. We construct demonstration graphs by splitti
model and the robot's dynamics. Consequently, research has explored models incorporating the second-order derivative of fea
scheme achieves improved computational efficiency and extended flight duration while generating new control sequences for t
general form (linear, thin-shell, volumetric) in 3D space. Furthermore, it decouples the runtime complexity of the approach from
mpted Tracking (DCPT) that achieves robust UAV tracking at night by efficiently learning to generate darkness clue prompts. Wi
raining categories. Integral to our approach is a multi-layered tracker integrated with the foundational detector, ensuring continu
bjects are generally small in the image plane, often blurred, and frequently occluded, making them challenging to detect and rec
s. We integrate these rules into the particle filter, resulting in the Object Permanence Filter (OPF). For multi-object scenarios, w
ges in the wild. This raises the question: can we re-purpose these large-scale pre-trained static image models for open-vocabu
k, in this paper, we present a multi-correlation Siamese Transformer network that has multiple stages and carries out feature co
ms, and methods that encourage cycle-consistency in the estimates (i.e., tracking backwards should yield the opposite trajecto
accurate association and state estimation by batch processing many frames of sensor data while being capable of running onli
tainty awareness poses a problem in safety-critical tasks such as autonomous driving where passengers could be put at risk du
nd used in conjunction with a simultaneous localization and mapping software. Our approach allows humanoid robots, targeted
transformer-based tracking model optimized for high-speed operations across various devices. It achieves a more favorable tr
be the robot’s trajectory along the crop rows. We design a driving state quantification algorithm and accurately separate the
tion conditioned on future control inputs. The single-track dynamics model approximates wheeled vehicle motion under specific
ercome these limitations, we propose an adaptive interframe alignment method that allows for an adaptive selection between th
tightly coupled fashion, which thus is termed SRI-GVINS. In particular, for the first time, we deeply fuse the GNSS pseudorange
k based on multi-basis depth representation. The network takes panoramic images (obtained by stitching dual-fisheye images w
arallel processing and are based on the Single Instruction Multiple Data (SIMD) paradigm. In this paper, we present BIT-VIO, th
nformation in human-made scenes, while point-based event and image features complement each other through well-designed
eraging strategic graph partitioning and an optimized approach to Jacobian approximation. By integrating JacobiGPU into ORB
onential function of the automorphism of SE_2(3), which can effectively enhance tracking performance under fast rotational mo
d, and providing swift and safe assistance for people with impaired vision. In perception, we engineered a lightweight multitask
ategy keeps the distance between the grounding point and the CoG close, an upward force can efficiently work for the STS ass
f a supernumerary robotic tail to augment human balance. The actuation, geometric and inertial parameters of the tail are exam
the exoskeleton to the user, and quasi-passive mode for aiding in sitting down (power absorption), where the device absorbs an

ement habits) of a patient, hence guaranteeing effectiveness. To fulfill this requirement, this paper proposes a new motion plan
suming significant power. Some passive exoskeletons don’t require power, but their fixed stiffness constrains normal motio
ditionally achieved in indoor environments using costly force plates. To overcome these limitations, this study proposes an alter
eters exploration process and improving the assistant performance. However, there is no existing method that could utilize the h
discomfort and potential injuries. To reduce misalignment with the knee joint, it is essential for the thigh and shin harnesses of t
t of assistance to complement the human effort needed to perform the walking activity. The proposed controller exhibits adapta
tate estimator. The projection operator guarantees the convergence of the parameters while offering a limited amount of assist

on is designed based on a reduced-order model and is planned using an optimization method with multiple constraints. In additi
. In this paper, the initial estimate derived from IMU integration is refined. Following the computation of velocity and ground con
 e the rolling body and terrain as parametric curves, we derived rolling constraints and dynamic equations of motion for model-b
ing modeling to the planar, nonlinear rigid-body dynamics and subdividing the terrain into contact-phases, a tractable nonlinear
n by a coaxial hub arrangement. The analysis is used to inform the design of a central pattern generator to control the robot by
planned trajectory without rollover and slip are calculated using the passive decomposition of the WMR dynamics with the Pfaf
o modes due to the difference of their configuration spaces. However, the control inputs are redundant for planar motions, which
uation, and therefore our approach does not have singularities of chained-form transformations. Opposing current methods, our
sures for torque-controlled quadruped robots that free-fall on a flat horizontal ground, knowing neither the distance to the landin
paper presents a novel framework for legged loco-manipulation that considers whole-body coordination through a hierarchical
existential capability for animals. While some robots can be designed with multiple orientations for mobility, most seeking to aff
ontact conditions between the environment and the robot body. We apply Morphable StateNet to the standing-up motion of hum
-body control. The latter firstly maximizes the push forces applied to the robot's center of mass to make agile whole-body recov
f SDFs. We further leverage the kinematics chain of the robot to produce the SDF representation in joint space, allowing robust
ctions. Validation has been conducted in MATLAB/Simulink and MuJoCo simulations with Rollin’ Justin.
Integrating force control with the coordination of a robot's body and arm, we present an end-to-end policy for legged manipulat
ecision-making to aid humans on low-level tasks, further reducing operation effort. However, shared-control methods for wheele
 it can directly generalize to target domains rather than being specialized for a specific domain or scenario. In this regard, we p
f-supervised manner. Specifically, our method combines masked feature prediction and masked sample consistency to encode
at the robot can seamlessly cooperate with the appropriate individual despite its appearance changes or partial or total occlusio
work, we propose an efficient spectral method for geometric verification (named SpectralGV) that does not require registration.
arious graph augmentations highlighting different compositional aspects of action scenes. Furthermore, our research reveals th
n units (BPUs) to extract multi-scale long-range contextual information and parameter-shared NetVLADs (PS-VLADs) to aggreg
ddress response degradation, we propose a module for learning temporal consistency and reversibility of response, supplemen
ng procedures. MegaPose for example is trained on a dataset consisting of two million images showing 20,000 different objects
mised by low frame rates. Spatial information is also compromised by reducing the number of embeddings as the transformer la
nd, and stand-to-sit. The VTSEA features a SEA-coupled variable transmission ratio adjusting mechanism and works between
n, we show that the optimal configuration in terms of electrical energy consumption and complexity consists of rigid actuation on
lever mechanisms have been implemented in numerous variable stiffness joints. Nonetheless, the integration of the lever techn
(ASRM), so that the stiffness of the actuator can be adjusted (softening or hardening). By designing the lever length, the range
via a switch-and-hold mechanism. Thus, the user would be in full control of the potential energy storage and release timing. In
on error, play, friction, and stiffness. Furthermore, we assess overall actuator performance through position control experiments
us can provide bidirectional actuation. The workflow including design, fabrication and mathematic modeling of the pneumatic ac
lemented and characterized. A specific step-by-step control sequence is then elaborated based on its characterization. Testing
forming cyclic tasks. However, conventional springs can be limited by their non-customizable force-deflection behavior -- for ex
and tether. First, we present a path planning solution based on optimal Rapidly-exploring Random Trees (RRT*) with novel sam
rogeneous UAVs by ensuring steady communication based on the shortest total mission time. To accomplish this goal, a relay
ensor mounted on the UAV. Before online navigation, a high-fidelity simulator acquires a flight dataset to learn energy consum
elation through the mixer matrix. Furthermore, the effective workspace of the testbed can be extended to the dimensions of the
here the first part of the in-field mission is dedicated to self-supervised fine-tuning of a pre-trained convolutional neural network
he stall regime, sideways uncoordinated flight, inverted flight, etc. The method uses the differential flatness property of the glob
tionary model-based disturbance observer. The resulting observer can guarantee a provable convergence in optimal conditions

indoor environment. Current approaches utilize vision-based or 1-dimensional sensors to support nano-drone perception algor
ach trains painting policies that operate effectively in both virtual environments and real-world robotic sketching systems. We ha
o generate manipulator joint trajectories with environmental safety constraints and scale the trajectories to satisfy joint moveme
estimation, through repeated attempts to pluck individual strings. A subsequent stage implements a Safe Active Exploration pa
des have led to a significant demand for skilled roboticists across industry and academia. This demand has sparked a need for
tries (i.e. symmetries) present in the problem. We demonstrate that when trained with only 5 minutes of real world flight data, o
red or controlled during training, which we call ``manageable''; and 2) those that cannot be directly measured or controlled (e.g.
s system, the wide-angle camera offers panoramic images as prior information, which helps the search camera in capturing de
onomous robot perception and navigation within very dense obstacles and dense human crowds. First, we propose a framewo
rs in a language-guided framework. We propose a Vision-Language Interpreter (ViLaIn), a new framework that generates PDs
e achieve this by providing the second-order derivatives of MHE, referred to as the MHE Hessian. Remarkably, we show that m
a defect library. MCDEN decomposes abnormalities to basic elements from the library while editing outlier features to reconstru
(LLMs) to achieve generalizable robotic manipulation, but most of these researches focus on high-level task planning, sidelining
LLMs often lack feasibility and correctness. To address this challenge, we introduce ISR-LLM, a novel framework that improves
babilistic inference to address these demanding requirements. Our approach implements a combination of a convolutional spiki
ns and time-sequence models to efficiently capture spatio-temporal relations between interactive agents and predict their dynam
n from voxel-based features to point features using Gated Recurrent Unit (GRU) refinement. To further enhance scene flow esti
e learning process under gravity influence, which aids in modeling the nuanced relationships between joints vital for motion cont
nging this work one step closer to being a practical and useful tool for roboticists. First, we formally introduce the PO-ECAS fra
ethod enables the agent to learn reusable skills with a single policy and solve zero-shot tasks under the guidance of LLMs. In pa
ings state-of-the-art Inverse Reinforcement Learning (IRL) techniques to solving bipedal locomotion problems over complex ter
ely (with probability < epsilon) to falsely issue an alert; any alerts that are issued should therefore be heeded. Our method is sp
ecently introduced probabilistic belief-dependent constrained POMDP. We present a technique to adaptively accept or discard
ng a policy observation with a quantile and training it with reinforcement learning. Theoretically, we show that such policies achi
y forgoing the traditional resampling step. This ensures a broader and more diverse particle set, especially vital in high-dimensi
application scenarios. Moreover, evaluating test cases solely based on their fault-revealing capability is insufficient. To addres
eterogeneous robotic actions according to the information emerged during the interaction. The AI-enhanced telepresence robot
properties using compliance control. Our approach involves using Reinforcement Learning (RL) to train a robot to compliantly m
nse knowledge with automatic generation capabilities. SG-Bot employs a three-fold procedure—observation, imagination, and
erogeneous multi-robot planning introduces new challenges of coordination while also pushing up against the limits of context
augment highly semantic and information-dense language instruction with position cues. We utilize a Multi-modal Large Langua
of artificial intelligence and machine learning, notably robust learning, while also extending its applicability to industrial robotics a
asks and makes decisions on two systems based on instruction types. Our RFST consists of two key components: 1) an instruc
h outside robotics has successfully studied instruction-based prompting, where providing LLMs with API documentation and de
s on the overall behavior that results from their composition. By verifying the performance of these subtask policies using a mul
success of safe and efficient navigation. However, determining the ideal timing to execute replanning in such partially unknown
er is both time-consuming and labor-intensive. This paper introduces a novel loop closure detection method that leverages pret
al simulator (i.e., the robot's "imagination"), the robot assesses whether the object is a chair and determines how to reorient it p
skill learning in simulation by automating the generation of 3D assets, task descriptions, task decompositions, and reward func
ss of LLMs for complex long-horizon tasks. Inspired by human intelligence, we introduce a novel framework named FLTRNN. F
rness these multimodal foundation models to enhance the robustness and adaptability of autonomous driving systems. We intr
planner to execute each sub-goal. When complex environmental and temporal constraints are involved, inference over plannin
y on onboard visual and proprioceptive sensing, without requiring detailed maps with obstacle layouts or precise object location
n from human forgetting this paper implements and evaluates novel forgetting techniques that allow consolidation in the robotâ€
tional spatial awareness and a multi-modal virtual-physical interaction approaches. The former seamlessly integrates the physic
cing prohibited and restricted areas, avoiding flights over humans, and managing maximum limits of altitude and speed. Our fra
ficient method to plan motions that are consistent with the dynamics of over-actuated UAVs. With a hierarchical control structur
nario of a team of quadrotors with heterogeneous GNSS capabilities. These simulations answer an open question, showing that
tegy, which provides additional normalization properties to the field. This allows the following of unbounded curves and reduces
spired by kites, this paper proposes a new landing method for tethered lifting-wing multicopters, which can be used without loca
of the main problems being the lack of publicly available omnidirectional datasets in relation to drone flight. With this paper, we a
accurate distance estimation as well as better scalability with respect to the number of neighbors. Additionally, a novel swarm c
tings. To overcome these limitations, we propose the Heuristic-based Incremental Probabilistic Roadmap Exploration (HIRE) pl

estimation without needing an analytical model. They struggle, however, with high-dimensional systems as they operate over a
 ructure of the MPC problem for efficiency. We demonstrate TinyMPC’s effectiveness by benchmarking against the state-of-
erate gaps between each other. The gap distance can reach about 16 times the diameter of oocyte. In addition, a capacitive se
me-division multi-channel (TDMC) CLRC strategy such that parameter convergence is achieved even without the IE condition. I
uch virtual events. The full pipeline runs on the actual vehicle and interacts with the physical world, but the safety-critical events
ng of both risk and plausibility objectives. The plausibility objective leverages generative models, such as a variational autoenco
can lead to a loss of tracking and catastrophic crashes of the controlled vehicle. To this end, we propose an effective data augm
aset, and use it within training to generate better predictions in high risk situations. Through these location-based and speed-ba
e the realism of existing traffic models. This work also identifies two main challenges: capturing the nuances of human preferenc
ut prohibited actions, assessing the robot's comprehension of these restrictions, and, most importantly, ensuring compliance. M
data. This approach can result in suboptimal system performance due to the amplification of noise and the large transmission d
s feature distillation as the fine-tuning task to transfer knowledge from the teacher model, which has been pre-trained with comp
trajectory overlaps. In this work, we propose Future Interaction modeling for Motion Prediction (FIMP), which captures potentia
this modeling approach, perturbation-based augmentation can be applied to enrich scenarios further. In addition, the represen
work and a transfer strategy called KT-BT: Knowledge Transfer through Behavior Trees. The KT-BT framework follows a query-
s and exchange their local knowledge base until they converge on the global knowledge base. In the assignment phase, each a
d controls over multiple prediction horizons. Model training is performed in a self-supervised manner using a synthetic dataset.
using a pre-trained Vision-Language grounding model. We achieve state-of-the-art zero-shot object navigation results on Robo
a volumetric representation can be reduced to form a suitable input. From the input, tiles are selected and embedded by differen
arser module based on Bidirectional Encoder Representations from Transformers (BERT) to parse these steps into initial BTs. S
on network to capture the high-order interactions among agents, which not only models the evolution of hypergraph over time b
airs generated by teacher LLM (GPT-3.5). A distinct pretraining strategy is devised to align numeric vector modalities with stati
ity, or fail to tackle agents' heterogeneity. In this paper, instead of focusing solely on pairwise interactions, we propose a Hetero
ter. We introduce knowledge about the previous motion state by a status-quo-ante filter and hereby identify further false positiv
they see a substantial drop in efficiency in low-light conditions, consequently hindering both scene perception and path plannin
s: the essential features for ego planning and the effective data augmentation techniques to reduce compounding errors. Furth
mework to automatically produce the topological map of complicated intersections. This framework adopts the crowdsourcing w
he latter assuming full environment observability. We introduce the Scene Informer, a unified approach for predicting both obse
e efficiency and accuracy, we propose a novel lightweight visual semantic localization algorithm that employs stable semantic f
ments. Previous interactive prediction methods use an encoding of k-mode-samples, which under-represents the full distributio
ally, MARC realizes a critical scenario set that reflects multiple possible futures conditioned on each semantic-level ego policy.
hod based on maximum entropy inverse reinforcement learning (IRL) is proposed to recover the cost functions of the all demon
two wheeled platforms with little hardware modification compared to conventional wheeled robots; we collect datasets of our wh
ble) and efficient (the frozen robot problem). With recent advances in deep learning approaches, the common reaction has bee
lity is introduced to correlate the tracking performance for guiding robotic motion planning. Based on the detectability measure,
riving tasks encompassing visual and spatial reasoning, predicting the intents of traffic actors, and decision-making based on v
s of waypoints for multiple time steps. On the other hand, a cross-attention mechanism implicitly associates the latent feature re
self-correcting visual navigation method, SCALE, that can autonomously prevent the robot from the OOD situations without hum
n the existing scene. Our unified modeling approach, combined with autoregressive agent injection, conditions the placement an
image super-resolution method, ignoring the 3D geometric information of the point cloud obtained from a LiDAR. To solve the a
em and focuses on predicting the border of traversable area. Since it encodes the spatial information of the surrounding environ
including human positions, head orientations, and 3D skeletal keypoints from onboard in-the-wild sensory information. The res
 eraging multimodal human behaviors in the setting of N-driver concurrent interventions. Specifically, we first learn the N numbe
mitigate the posterior collapse problem, we introduce a two-branch sequence encoder to capture both discrete options and con
d for unstructured environments. It involves pre-training a uni-directional traversability classifier, termed UniTraT, through self-s
it addresses the challenges of feature-free object modeling and continuous registration. However, achieving complete environm
ng robot navigation techniques grapple with the intricacies of navigating through unknown environments. To tackle these comp
ic graph, which improves navigation efficiency by handling task-driven features taking both image correlation and topological la
riments we distribute 72 robots to a crowd of workers who operate them in homes, and show that OpenBot-Fleet can learn robu
only a subset of the semantic information present in the original sensory input. To study the effect of these modular approaches
ithout suffering from catastrophic forgetting. Furthermore, the model is able to determine confidence intervals of the prediction v
ed visual representations at the categorical level. We identify the key bottleneck in using a frozen pre-trained visual backbone fo
sk manipulation skills using (a) semantic augmentations that can rapidly multiply existing datasets and (b) action representation
h pick-and-place. This enables language-conditioned rearrangement to be performed zero-shot, without needing to collect a tra
l reward specification or environment resets, even if the policy is pre-trained. We introduce RoboFuME, a reset-free fine-tuning
significant performance gains by creating a large ensemble of optimized single models. We then develop a generalized framew
al unit mismatches in robotics code are often intentional (e.g., when operating a differential drive robot), resulting in false positi
r to explore the differences in the take-off process under different conditions, we conduct the flight experiments under different
amics without affecting the propulsion efficiency. Kinematics and simulation analysis are performed to characterize the influenc
type and the turning radius is also stated. Then, an approach to the attitude is proposed, defining the relation between the defle
wimming motions. The model is synthesized by a nonlinear oscillator and a traveling wave equation. By changing four paramet
mud with intermediate levels of water content (25%-26%), but decreased significantly on higher or lower water content. Measure
obtain hydrodynamic benefits, such as in increased thrust, or improved movement efficiency. One main bottleneck has be
y and timely estimating the swimming direction. Firstly, we effectively compensated for the yaw angle drift of the IMU sensors th
y interchangeable caudal peduncles to our robot to explore the effect of a tunable stiffness mechanism and determine the optim
d neural network (BINN) is introduced to generate collision-free escape trajectories through the dynamics of neural activity and
We investigate several open robotics and neuroscience questions: 1) What is the role of explicit interoscillator couplings between
e concept of form closed region is used to produce predictable constrained motion in a cluttered environment using lateral undu
dinal (x) axis, and then tweaked the design to reduce the number of faces. We then attached the shell to the robot by modifying
as the motor dynamics. In contrast, in this paper, we consider a quadruped jumping task that we rapidly optimize online. We d
low-up to the original Dima design that uses aerodynamic drag on the tail as the primary mechanism for generating the stabilizi
f spinal movement, such as twisting, extension, retraction and rotation, in helping STRAY realize efficient locomotion. Prelimina
ance the wheeled motion's agility. Additionally, we establish comprehensive mathematical models for adaptive dynamics mode
memory alloy, magnet-locking mechanism, and corresponding support structures, is fabricated. Experimental results demonstra
alog for depth, a novel deep learning architecture was proposed based on sinusoidal and random data in order to obtain a mor
tric actuator, and an optimized control circuit for high-impedance-width modulation and charge recovery, this electronics realize
parameters to achieve high jumping energy (first step) through optimizing the storage and dissipation of energy and then high
stic shape memory alloy (SMA) spring that is manually loaded to store elastic energy. The on-board sensing and computation t
force parallel to the water surface, and three passive legs utilize surface tension to support the body mass and maintain contac
res, which deform cyclically when excited periodically using Joule heating. The VLEIBot can swim at speeds as high as 15.1 m
l travel away from the nest, come back by means of odometry, and eliminate the resultant drift by inferring the home vector orie
orm employs a four-wing configuration, each of which integrates an extra servo motor to enable the rotation of the flapping plane
gh thrust-to-weight ratio, and cost efficiency. Regarding design, multiphysics simulations guided structural optimization to incre
 l chemical stimulation and calcium recording. In this system, reliable evoked sensory-motor mapping (success rate > 89%) from
sed mechanism can change the effective length of the robot from almost zero to any length by moving the robot's body back an
t describe the hardware design that allows the robot to achieve a modal switch between flight and wheeled modes through fold
lification via mechanical intelligence in lateral undulation, which offloads feedback control to passive body mechanics and intera
ed on terrestrial robots in the air, e.g., either using a tail on the hexapedal robot RHex or using a reaction wheel on the jumping
previous work, we proposed a method for the rapid production of 3D objects using the centimeter-sized modules (referred to as
as proposed; however, stiffness calculations, derivation of inverse kinematics, and validation of stiffness and load capacity were
ck of the SCP-driven robotic eye, measuring the temperature states of the slender SCPs is generally impractical for the ubiquito
s markerless, white-box, and delivers superior accuracy and robustness. We propose to use two key technologies: differentiab
Anything Model(SAM) without additional training. With the automatically generated masks, we optimize the extrinsic parameters
calibration algorithm, for cases where only a single pair of images is available. Furthermore, we introduce a simple yet effective
tion with a memory of past evolution. LTO uses a fixed-grid search for the optimal parameters per frame and without memory.
fic tools or highly depend on the accuracy of camera motion estimation. To address these issues, we present PeLiCal, a novel
resolution and Field of View (FOV), which may degrade the reliability of the calibration results. Therefore, our research introdu
e the extrinsic parameters based on LiDAR-mapping space occupancy while removing motion distortion. In the final stage, the
d robustness. In the first stage, we detect objects in the image and point cloud and build graphs on the objects using Delaunay
tern can be detected. A blinking pattern can also be displayed on a screen. Then, the pattern can be directly detected from the
et expectations. Moreover, most methods require a precise initial guess for calibration to succeed. In this paper, we introduce L
ormed neural network (PINN) to predict and identify collaborative robot joint dynamics. The procedure involves deriving the state
ation separately and using only the parameters of the objects in that scene as inputs. The resulting lower-dimensional models g
ch limits the calibration accuracy. To this end, we introduce a novel extrinsic calibration network (HIFMNet) in this paper. It esta
d End-Effector Calibration (GBEC) that enhances the repeatability and accuracy of the derived transformation compared to trad
nfigurations. Measurements are collected when the robot reaches each discrete configuration and stops. A neural network then
ethods have been proposed, but their performance is barely satisfactory in most cases. These methods usually suffer from issu
meras, calibrating them with LiDAR presents a significant challenge. In this paper, we propose a novel method to calibrate the e
wpoint-difference. Current works are mainly focused on unmanned vehicles, which is short baseline and small viewpoint-differen
ation using a heightmap expressed in the robot's local frame. Next, data is collected in simulation to train a heightmap predictor
works relying on heuristics and anomaly detection to update navigational information, we train our navigation policy to reconstru
rages RGB observations and a pre-trained vision-language model to generate a language-grounded value map. VLFM then us
and goal-agnostic exploration, with the latter providing the ability to search novel environments, and the former providing the ab
ce novel network structures for single-agent control learning that explicitly capture these symmetries. Moreover, we investigate
provide a distribution as the output, obtaining a mean and variance for the collision check. The obtained variance is processed
cenario Replay (PSR), that stores and partitions previously observed human trajectories, referred to as scenarios. During plann
me-step multimodal motion predictions are obtained. A nonlinear Model Predictive Control (MPC) solver utilizes the prediction o
atic programmings (QPs) following the paradigm of alternating direction method of multipliers (ADMM). Furthermore, these low-
Furthermore, there are discrepancies between simulated and real environments that impede real-world implementations. To o
works have drawbacks in interpretability and stability. To address these limitations, we propose a trajectory planning method for
sive numerical results to illustrate the advantages of our proposed approach over uniform sampling in terms of the quality of th
pt to form a sparse graph at the global level, which records the history of the exploration. A series of subgoals that can be obtai
obots. DiPPeR performs on average 23 times faster for trajectory generation against both search based and data driven path p
ology known as Sum-Of-Squares (SOS) Programming. SOS programs, however, are notorious for their poor scaling properties
ath directly. Then the robot can follow the generated path smoothly using sophisticated velocity-planning and path-following con
ibutions with infinite support. On the other hand, safe planning typically requires very fine time discretization, paired with distrib
ons is collision-free. We demonstrate that our formulation is practical for real world deployment, certifying the safety of a twelve
re interrelated details between images at multiple hierarchical levels. Finally, a spatial refinement module, constructed on the fo
nt the temporal information and use learned time estimators based on different statistic distributions to examine and select the
o solve problems with a very high number of parameters. Therefore, our architecture includes neural image encoders, which tr
on. Notably, our formulation avoids hard constraints by construction, which in turn avoids abrupt transitions in robot behavior ar
h (e.g. beaded/string curtains, pliable tall grass), from ones that must be avoided (e.g. transparent surfaces such as glass walls,
ariations, including height, uncertainties, and slope characteristics. Subsequently, we construct a traversability map based on th
esign a collision-based observation model, which is differentiable and can be easily applied to various estimators. Secondly, we
-R, a novel 3D graph-based estimation network.GDN-R employs a layer-wise probabilistic graph-rewiring algorithm leveraging
ch involves a two-step process. Firstly, we partition the passable space into blocks until there is no vacancy for a minimum-size
agents’ trajectories, 2) use statistical techniques to provide uncertainty intervals around these predictions, and 3) learn an
works and represent them as a connected graph. The graph is then used to train an attention-based Graph Neural Network (GN
ions on node selection during path planning. Secondly, a slow-speed informed guiding approach is introduced to optimize the s
ational complexity. We address these challenges by designing an online non-parametric Lidar-based safety function using the G
olve interaction and physical reasoning about such objects. To address this limitation, we propose PhysObjects, an object-cent
on the Prediction of the Required number of Views (PRV). The key idea behind our approach is that the required number of vie
e limitations, we propose to integrate implicit representations into the One-Shot View Planning (OSVP). The key idea behind ou
t latency. To overcome these limitations we propose EfficientQ3M, an efficient, modular, and multimodal solution for object que
M3, which leverages multi-view RGB images captured before and after interaction to model articulated objects, identify movable
l motion-focused model with a dual-branch structure for Lidar moving object segmentation. Novelly, we decouple the spatial-tem
etrically complex regions. In this paper, we propose a two-stage training process, decouple view-dependent and view-independ
use the estimates to train a neural volume, which serves as an intermediate representation ignorant of challenging appearance
scribes the positions and visibilities of the entire cloth mesh vertices, enabling more efficient dual-arm and single-arm target-orie
Drive dataset to include pixel-level labels for amodal optical flow estimation. We present several strong baselines, along with the

scene flow estimator, enhancing scene flow estimation. (2) developing a symmetric inter-frame correlation feature extraction m
Circum-views provide a novel solution for effectively addressing the representation of dense and fine-grained scenes. Specifica
king pixel correspondences, obviating the need to rely on gradients of explicit correlation volumes as such search directions. W
evance of the collaborators and their associated cameras according to a learned spatial prior. This proactive understanding of t
te diverse and high-fidelity 3D scene point clouds based on the image representation of range and reflectance intensity. Our m
3D object detection, semantic segmentation, and dense depth estimation. To counter the challenge of negative transfer — th
benchmarks. Our novel transformer-based framework includes a cross-space transformer module that learns attentive feature
articularly fabricated for large deformations of flexible beams. Using a discretization-based approach, the contact condition can b
o determine the boundary between stable contact and macro slip. This paper describes a generalized 2-d contact model under
lity enables us to estimate contact positions and surface normals by analyzing its deformation during interactions with unknown
, compliant contact models are used, connected in parallel to model multi-point contact and constraints such as a hinge. They a
ers are distributed within the gap of papillae pins. Variations in light, dark and grey effects can be generated inside the sensor u
presents both the soft body and the rigid body within graph structures, where nodes hold the physical states of the meshes. We
shing constraint and the nonholonomic constraint of the robot can be simplified as a concise linear motion constraint. Then, the
 deriving the second-order rolling kinematics; deriving the 3-D rolling dynamics; planning rolling motions that satisfy the rolling dy
w that rubber friction varies strongly with normal force Fn and contact velocity v, and includes a significant stochastic componen
ets, which limits their real-world applications. In this work, we combine the benefits of these two paradigms and propose an inte
tributional changes and often lead to overfitting of TTA models. To address this problem, this paper first presents a robust TTA
class-incremental manner, and can be evaluated on-the-fly (anytime inference). To accomplish these, we propose a novel emp
perience utilization of old tasks is inefficient. To address these issues, we propose a experience consistency distillation method
limitation, we propose an enhanced hybrid system with a nested hierarchical action abstraction that can utilize previously acqu
ry to tackle manipulation tasks of ever-growing complexity. LRLL achieves this with four novel contributions : 1) a soft memory
uture tasks, demonstrating the potential of open world and lifelong learning. We evaluate the proposed framework on multiple ta
VCL). In this framework, PSNN introduces uncertainty during spike emission and can apply fast Variational Inference by optimi
in unstructured demonstrations. Continual skill discovery updates existing skills to avoid catastrophic forgetting of previous task
ach is employed to compensate for the poor stability in their servo-controlled attitude of each bi-copter. In bicopter configuration
ing the spatially local action patch and the low-frequency actions in the frequency domain. With this reshaped action space, we
presenting a substantial hurdle in developing task-tailored modular robots. Previous approaches either lack adequate explorati
ght corners and curved surfaces. This paper presents a novel reconfigurable glass cleaning robot called "WiBot" to address the
nction for the resolved motion kinematic control for one or more robots to adjust their motion to avoid colliding with the obstacles
scribes both geometric and semantic aspects of the environment. Then, we leverage graph-theoretic tools, namely graph spann
f blind UGVs by introducing cooperation with one UAV, which is the only robot that has full perception capabilities in the group.
energies, e.g., motion smoothness, are introduced in the proposed design to improve the flocking behaviors. The optimal contro
ese robots are enhanced by FBG sensors, which sense the small force interactions between the surgical instruments and retina
r tool is designed to be cost-effective and straightforward to use without requiring human intervention during the labeling proces
ltimodal Handover Failure Detection dataset, which consists of failures induced by the human participant, such as ignoring the
, and images where objects are observed close enough so they appear larger than the field of view of the camera. We also pro
ompletion. These scenes exhibit long-range perception and minimal occlusion. We develop an automated annotation pipeline le
eate an RGBD dataset tailored for robotic industrial grasping scenarios and evaluate it across various tasks, including object de
OOIs in effectively representing dynamic environments. This oversight creates a significant gap in the coverage of interactive ele
RTS-GT) dataset to support localization research with the generation of six-Degrees Of Freedom (DOF) ground truth trajectorie
strategy is further introduced to enrich data diversity. Extensive experiments show that models trained with RaSim can be direc
nthetic data, collected with a real robot platform and a physical simulation platform, respectively. Our current dataset includes 1
d leverage large-scale human-human interaction data that is more easily accessible? Our key insight is to exploit a correspond
cale dataset that consists of 1282 hours of diverse driving scenarios from 4 cities (Las Vegas, Boston, Pittsburgh, and Singapor
pedestrian tracking outcomes. Our system enables large-scale data collection in diverse environments and fast trajectory labe
form better than ones trained on the robot data only, even when evaluated solely on robot episodes. We explore the economics
HGNNs) to transform regular driving scenarios into corner cases. To achieve this, we first generate concise representations of
n simple cases, such as push or pick-place tasks, relying solely on visual guidance. In reality, there are many complex skills, so
al navigation using only onboard RGB observations. We introduce a definition for such behavior based on the counterfactual pe
ndering collective progress toward safety performance enhancement. To address this limitation, we introduce RiskBench, a lar
74 stereo images and lidar point clouds with 3d bounding boxes. Furthermore, we provide an initial approach and a suitable me
fill this gap, we present IDD-X, a large-scale dual-view driving video dataset. With 697K bounding boxes, 9K important object tr
chmark leads to the common practice of training a model on one benchmark (e.g. Waymo) and then assessing it on another da
on of remotely operated vehicles (ROVs). The training subset consists of a large number of synthetic images with 6D pose grou
 utdoor environments. The GOOSE dataset incorporates 10000 labeled pairs of images and point clouds, which are utilized to tr
r suction-grasping in bin-picking tasks. This pipeline consists of a series of methods that progressively transit from a single obje
arance changes. Each scene has multiple revisits across a year. To establish the ground truth for VPR, we propose a semiauto
stry and orchards. TreeScope provides LiDAR data from agricultural environments collected with robotics platforms, such as UA
first project 3D point clouds onto 2D image domain to investigate the distribution of the LiDAR's echo pulses, which is then use
ws the popular pseudo-label based self-training framework. Unlike previous works using single-model uncertainty to refine pseu
roduce a cross-modality optimization process, including a local adaptive key region aggregation module and a global modality c
obustness. The paper also introduces a motion loss approach, which improves articulation estimation performance by emphasiz
cally, our method learns to associate low energy scores with inlier points and high energy scores with outliers allowing for robu
nable efficient storage or communication, we also quantize the generated descriptors into integral values. We perform EdgePoin
t the Tree-based Transformer (TrT), which is able to extract abundant local and global features with linear computational compl
ds EgoPAT3D, the sole dataset dedicated to egocentric 3D action target prediction. We augment both its size and diversity, enh
parameters to fit target domain knowledge through self-training. Unfortunately, these approaches often amplify the challenge o
per, a new spatio-temporal sequences fusion-based self-attention, STNet, is proposed to perform slip detection by allocating m
he model. Our model uses the raw point cloud representation of the surrounding objects in the environment as input. Using raw
D images annotated with camera poses and configurations. This enables significantly greater applicability than existing approac
e joint learning in 3D point clouds. Given a 3D point cloud object, our method detects the affordance region and generates appro
e then merged with the final planes by post-processing. DPC rearranges the point cloud by voting layers to make subplane clus
images for the keypoint locations. In this paper, we remove these dependencies by utilizing inpainting methods: In the training
work, we leverage recent advancements in neural rendering to improve static and dynamic novel- view UAV-based image syn
rategy for visual representation learning that enables zero-shot scene transfer and real-world deployment. Control policies rely
OS) allow detecting them only against a known background and from a static camera, making these approaches unsuitable for
ns is represented by lightweight deep learning models directly deployed aboard nano-drones. This work tackles the challenging
nerates both point-level and cluster-level pseudo labels to harness knowledge from unlabeled data. It notably enhances the tra
panoptic representations with consistent instance IDs. We evaluate this novel system in a very challenging horticultural scenario
ables adaptive and robust supervision for both key areas and textureless regions by allocating weights based on gradient magn
era gap problem. This paper presents a novel framework that bifurcates the high-dimensional prediction challenge into two man
es of surround-view LiDAR point clouds with enriched semantic annotation. Unlike existing datasets confined to a frontal perspe
nt cloud sequences. We also design a semantic auxiliary training strategy to make the predicted LiDAR point cloud sequence se
or in task space and a trained reinforcement learning agent for the mobile base to account for the kinematic feasibility of the mo
osed of a 3D back-end that extracts rich features and enables the semantic segmentation of the lidar frames, and a 2D front-en
us with regard to their location. To tackle this problem, we present BASK, a Bayesian approach to tracking scale-invariant keyp
ak inductive bias for predicting the whole trajectory, often leading to poor map compliance, i.e., part of the trajectory going off-ro
ction hysteresis approximation, which is crucial for tasks involving extensive dynamic to static state transitions. In this paper, we
ormers in parameter space trained on different MuJoCo locomotion problems, forming multi-task models without centralized trai
We use an asymptotically stable dynamical system to model human goal-targeted motion by enforcing the human walking traje
ic Q-learning (POAQL) method. POAQL considers the cumulative reward of the observed subteam instead of the whole team, w
ch are often unobservable, and track its transition as new actions are taken. Our framework then conditions each action on the
atures from an off-the-shelf LiDAR feature extractor. We evaluate MGTR on Waymo Open Dataset motion prediction benchma
able texture recognition. Along with the active sampling strategies, we evaluate neural network architectures, representations o
architecture, called Neural Circuit Policies (NCPs). However, a comparison of these two synaptic models, within the same arch
rformance while delivering high accuracy on edge devices. Our proposed method involves three key steps. Firstly, we review st
ramatically with the growth of training data. On the other hand, implicit methods like Neural Radiation Fields (NeRF) suffer from
fusion Model and automatically labeled to train a classifier. During this training, we further align features of the two domains usi
e to predict the future motions of the agents encountered in social scenarios. Contrary to prior works, we reformulate the social
ensure both the perceptual quality for the human and the performance of downstream tasks during the transmission? To addres
approaches such as differentiable programming, global search, and optimization. In combining approaches from these two para
kles the problems of policy adaptation and safe reinforcement learning. SafeDPA jointly learns adaptive policy and dynamics mo
he Cosserat rod theory and incorporated into the PINN’s loss function. The boundary values for the training are obtained fro
complex constraints, such as planning on a lower-dimensional manifold of the task space while considering the robot's dynamic
and path planning in complex environments. The proposed method iteratively selects objects, determines their relocation regio
n-depth perspective of the field, revealing surprising insights, including the effectiveness of a straightforward baseline approach
ormance to unseen environment settings. To this end, we first propose to formulate online grasp learning as an RL problem tha
ng pixels with reward feedback as labeled data and others as unlabeled, it efficiently exploits unlabeled data to enhance learnin
n using monocular images and keypoint information. Our method employs a two-level nested U-shaped architecture, featuring a
changed, while the images from a failed task will change significantly. Using this insight, GANDER enables inspection and task
oriented grasping called the One-shot Task-oriented Grasping (OS-TOG) framework. OS-TOG comprises four interchangeable
mer (PAST) -- a sequence-to-sequence neural network -- to infer assembly sequences recursively from a target blueprint. We th
ramework formulates the adaptation design and control as a dual Markov decision process and is able to improve robot-object
sual pretraining models as vision encoders. Our study investigates the robustness of this framework when applied to a dual-arm
n forestry. To address this challenge, we introduce DeRi-Bot, the first framework that enables the collaborative manipulation of
r'' policy pre-trained with multiple camera viewpoints guides a ``student'' policy in learning from a single camera viewpoint. To e
(SVFL), a novel method for robotic control which leverages visual and force feedback. We demonstrate that SVFL can significa
enging tasks including reasoning with occluded objects, novel objects appearance, and object reappearance. Throughout our e
p tangled-prone objects for industrial bin picking. First, we learn PickNet - a network that maps the visual observation to pixel-w
ork, we explore how demonstrations for fabric manipulation tasks can be collected directly by humans, providing an extremely
ffector pose evaluation, and motion planning of the robotic arm into a single model, enhancing the success rate while reducing
I selects the best course of actions based on the current state derived from visual interpretation, whereas Critic II evaluates the
this paper, we solve this problem by developing a novel neural network DefGoalNet, which learns deformable object goal shap
o the generalization gap. Towards an answer to this question, we study imitation learning policies in simulation and on a real rob
d on multiview stereo. Compared to existing works, our method not only reconstructs the whole 3D scene in order to obtain high
o isolate the relevant motion in a demonstration, and parameterize a low-level controller to reproduce this motion across chang
arameterized manipulation primitives. By learning the low-level manipulation policies, our approach can control the object's stat
ject. We find similar objects in food, marine, and manufacturing domains. The aim is to actively manipulate the object into an a
certain types of articulated objects, they lack generalizability to unseen objects, which significantly impedes their application in
icking with restricted reachability. In this paper, we propose a novel approach for learning dense and diverse 6-DoF grasps for
nto account which ones might be the most valuable for learning. In this paper, we address this problem and propose a novel ap
asping in cluttered environments. In contrast to existing RGB-D based grasp detection methods, which heavily depend on depth
ge according to the tentative action. The correction stage utilizes task properties to obtain the cropped image with task-relevan
n imaging principle. On the other hand, RGB images, while rich in texture information, lack essential depth and 3D information c
mentation and 3D RL to improve the efficiency of RL policy training. To improve the stability of the policy on real robots, we des
mization alongside a long-horizon control policy using RL. MORPH is designed to ensure that the optimized hardware proxy rem
ata-efficiency by using all the collected experience, which empowers learning from real robot experience only. Moreover, the res
ustification of our approach and we empirically validate the approach both in simulation and real-world tasks. Our method, whic
d require tremendous effort and computational cost to generate new datasets and relearn the grasping task. In this letter, we pro
sp heatmaps as guidance, which infers in a global-to-local semantic-to-point way. Specifically, Gaussian encoding and the grid-
nd the controller uses the 2D keypoints error for control like traditional image-based visual servoing (IBVS). HPN-NC can comp
ns a policy from a binary reward and a few simulated example grasps. It does not need any real-world grasping examples, is tra
en a sequence of previous time step point clouds obtained with LiDAR sensor. ATPPNet leverages Conv-LSTM along with chan
upervised Learning (SSL) approach, called Transformer-CNN Cohort (TCC), that consists of two students with one based on th
t scenes. In addition, conventional approaches require many annotated low-light crack images which is time-consuming. In this
ed to full-precision point cloud networks. However, achieving a fully binary point cloud Transformer network, where all parts exc
h employs the diffusion model defined by mean-reverting stochastic differential equations (SDE). Using our proposed new obje
isual cues can effectively be bound. We investigate how state-of-the-art models perform in recognizing nouns and modifiers fro
espondence Loss (CCL) for view-invariant dense descriptor learning, which adopts the concept of cycle-consistency, enabling a
MLP decoders (MLM) and effective keyframe selection strategy. Experiments conducted on the commonly used Replica and T
OV-NeRF), an object-composable NeRF model that is the centerpiece of a real-to-sim pipeline for synthesizing training data targ
policies using vision-based rewards. First, we use a contrastive-based objective to learn visual representations. Next, we cons
per, we propose a real-time approach for estimating both contact establishment and subsequent changes by leveraging the dep
lation and real-world scenes. To evaluate the proposed method, we conduct comprehensive simulation experiments in three di
oposed controller depends on the sound control theory and its convergence is guaranteed from a computational perspective. It
e appropriate grasp configuration, and inferring the necessary parameters for an accurate throw into the basket. This empowers
nd-object system models is typically infeasible. We present a method for controlling in-hand manipulation using inverse Jacobi
ystem for collecting visual and tactile data, featuring a cost-effective tactile glove to capture human tactile data and Hololens2 fo
ct formation between the object and its environment. In particular, we estimate the contact patch between a grasped object and
connecting the substrate, screw, screwdriver, and fingertips. Considering rolling contacts and finger gaits, our force control sch
task of a simulated cranial window procedure. Preliminary results show promising performance comparable to those of human
stimation of contact state. The efficacy of our proposed framework is validated with a series of experiments that involve the exe
n an end-to-end way. By abstracting several common DLO manipulation tasks into goal-conditioned tasks, DexDLO can perfor
te in picking up and placing dishes in a rack, and the ability to pick up flat objects like napkins and delicate ones like strawberrie
s cannot be accomplished. We propose designing the maximum permissible errors in advance by expressing the target relative
n open topic. We develop a new model-free framework for the modal-based deformation control. Physical interpretation of the m
ds smooth contact dynamics analytically. Our first contribution establishes the theoretical equivalence of the two methods for si
nt contact modes; 2) planning robot contacts; 3) path evaluation and control optimization. This framework offers two main adva
ulate and orient small objects for subsequent classification, effectively. We contribute a specialized dataset designed for classi
problem of the R-UAV has not been investigated. In this paper, we propose a dynamic multi-agent deep deterministic policy gr
te hardware without modifying the existing application in ROS2. In presence of multiple identical services, it dynamically identif
position with respect to their neighbors. This presents a significant simplification over traditional formation control which, at a m
uted DDP (MD-DDP), is a two-level architecture that addresses both consensus and local constraints with ADMM, further reduc
ordinate jointly these robots such that the overall utility is maximized. This work proposes a generic coalition strategy called K-s
stems to supplement GIS (Geographic Information System), offering a more comprehensive and real-time understanding of en
xponential convergence when the objective is strongly convex and has a Lipschitz continuous subdifferential, and (2) o(1/k) con
ommunication manners are separately utilized in different allocation scenarios for analyzing the effects of non-ideal communica
 lly vectored propulsion of the robot and construct software architecture based on the robot operating system (ROS). Then, we
orphing quadrotor is introduced with the kinematic and dynamic models established and corresponding issues and challenges p
sampling-based approach hinges crucially on the chosen sampling strategy, and a well-informed heuristic can notably enhance
ustness of the synthesized policies. The first objective is to maximize the probability of being temporally robust for a given thres
his paper introduces an augmented DeLaN architecture for linear viscous and non-symmetrical Coulomb friction identification,
puting both the residual vector and the Jacobian of the residual vector of the dynamics constrained by the time integrator. Thes
inal PV solver and extend it to floating-base kinematic trees with constraints allowed on any link. We make new connections be
sent the dynamic behaviour of the subsystems at the interface in co-simulation. Such a model-based co-simulation scheme was
ed Inferred Dynamics via Emulating Rehearsals (RIDER), that learns the dynamics of robot-terrain interactions within a compa
stly learning manipulation skills for real-world tasks using only the simulated data. Our framework consists of two main compon
ng process using the spherical helix trajectory, it is guaranteed that the peg can automatically be inserted into the hole due to th
and fed into an encoder-decoder network that leverages edge maps and RGB images for the purpose of initial pose estimation
tree search algorithms to reduce the combinatorial complexity of determining such an assembly sequence. The search can be
description of tasks, dispatching algorithm, etc. The generated solutions are evaluated with a Multi-Agent Path Finding (MAPF)
ometric features; or depend on the given guess of the initial transformation matrix of multiple single-session maps, which is not
riven approach. We cast the problem in an RL setting where the reward function is based on the ability to neutralize attacks tha
le the second approach partitions the environment into disjoint areas that are then assigned to individual robots. Starting from t
vel framework called Longitudinal Control Volumes (LCV) to model the flow of material in a recycling facility. We then employ a
on a cascaded-PID. The numerical simulations are conducted to demonstrate that position control is possible, even with comple
ct-environment model. To generate a classification error bound, it employs a causal envelop filter that is needed due to the una
ethods for robotic applications based on machine data which will be made publicly available to the research community. As a ty
ng accidents or damages to the objects or environment. Situation awareness becomes important in such cases to enable the ro
complete motor failures. Unlike conventional FT control methods that need to pinpoint the failed location, our controller identifie
prove its accuracy. Our method is evaluated on a digital twin of the 3D printer, demonstrating its capability to enable the printer
ons are often incomplete and restrictive. A novel metric, based on actuator capabilities, is introduced to detect LOC in quadroto
troduces a framework that can extract scene graphs (SGs) from sensor inputs to capture the entities related to the AV, and a d
(HTNs) to handle contingencies. The results demonstrate that our approach is able to handle a wide variety of contingencies in
ctive. In particular, this paper presents a novel topological map generation approach that fully utilizes Line-of-Sight (LOS) featu
g force, the perching mechanism requires no power through the use of a ratcheting serial elastic transmission, allowing the UA
pability. These properties address perception needs for diverse robotic applications, including multi-robot exploration and dexter
dimension is reduced, which facilitates real-time planning. The hybrid trajectory optimization is transcribed into a nonlinear prog
osed localizability detection is achieved by utilizing the correspondences between the scan and the map to analyze the alignme
address these limitations, we focused on the use of Underwater Legged Robots (ULRs), which offer greater stability and agile
the GMRF's concentration and flow field belief that are continuously updated by gathered measurements. We introduce the up
sibility and effectiveness of the proposed scheme have been validated in both simulation and experiments (conducted with Ocea
vel of autonomy, this paper proposes a vision-based UAV inspection framework for dynamic tunnel environments without using
This paper presents SOL, a soft-robotic, dual-axis, sun-tracking mechanism for improved solar panel efficiency. The proposed
novel data augmentation technique by rotating joints and adding synthetic image background, and demonstrate its impact on t
lightweight robot manipulators which require a new investigation on the disaggregation of robot energy consumption. In this pa
elf UAVs. Our approach centers on a small landing platform affixed to a robotic arm, adeptly correcting UAV pose errors in wind
a single robots to quickly and sparsely sample the region of interest to generate a low resolution gas distribution map, followed b
model of the phenomenon. However, deploying ASVs has a significant logistical overhead in terms of time and cost. It requires a
t omnivariances within the point cloud. Moreover, we present a novel factor graph-based pose estimation method that fuses tot
involving an unmanned aerial vehicle (UAV) coupled with a digital inline holographic microscope for in situ particle diagnostics.
(Waypoints-Free Autonomous System for Traversability with Enhanced Robustness), uses experience data recorded from a rec
me, the biomimetic tips embedded in ViTacTip can amplify touch motions during tactile perception. For comparative analysis, we
pressure sensing. The sensing behavior of the piezoresistive sponge was experimentally evaluated, showing high sensitivity an
nse of real-time processing. In order for these tactile sensing systems to find use outside of experimental laboratory environmen
eceived relatively little attention when compared to other sensing modalities, such as force, texture, or slip identification. In this
d formulation that allows both model-free and model-based RL to obtain accurate policies for pushing an object to a goal. To ac
from unilateral teleoperation in free motion to bilateral teleoperation in contact. We also present a neural network based two-cla
fine-grained contact geometry sensing; and (ii) a mechanically simple suspension structure to enable three-dimensional elastic
g and proprioceptive sensing capabilities. In order to reduce the amount of embedded hardware, i.e. the cameras and motors, w
e the illumination scheme to different sensor sizes and shapes to fit on a variety of end effector configurations, the sensors can
n effectively utilizing tactile information due to the diversity of tactile sensor form factors, manipulation tasks, and learning object
g a novel and simplified fabrication processes. This makes it low-cost, modular, durable and in the size of a human thumb while
ms. Moreover, it is low-cost and easy-to-fabricate, requiring minimal assembly skills. Functionally, 9DTact builds upon the optic
allows for the adaptation of the camera pose and the scanning of a large contact area. During the detection process, the camera
elies on stereoscopic vision to capture a 3D representation of the environment before contact and uses photometric stereo to re
array (8-9 mm spacing) of distributed MEMS microphones (
e integral of tactile flow, which describes the movement of iso-strain surfaces within the fingerpad. This work presents the first a
e sensors (TacTips). We present a suite of bimanual manipulation tasks tailored towards tactile feedback: bi-pushing, bi-reorien
object surfaces in a limited number of steps. Through sufficient exploration, our algorithm incrementally collects tactile data and
ta recorded by the robot's proprioceptive sensors. However, it faces challenges in being specific to certain robot leg configurati
imizes object velocities from marker flows when slippage does not occur, or regresses velocities using a slippage estimation ne
les of different sizes and textures. After unfolding, at least three corners are visible in 83.3 up to 94.7% of cases. Given these s
ent a finite element model of the force response of the bubble sensor that uses a linear plane stress approximation that only re
precise and secure grasping. To address this challenge, we present a new solution: a detachable Fiber Bragg Grating-based fl
e to sense a wide range of temperatures with exceptional resolution. Additionally, the pressure and shear can also be perceived
uares objective function—which estimates the resistance distribution of the skin—we enhance the touch sensing accuracy an
ate. We detail the sensor design and fabrication process, provide characterization results showcasing its sensing proficiency, a
directional Generative Adversarial Network (GAN) termed SightGAN. SightGAN relies on the early CycleGAN while including tw
ting conditions throughout experimentation. Additionally, constrained by shape and structure, current vision-based tactile senso
tion deducible from markers. In this work, we propose a generative adversarial network (GAN)-based method to generate realis
and a continuous pose estimator-controller to refine the pose estimate and accurately manipulate the pose of the grasped objec
re conducted within eight spatial complex configurations printed from a proportional bronchial model. The 3D experiment result
ore, we combine the primitives into a closed loop hybrid control strategy using tactile and proprioceptive information to leverage
mpedance-separating method. The precise ranges are measured, showing non-rectangle shapes suitable for the response of io
ed. We introduce a model-based approach, where a deep visual-tactile predictive model iteratively learns to perform servoing f
thods utilizing datasets labeled by humans. This paper introduces MViTac, a novel methodology that leverages contrastive lear
pared with several baselines, and intelligently handled current collisions. Our proposed framework is generalizable to a wide va
e point with unknown position for the robot end-effector. We introduce the interactive perception to generate prescribed but tuna
 out aiding to extend the understanding to other heterogeneous systems. To this end, in this work, we provide degenerate motio
with deformable objects beyond the immediate contact area. The biomimetic GTac in use is an improved version optimised for re
sensors. To ensure the safety of the proposed sensor design, an approximate equation is derived to predict the maximum von
n reduce high-order contact mechanical models that need to be solved iteratively to a set of linear equations, and calculate con
oceptive devices or observers. Second, we present a hierarchical optimized whole-body controller (WBC), which takes into acco
asurement equation is established for contact force estimation and a Gaussian process adaptive disturbance Kalman filter (GP
onstraint-based unified force-impedance control. In this study, we first explain the basic concept of abstracting the task constrain
ocess using a Hello Robot Stretch satisfying an Event-based STL specification.
under all possible moves of the adversarial player. The approach is based on growing a (search) game-tree in the hybrid space
tions for the first class and an over-approximate method for the second class. Our experimental results demonstrate the feasib
these issues, we propose a notion of model predictive robustness, which provides a more systematic way of evaluating robustn
rmulation of machine learning methods. In this context, Riemannian manifolds emerge as a powerful mathematical framework t
LP) formulation that combines the motion of the team of robots with the relaxed specification automaton. Our approach combine
ation robustness is maximized, (b) travel time is minimized, and (c) the success probability is maximized. We tackle the problem
delays could potentially lead to the violation of essential safety properties, such as collision avoidance. This paper develops m
strategy that maximizes the temporal robustness of the robot's MITL task. As the first step towards a solution, we define a Mixe
rmer twice, i.e., first encoding the LTL instruction via the Transformer module for efficient understanding of task instructions du
ent is modeled via a Markov chain. While they provide strong theoretical frameworks, there are still many aspects of human-rob
obot to follow a global task planned offline while taking proactive decisions and generating temporal logic specifications for loca
entional decision tree specialized for temporal logic and inter-task constraints. The TB-PDT is built incrementally to represent th
ring large amounts of training data to generalize. We propose LTL-Transfer, a zero-shot transfer algorithm that composes task-
together. For higher accuracy, a special function was proposed to convert the three-dimensional vector into two-dimensional co
erates behaviors by synthesizing a directed graph over the provided motion primitives. If these component motion primitives are
ration of concerns so that managing 'what is known' is decoupled from 'how it is learned'. This idea is realized in a graph-based

a grasping network named transparent object grasping convolutional neural network (TGCNN) for grasping position detection,w
es have been proposed in the literature mirroring this capability, either using an adjustable internal tension mechanism similar t
ps associated with implicit models of local shapes shared across object categories. Specifically, we employ a single expert gra
ently adapt to new objects by estimating their unobserved properties through interaction. The encoder predicts a distribution ove
zation. Specifically, our method generates high-quality hand grasps efficiently without complete object models by learning gras
ral representation defined on the surface to densely encode grasp configurations. Further, we exploit function-to-function transf
ccess to the workspace (to place or attach a new object) and another assuming digital access to the camera software (to inject
eviously seen objects from a database. Nevertheless, it is often limited by the availability of reference data or coverage for obje
multi-object stacking and partial occlusion. To address this issue, we have developed a grasp manipulation relationship detectio
can be affixed to any robot part, e.g., end effectors or arm. We tested several regression and classifier methods to learn the re
ted grasps can be exploited in the real world. More than 7000 reach-and-grasp trajectories have been generated with Quality-D
ameters, as for example the grasp type and final posture or the grasp speed. Therefore, it is difficult to control these grasp chara
osal to promote temporal consistency. Additionally, we present two mechanisms to improve the efficiency of our sampling strate
to a broader class of gripping mechanisms and address the limitations that had led to its obscurity. Two Fractal Fingers, coupl
e robot must compute a collision-free grasp trajectory while taking into account the geometry of the target object. Most grasp d
n multiple objects enabling robust dataset collection with no human intervention. We also present a standardized state machine
 deployment data. Nonetheless, existing methods typically require constant human supervision and preemptive feedback, limitin
InterRep, a novel interaction representation method that possesses not only the strengths of pre-trained models, known for the
ypically restricted to simulated or heavily engineered environments since agent exploration in the real-world without the guidanc
ning. SE(3) equivariance ensures that the sampled grasp poses are equivariant to the object's pose changes, enhancing robust
s to thousands of training samples. We demonstrate that our representation performs well on a real robot and achieves a succe
grained contact guidance for each fingertip. In particular, we employ a generative model to predict separate contact maps for ea
urfaces. This paper studies regrasping control to enhance suction cup grasping performance on PCBs, addressing issues arisi
ons on objects that need to be grasped, especially when objects are partially visible. In this work, we leverage human-like objec
his issue, we classify the observed performance into qualitatively ordered outcomes, thereby stratifying the results. We then sh

nterference between the grasped object and obstacles. To achieve the three modes using a single motor, the developed grippe
rcome the complexity and limitations of finger surface sensors, this paper proposes a force-controlled two-fingered gripper that
utilizes 3D-printed infill structures (periodic cells) with different infill densities to represent regions with different equivalent mech
m allowed initial impact velocity are identified. Results show that by properly optimizing these parameters, robot picking of rigid o
librium. We also show that for a given maximum acceleration, there is an optimal normal force for transport. To test our analytic
s a memory alloy skeleton and a water-cooled recirculation system, leading to a substantial reduction in the time required for ea
rnal magnetic field, facilitating the manipulation of small objects in confined spaces. Modeling and analysis of the magnetic grip
or vision-based error compensation, vacuum level monitoring, part contact detection, and detection of abnormal vibrations. We
head. By contrast, this paper presents a generalized approach to destructive disassembly planning that can automatically deriv
hand are developed. These analyses remarkably show that under very mild constraints on the hand design, the hand is complia
ners in narrow shelves). The DSG can delicately grasp objects of various shapes by employing two scoop-shaped fingertips tha
re, this underlying representation renders a co-design space amenable to gradient-based optimization. For efficient search, we
ufficiency region. Specifically, friction appears to enlarge horizontal error tolerance, while joint stiffness and pad damping allow
of in-hand manipulation by rotating its two constituent sticks relative to each other and offers control of objects in all three axes
ing paradigms. Studies on human manipulation point to the criticality of low-level tactile feedback in performing everyday dexte
iring human demonstration and without programming. However, owing to hardware limitations, its grasping performance remain
allest items (down to 5 cm in size) but cannot provide the shear force needed for handling bulky items. Gecko-inspired adhesio
zed Gaussian mixture models and probabilistic movement primitives to address these limitations and satisfy all the aforementio
s. Comparative studies show that our model has resulted in a significant improvement of the recognition accuracy by 0.0223, 0.
encounters difficulties in measuring human-finger's joint angles. While the joint-space is important in many scenarios, such as te
limb loss validated the system, accomplishing an object discrimination task in two different modalities (with and without the feed
challenge. In this work, we develop a tactile feedback board in the size of a smartphone with only five actuators, and achieve th
nment is augmented by the introduction of high-level haptic interfaces for robotic assistance that aim at re-directing the motion
r jamming to modulate its stiffness and provide feedback to users during teleoperation tasks. We characterize the performance
ctile sensation. The sensory drive voltage threshold will decrease with future refinement of mechanical impedance matching an
ustive data sets even for fairly simple objects. Active learning provides a mechanism for focusing on only the most informative a
ons. We then generate actuation instructions based on the circular spline initialization to drive the physical display. We demons
nstruction. We propose an incremental mesh extraction algorithm to achieve lossless fusion of streaming incremental mesh data
ossible. This paper introduces a novel approach to transparency control based on a human-in-the-loop velocity optimisation fram
pre-processing blocks. These include two indirect Kalman filters and an additional LSTM network, which together enhance the i
MG signals, combining Non-Negative Matrix Factorization (NMF) with Deep Neural Networks (DNN) in order to both avoid explic
r transferring the psychological principle that "humans act coefficiently as a group" (i.e. simultaneously maximising the benefits
he improvement of these datasets a substantial research challenge. Consequently, several issues have been identified in the an
nfidence in the eventual success of the task. The confidence level is computed by estimating the variance of the return from the
Baseline study (21 participants) and the Training study (16 participants). Our work is of interest to sUAV operators, regulators,
ction to the robot motion and the partial observability of the human fatigue state. We model the problem using continuous-state
e latent variational autoencoder (VAE) representations of trajectory segments (sequences of state-action pairs). Our approach
nerate enough demonstration data to achieve satisfactory task performance. Interactive approaches allow for humans to provid
ystem enables efficient, affordable, and safe collection of long-tail pedestrian-vehicle interaction data. Using our proposed simu
hance of being selected, which is unrealistic based on our findings. To address this problem, we design PERCEIVE, a human p
tory is generated as the expert reference, and a dense flight corridor is constructed to provide sufficiently large safe region. In t
ernal loading in the force reference. The adaptive compensation of the external loads is achieved through an electromyography
displacement relationship of gel deformation and combine the model with the highly impedance motion of robots to design an e
botic system is ensured. The paper proposed two approaches to ensure the passivity of the system when subjected to workspa
inexperienced users. In this paper, we present a framework that uses a tablet as augmented reality (AR) device to modify a rob
man-robot collaboration. Therefore, in this paper, we introduce FlowMNO, an Optical Flow-Integrated Markov Neural Operator
nt a significant challenge for these data-driven methods. To overcome this issue, we propose modifications to the learning proc
, we conduct a novel study that recreates a COVID-19 pandemic grocery shopping scenario and asks participants to select amo
urther complexity. Conventional state estimation approaches can be adapted to handle sparsity and even mobility. However, the
gent reinforcement learning with local coordination for individual diversity (MAC-ID), which can synthesize diverse pedestrian m
o-conditioned prediction remains challenging in downstream planning due to the complex nature of neural networks, limiting the
stacles are extracted as nodes and their relationships are built as edges. Three contributions are presented. 1) By assembling d
Concretely, we propose two methods - one based on evolutionary optimization and another on joint trajectory-and-environment
he human demonstration from the AR environment and contrastively learns the grasping strategy from the demonstration. For
e scenario is used to stimulate various human trust levels when working with robots. To enhance recognition performance, the
in distributions in a high-dimensional space via kernel two-sample testing between two sets extracted from multimodal data (e.g
n group preferences for complex collaborative robotic tasks (cobotics). We develop a sparse optimization formulation to introdu
-confidence from experience using semantic action embeddings. Using an assistive cooking setting, we show that the semantic
h the curtains to deliver medicines to me?”, to bounding boxes (e.g., curtains) with action-aware attributes. They can be use
owever, factors like small electrode size and unstable electrode-skin contacts make HD-sEMG susceptible to pixel electrode dr
nces during the task. The proposed approach was experimentally validated through user studies. Results indicate that the prop
ent. We show that by only focusing on team performance, the robot may resort to the reverse psychology trick, which poses a s
ole in allowing Haru to obtain a better performance via our method than human player's demonstrations. Finally, results of our
back with people’s desired movements, are proposed for this challenge. On the basis of these studies, we propose a syste
operative control with the rider using the divergent component of motion is also proposed. The results of an extremely low-spee
arm and an assistive cutting board, inspired by how individuals with one functional hand do meal preparation. Our system incor
ime windows and generated their condensed latent representations to predict multiple sequences of the future system states. T
ve this problem, we proposed a method that creates a map as a topological map from a textual path and automatically creates a
ets, which then guides a probabilistic target search planner. Operators provide a set of diverse inputs, including priority definitio
t interaction system, designed to augment the visual inspection performance through mixed reality. Through holograms from a m
2 (ROS2) and Bluesky to provide extensible robotic applications, while working under the operational constraints of a large-sca
aires. We performed two pilot tests of our framework with human-robot collabotation work cells at two test sites as part of the D
muscle activity of the forearm at the wrist joint while analyzing sixteen surface electromyography (sEMG) sensors. The study ex
ensing, i.e., perceiving the environment from the outside without rearranging it for dense sensing. Therefore, in this paper, we in
econd approach. These controllers are linearly combined to shift the controller behavior between pure kinematic state-depende
l realization, and report on a user study of 80 participants to learn what people might want of and expect from robot rooms, the
large cut pieces and the actual draping process in close proximity to the robot. The data is used to train and compare different
get person's velocity and direction to coordinate the motion between them. The inner-loop impedance controller is employed to
atically tuned MPC shared controller with respect to a hand-tuned baseline version as well as its generalization ability.
ns and robots. To specify the desired behaviour, we propose to use syntactically co-safe linear distribution temporal logic (scLD
m of optimally choosing a collaborative robot's behavior based on a conditional model of the human that depends on the robot's
an online teaming assistance context and fails to account for the real-life probabilities of potential team beliefs. We propose a n
h other HRC assembly approaches.
mapping process: mobile robots, human operators, and AR head-mounted display (AR-HMD). A 3D complementary mapping p
based on GANs that utilizes a recurrent, LSTM-based architecture for intention recognition in robotics. The resulting networks p
arative learning to learn a task, thus eliminating the need for programming expertise or long development times. The proposed
ch to address together these two issues is proposed. Starting from a desired final configuration, the framework plans a collision
bimanual peg transfer procedure that alternates manual and autonomous control correspondingly. The authority allocation in th
m's capabilities with respect to underwater communication, we perform a substantial in-person human study with 12 participant
trajectories, and additionally incorporates obstacle avoidance. Additionally, the paper addresses shared autonomy (SA) which
a case, a trained caregiver’s policy may not work for different care-receivers. To alleviate this issue, we propose a framew
mpetitive generalization ability compared to inverse reinforcement learning and conventional GAIL. We then propose a priority-c
ences to low-level adjustments to skill parameters. In this work, we present Distillation and Retrieval of Online Corrections (DRO
ence operational intuition. A Virtual Fixture spatial guidance framework with intuition maintenance is proposed. This novel guida
of a kinematics-based index. The resulting motion is then integrated into the learned task trajectories. The framework is implem
tions with fast convergence, subsequently enhancing the policy with a potential field method. The control authority is allocated s
approach ensures stability without relying on energy tanks, as demonstrated in our prior research. In this paper, we further ext
(NL). However, the translation of NL information into constrained robotics domains remains a challenge. In this paper, we use L
policies. In this work, we investigate how to generate multi-agent strategy explanations for human-robot collaboration. We formu
ying with safety regulations, this work introduces a novel optimal nonlinear control problem. It prioritizes path preservation, expl
physical contact to enter physical HRC, in which he/she can hand guide the robot or take over its gripped object. First, human s
nner which, by exploiting the variability of the robot's inertia as a function of its configuration, is able to return trajectories that ca
n meeting, we introduce a novel zero-shot VLN framework. Within this framework, large models possessing distinct abilities are
and memory-based cognitive processes. Indeed, the artificial personality manifests itself not only in various aspects of the beh
y agglomerated steps, key pieces of information are lost in the process. We amplify the initial definition of preferences to accou
RIDA's major weakness, our system uses pre-trained text-to-image models; however, pre-trained models in the context of real-
MateViT to perform pixel-wise semantic segmentation, simultaneously recognizing both objects and materials. Our methods ac
dynamic obstacles through an obstacle avoidance unit and a novel human guidance planner. With this aim, we also present a l
nize MP libraries. To showcase their applicability, we provide explicit formulations of the spatial operations for libraries compose
a simplified kinetic motion applied about a specific point on the body. Following the characterization of post-fall recoveries, we
limb using information about human body shape and pose that only needs to be provided at run-time. We show how this appro
robot involves a task failure, this will disproportionately negatively alter their perception of the robot's competence, as well as t
MoDs), which encode spatial motion patterns learned from (possibly scattered and partial) past observations of motion in the en
 ning develops methods that can learn and adapt over time, most of these methods are tested in the narrow context of object cla
stimate the complete version of a portrait drawing, after the human user has drawn some initial strokes. In this way, the robot c
t is required before they can be adopted for use in the ‘real world’. One key challenge on that journey is the user training
we conduct extensive experiments to evaluate the model and system. Experimental results show that our new method and syst
contours. Then, expressive digital strokes were generated after interactive editing according to user's aesthetic understanding.
common reference frame. Visualizations in the mixed-reality device then allow operators to see a virtual life-sized representatio
duce dimensionality. To improve planning reliability with the learned EED, we design a control space based on a walking trajec
shape, walking gait speed variation, and energy expense. The model is validated by extensive foot intrusion experiments with a
ns on the mixed uneven terrain consisting of varying slopes and steps. The results suggest that the controller shows comparab
mponent is integrated with Reinforcement Learning (RL) to guide the control policy to produce behaviors aligned with the data d
ible whole-body template behaviors, which can provide reliable warm-starts for NC-MPC, thereby greatly reducing the computa
bio-plausible reward modulated spike timing-dependent plasticity learning algorithm. By monitoring the spikes from the motor n
evelopment of an algorithm capable of detecting abrupt, incipient, and intermittent faults in full-sized robots, its implementation u
ucturally, the inner skeleton layer is composed of rigid framework and linear actuators for changing the global body dimensions
tributed, non-collocated force/torque sensors, thus being a flexible framework with respect to the underlying robot sensor suit. T
operative perception using a novel Vision Transformer, named as S2R-ViT, which considers both the Deployment Gap and Fea
onflicts Fusion (ECFusion) method to explicitly eliminate the extrinsic/inherent conflicts in BEV space and produce improved m
om limited communication bandwidth. To address these issues, we propose a novel camera-based 3D detection framework for
ormer-based architecture to classify visual intentions in image sequences of tracked traffic participants. We use a vision transfo
re distillation(CFD). This strategy combines feature distillation from merged clean-fog features with contrastive learning, ensurin
posed method comprises multiple spatial and channel-wise attention blocks available to learn global representations across spa
significantly promotes the development of LiDAR networks. However, current research operates under an ideal situation where
wever, the inference speed, crucial for running on an autonomous vehicle, is neglected. To this end, a new method, dubbed Fa
man pose estimation, named LPFormer, which uses only LiDAR as its input along with its corresponding 3D annotations. LPFo
zation of a small backbone. These challenges, in turn, hinder the optimization of the framework and yield inferior prediction resu
lifecycle management more challenging. Additionally, occasional false-negative detections due to difficult scenarios like occlus
nce decline in multi-agent perception. In this paper, we thoroughly examine the impact of the distribution gap on existing multi-a
ce, it becomes imperative to investigate whether specific GPS signals can easily mislead the multiagent perception system. To a
erging end-to-end paradigm, that tightly integrates the perception and forecasting architectures into joint training, promises to so
ery stream flow among agents. The cross-agent queries are interacted via fusion for co-aware instances and complementation
knowledge. Subsequently, our method integrates the fog-induced noise distribution inferred from the LiDAR imaging model with
e two sensors would lead to near-duplicate detections, making trivial improvement to cooperative task while causing unnecessa
tasets, such as V2V4Real, provide 3D multi-object cooperative tracking benchmarks. However, their proposed methods mainly
sent during training. This causality gap typically emerges in robotics applications such as autonomous driving, where a network
environment difference. To this end, we present Hierarchical Human-to-Robot Imitation Learning (H2RIL) that learns the mappin
ng ahead (POLA) to mitigate the gap. Concretely, we optimize the policy on both current and future states where the future stat
rld experiments on everyday tasks, we show that exploiting both the image-level and pixel-level properties of visual foundation
d "how" to do it. A powerful way to encode both the "what" and the "how" is to infer a well-shaped reward function for reinforcem
e reward, and many rewards may result in the same behavior in the downstream tasks. Instead, we show that it is possible to op
ation. The developed framework comprises two key components (i.e., behavior cloning to learn human unfolding policy and lea
nsorRT optimization for specific cutting-edge hardware devices has been also explored. For the comparison, offline metrics suc
e lower bounds or analytically exact estimates of escape energy for a very restricted class of known objects with low-dimension
essed by a linear approximation. Then, the input to generate the desired rolling orbit is designed. Next, as an example of utilizin
designed to provide 2-degree-of-freedom actuation, driving the f/e and a/a motions of the proximal joint simultaneously. Based o
y remove the actor’s features, which may still cause difficulty setting up the physical system, while large feature reduction m
rds these action targets in task or joint space. However, straightline motion in these spaces for the most part do not capture the
n cone assuming all parameters are precise. Then, a robust planning method refines the motion cone to maintain desired conta
demonstrate improved sample efficiency using a framework that guides exploration toward relevant state space by following ac
s integration of both sensory inputs, offering richer spatial information and facilitating better reasoning about robot actions. The m
hen the internal dynamics of the manipulated object are unknown, such as when it is filled with liquid moving during manipulatio
lel mechanism. The climbing motion is made by attaching and releasing the perimeter and middle cups one after another. The
ttempts. The CG-CNN is refined using this simulated dataset and enhanced through domain randomization techniques. Subseq
ntinuous execution of trials, a straightforward generalization of the learned data to different execution speeds, and learning from
zation of robot programs, and an automatic planning and execution pipeline for force-controlled robotic surface treatment. Robo
acilitating coordinated navigation among Heterogeneous Vehicles (HVs). To address these issues, we propose the LLM-driven
allenges, we present a novel low-light image enhancement model, termed Spatial Consistency Retinex Network (SCRNet), whi
measurements coupled with a U-Net based tool tracking to adapt the complete robotic model during task execution. As a proof-o
erature with a novel greedy algorithm and multi-robot extension. We close with numerical results for a police district in downtow
a teleoperator in the loop makes the system widely applicable to unstructured environments. A human can take care, with little
erived sample complexity bounds to certify the performance of a feedback policy without making assumptions about the system
asteful since only certain spatio-temporal regions are reachable and relevant to the self-driving vehicle. We present a unified, in
which directly provides safe input, and can be regarded as a real-time safety filter. The input expressivity is also enhanced by le
with over 20,000 km of historical data, the novel proposed NVFormer implicitly models the relationship between vehicle dynam
obustified problem. Additionally, a strategy for selecting an alternative terminal state to reach is introduced, which comes into pl
hms. Our method leverages a neural network based multimodal predictor to generate multiple trajectory proposals for the auton
much fewer samples compared to existing reinforcement learning algorithms. In addition, our methodology streamlines complex
ents of dynamic autonomous driving of bots around a track are provided to show that the proposed o-MPPI is more efficient and
ences that decrease the cost of physics-based simulation; and iii) LazyBoE, a robust kinodynamic planner that leverages these
mptotically optimal RRT* method, as well as a relevant reactive method and provides satisfactory performance, closely matching
vely selecting trajectories to learn a model precondition for an inaccurate pre-specified dynamics model. Our proposed techniqu
hm which plans trajectories while anticipating future observations. In particular, we introduce a model which learns to predict the
e structure using a reactive planner and learn specific parameters with Bayesian optimization. The method is evaluated on a se
nces. Treating existing motion planners as either a closed or open box, we present a variety of ways that Motion Memory can c
e states of surrounding vehicles can be a nuisance shortcut. In our work, an off-policy approach is proposed to alleviate the iss
n techniques. However, these methods often overlook the important multimodal property of the optimal paths in the training set
an MPC method to accomplish long-horizon manipulation tasks by dynamically specifying sequences of subgoals for the MPC
he bandit correspond to sets of possible transitions. High-reward regions are identified by clustering transitions from sequential
ondence-free alternative: The user simply measures a few points from the fixture's surface, and our method provides a tight su
proposes a generic prognostic tool for probabilistic condition monitoring of mechatronic systems, with the aim to improve the pro
through performance circuitry, DR.J provides an alternative to transporting equipment to in-person competitions and serves as
del transformation is assigned as a hybrid exponential– polynomial function whose coefficients are conveniently calculable wi
age and generates a stitching path alongside the seam line, ensuring a consistent distance. The sewing machine then accurate
 ng the workpiece surface are used to optimize the surface finishing process and ensure the coverage of concave areas. Theref
heduling, and product quality. To demonstrate the effectiveness of this approach, a control problem is solved in the Decentralize
This paper presents a minimalistic safety fallback mechanism when offloading mobile robot navigation to the edge, that ensure
t-Based Search(CBS) theory. Firstly, HDVs are accurately modeled as constraints to enable usage of CBS in the mixed traffic.
act of divergence factors from transmitted models during the local learning process of each silo. We also analyze the effects of
, and each ODD query matches only a small number of scenarios in the database (0.2% in our case study), rendering most of t
cle focuses on the safety-critical context of autonomous vehicles (AVs) merging into traffic consist- ing of a mixture of AVs and
k with LSTM-based social pooling to predict the trajectories of surrounding vehicles. These trajectories are used to compute for
approach 1) dynamically predicts agents' goals irrespective of particular road structures, such as lanes, allowing the method to
s with the travel efficiency improved and the hard safety constraints satisfied. Specifically, all constraints involved are addresse
racteristics in the simulation environment. We propose a socially-controllable behavior generation (SCBG) model for this purpos
style. In this paper, we present a model of the driver's decision-making process that explicitly captures their perception of the dr
t DNN penalize all misdetections equally irrespective of their importance for the safe driving task. Hence, to mitigate the occurre
ace, which prevents the model from capturing sufficient information from points in different distance ranges and reduces the mo
l built with Transformers, that can consistently track objects in the scenes while also predicting their states accurately. STT con
detrimental effects on collaborative perception. In this paper, we introduce SmartCooper, an adaptive collaborative perception fr
it outperforms the learned policy alone by up to 20% and a plain evolutionary algorithm approach by up to 53% on realistic ben
ture of the UDE is its ability to convert the robust control problem into a low-pass filter design in the frequency domain, which g
atically profile autonomous vehicles and differentiate those from human-driven ones is a must. In this paper, we present a fully-f
nnovative solution tailored for radar-based tracking. Bypassing the typical reliance on specific object types and 3D bounding bo
omain expertise and hand engineering for each road network’s observation space. Additionally, precise observations use gl
autonomous driving by combining basic driving imitation learning with LLMs based on multi-modality prompt tokens. Instead o
a query-centric Transformer model that performs efficient ego-conditioned motion prediction. For planning cost, we propose a
distance-based metric alone is insufficient to select relevant agents and accurately perform their predictions. To resolve these i
ce influences human drivers' behavior, in the Waymo Open Motion Dataset (WOMD), and we use these labels to perturb the da
locity bounds on occluded traffic participants. Compared to state-of-the-art methods, we achieved velocity increases in more th
Multi-residual Task Learning (MRTL), a generic learning framework based on multi-task learning that, for a set of task scenarios
pproaches are well-suited for learning cooperative multiagent policies with considerations for future demand, applying such me
eratively improve AV models. Moreover, it remains intractable and challenging to filter through gigantic scenario libraries collec
ovel domain-agnostic learning method for autonomous steering based on our analysis of training data, network architecture, an
roblems, we propose Map-Agent Coupled Transformer (MacFormer) for real-time and robust trajectory prediction. Our framewo
unction that penalizes lane changes and deviations from a desired velocity profile. Finally, we generate a reference trajectory in
NDT). The structure of eMem-NDT is constructed by hierarchically clustering the text embedding of vehicle behavior descriptio
urrounding objects, which reduces the flexibility and generality of the system over different scenarios. Additionally, the triggering
d on Model Predictive Path-Integral control to generate interactive trajectories. This framework incorporates a Bayesian inferenc
d asses them with a situation-aware reward design, leading to a dynamic reward that allows the agent to learn situations that re
licy. Furthermore, effective curricula are designed and integrated into an episodic reinforcement learning (RL) framework with p
mimics the dynamic allocation of attention human drivers exhibit based on factors like spatial orientation, proximity, and driving
tal context information and the dynamic agents' states in a unified way. By utilizing features extracted from semantic maps duri
studies addressed the inter-subject variability of sEMG signals, which limits the generalization capability of the intention estima
on history. Taking arm dynamics into account allows RT to capture the operator's intent better than other SOTA models that onl
 ion composed of the tangent space projection of time-windowed sample covariance matrices from EEG filtered signals and der
ptical flow consistency to model the extracted expression features as the optical flow field between the neutral face and microex
approach explicitly extracts object and road-level features from external camera data, which we demonstrate are important feat
h conversational interactions.We leverage the 3D simulator AI2Thor to create household query scenarios at scale, and augmen

s of different approaches. The scenario-based analysis highlights the importance of using multimodal sources of information an
3D vector field method is introduced for hand trajectory representation, leveraging minimum jerk trajectory predictions to disce
e impact energy. Our gripper simultaneously resolves the issue of smooth absorption of the impact energy and fast closure of th
ntagonistic robotic shoulder with two degrees of freedom powered by self-sensing hydraulically amplified self-healing electrosta
ever beams with adjustable supports, main cam mechanism, and symmetric support positioning architecture utilizing an Archim
which has rigid links and low-stiffness joints can perform friction-based cleaning of surfaces with complex geometries while mai
based on both the interval analysis method and the commonly used floating-point optimization algorithm are conducted under s
acement transmission is obtained using the two parameters related to the curvature of the transmission path: total bending ang
ew knowledge about robots composed of $n$ spacer disks. We show the existence of multiple robot configurations with differe

n cables, is utilized to conduct experimental tests under varying parameters. This work adopts an integrated approach of experi
o joint-specific parameters during encoding, enabling a stable optimization process. These parameters are subsequently proces
and water is established based on the modified Wagner theory with consideration of water surface deformation. Particularly, th
ossible to calculate a trajectory that leads to minimum progress along the path. This insight allows us to control the traversal of
segments, the redundancy of the inverse kinematics includes tips that lie on a sphere providing a closed-form solution to the inv
mber of solutions is not constant, with up to seven valid CSC path solutions even in non-singular regions. An implementation of
radients thereby rendering it unsuitable for online numerical optimization. In this letter, we propose a novel performance index u
m is a measure of its ability to follow recorded human arm and hand paths. The optimization is conducted using a modified varia
perties. This representation consists of a geometric progress rate and a third-order trajectory-shape descriptor. Concepts from s
can be built into closed-form kinematic deformation fields before even considering other aspects of constitutive modeling. We
epends on the robot's configuration, the robot speed and the properties of the environment. An approach to achieve the desired
nvironments, with a focus on system-level scalability and redundancy. We describe the design objectives of RASCAL and how t
ddle link as a 2-DOF virtual passive joint, a new trajectory planning method in the 4-DOF virtual passive-joint space is proposed
g the stability, it is then necessary to check two conditions: The Legendre-Clebsch and the Jacobi conditions, both well known i
h cables become slack, estimates the tension distribution among cables and computes unwanted MP motions, allowing to pred
y. In this study, new approaches to analyzing and commanding CDPRs by considering the coupling effect are proposed. First,
ue capabilities. The paper discusses the use of the Direct Collocation methodology for optimizing trajectories for the robot’s
trol with a Gaussian process data-driven robot dynamic model. Under the new enhanced EIC-based control, the stability and p
nservativeness inside the energy function, while keeping the promising provable safety guarantees. Specifically, we quantify th
unified approach to address these limitations with Adaptive Multi-step Control Barrier Function (AM-CBF), where we paramete
n skills, where base skills aim at completing tasks and failure prevention skills aim at reducing the risk of failures to occur. Then
predictions. More importantly, the resulting models are constrained to specific data formats and modalities. Moreover, connect
. We verified the interpretability and robustness of the task of inserting a furniture part, which consists of the ``approach" phase
ese effects into account. In this study, we propose a method for constructing a neural network that describes the mutual relation
en intuitively transferable to a new context, its probability distribution might change, requiring data-intensive relearning. In this w
art performance on standard QD benchmarks. However, CMA-MAE cannot scale to modern neural network controllers due to it
 ys its images on a monitor. Especially, regarding issue (2), most prior research has assumed the availability of depth images or
dynamical systems and ensuring that they are globally asymptotically stable. Hence, we introduce a novel Contrastive Learnin
onstrations include varied target positions but leave various obstacles alone. These MPs can not adapt the learned movements
k focusing on the structural relevance and the type of objects and goals, relying on an external primitive to move the robot to ac
n process and detects segmentation points based on the differentiated 6-axis force/torque signals obtained in the task demons
mutating from the original source. This can be beneficial when better skills might emerge or when a new skill is obtained to be u
ation encoder and a motion generation decoder. This enables us to pre-train the first using a large amount of unsupervised data
n probability in the complex environments. Firstly, we retrieve the nonconvex obstacles with arbitrary probabilistic uncertainties
ons carry no notion of which behaviors should be explicitly avoided. In this work, we propose Control Barrier Transformer (Con
nal knowledge and enable robots to reach goals in a zero-shot method by applying the rules independently and incrementally. H
™s weight, but this mechanism cannot compensate for varying payloads, as adjustable passive gravity compensation is relative
ation strategy, across six diverse tasks, showcasing its applicability in various environments. Notably, our method significantly
uld require a large number of trials and professional surgeons, which is generally impractical to arrange. Instead, we show how
the bell-shaped velocity profiles that are typical of human movement. As the novel formulation is more expressive, it is able to
onsuming, we further leverage AirExo to collect cheap in-the-wild demonstrations at scale. Under our in-the-wild learning frame
oposed approach are that it 1) infers constraints without calculating a new policy at each iteration, 2) uses a simple and more re
d emph{Diagrammatic Teaching}. Diagrammatic Teaching aims to teach robots novel skills by prompting the user to sketch out
se a method that retains the provably translation-invariant and relational properties of prior work but incorporates additional pro
eural architecture that enables the joint training of the policy and its associated Lyapunov candidate to ensure global stability th
rent from previous approaches in the literature, we propose to use polynomial representations for DSs, which yields an optimiza
time distribution. However, this approach is only effective when the domain randomization encompasses the actual shifts in the
ing a quadratic program that guarantees stability and safety using control Lyapunov functions and control barrier functions, res
al insights from demonstration data but also demonstrates the capacity to learn a candidate Lyapunov energy function that is co
eural energy function, which is essential for it to be a cogent stability certificate. Then, the neural energy function is learned to c
TCL additively decomposes the overall reward into a task reward and its residual as soft constraints, maximizing policy divergen
ramework is to only extract the comprehensive features of human motion from unreliable demonstrations and use them as a ro
ntroduce an offline model-based RL algorithm, IQL-TD-MPC, that extends the state-of-the-art Temporal Difference Learning for
by augmenting skill discovery rewards with additional rewards through a naive combination might fail to produce desired behav
d rate, depth of cut, and mechanical compliance during task execution. We show how a mathematical model of cutting mechani
base set of pre-training tasks: instruction relabeling via large language models and cross-trajectory skill chaining with offline re
ore efficient states, we propose a simple and effective vision-based RL method. It leverages the representations acquired throu
ts to learn highly dynamic behaviors from animal motion data. The learned controller is deployed on a quadrupedal robot and th
el Distillation for Sim-to-Real Transfer) to achieve efficient sim-to-real transfer of vision-based model-based RL using distillation
partially-observable pursuer policy depends on two key factors: the balance of diversity and optimality of the evader’s beha
nline fine-tuning phase. This sample inefficiency problem arises because the agent treats all state-action pairs in the replay buf
evel. We seek to develop a control policy that effectively breaks down the high dimensionality of snake robots while harnessing
asks. To evaluate the effectiveness of the proposed reward formulations, we conduct simulation experiments on CartPole, Ant,
nsuming process. To mitigate these challenges, our research theoretically analyzes the implications of data augmentation on M
orks yet can represent high-performing policies. We obtain multiple trained policies at the same time while maintaining sample e
pedal robot to efficiently learn to walk entirely in the real world within minutes and continue to improve with more training where
arameter tuning and mitigates the sim-to-reality gap by taking advantage of torque control's inherent compliance. Also, we acce
 , animals and humans can control their limbs separately. Inspired by this biological phenomenon, we propose a Decentralized m
luding our metadata system designed to further the utility of our data. Custom infrastructure allows end users to reconfigure the
images prior to training object detectors may not necessarily yield performance improvements. (3) The complex underwater en
ssian sampling (GS), thereby serving as optimal inputs for the mobility simulation. Subsequently, we introduce a co-simulation
o predict these high-fidelity features using input from lower-quality sensors. UNRealNet predicts dense, metric-space features d
zontal offset and angular deviation between them, thereby achieving accurate registration. To demonstrate the potential of our
thod utilizes DMVAEs to capture essential environmental information and robot properties, effectively modeling factors that influ
g framework is adapted to the three-dimensional space so that the planning result conforms to the geometric characteristics of
this work, we study trajectory-level data generation for multi-human or human-robot interaction scenarios and propose a learni
ental results demonstrate that the proposed approach can have improved performance in the multi-task training stage, and furt
i) a distance function to measure the reachability of subgoals. However, they struggle to scale to high-dimensional state space
earns to associate objects with keypoint coordinates and discovers a graph representation of the system as a set of keypoint em
. We evaluate our framework and related intrinsic learning methods in an environment challenged with exploration, and a comp
ories. However, this approach is usually combined with on-policy reinforcement learning (RL) and suffers from low sample effic
d multi-modal data and output picking and placing action. Besides, we have applied the visible connectivity graph to tackle nonli
ct pose information, TAPG facilitates guided, yet adaptive, learning of a sensorimotor policy, based on object segmentation. We
d manipulation tasks, we show that the structure included in CIPs substantially improves the efficiency of motor skill learning. W
rget prediction followed by planning, but seldom account for highly non-linear reactive behaviors. Alternatively, Reinforcement L
as well as limitations to overall performance. In response, we propose a novel strategy: using a policy gradient algorithm to find
od. It leverages a task-evoked planner(TEP) and a reinforcement learning approach with planner's guidance. TEP utilizes reusa
obotic furniture assembly problem in a more generalizable way, facilitating seamless chaining of skills for this long-horizon task
g simulation introduces the Sim2Real transfer problem. To mitigate this problem, we study different tactile representations and
gement tasks using a task state graph that captured spatial relationships and a cooperative cost matrix that provided details abo
construction. EquivAct is trained in two phases. We first pre-train a SIM(3)-equivariant visual representation on simulated scen
and propose a novel adversarial training framework for robust control. Our approach generates aggressive attacks on the state
rate global models. To address these challenges, this paper presents MultiAC6: a new multi Actor-Critic framework for robot ac
SLAM designs a gated generative replay to control the robot’s learning behavior based on the feedback rewards. Specificall
act intermediate representation is therefore preferred for dimension reduction. This work proposes an efficient HRL-based fram
constraints on the reward terms. In line with previous work, further control of the diversity level can be achieved through an attr
iculum learning for local navigation maneuvers. Due to the significant computational demands of collision detection in whole-bo
nition and metric localization. RaLF is composed of radar and LiDAR feature encoders, a place recognition head that generates
sparse voxel feature grid created by an octree and multi-resolution permutohedral tetrahedral feature grids to represent the sce
mance when facing complex wireless environments. However, the price to pay is the astronomical amount of training samples, a
s problem, we propose two quantitative metrics based on laser scans, which explicitly represent environmental complexity and s
vision networks can accomplish the same without any fine-tuning. Specifically, we use Masked Autoencoders, pre-trained on st
s a safety shield to provably guarantee human safety. This bridges a critical gap between theoretic RL research and its real-wo
e to their homogeneity. To tackle this problem, we introduce an efficient method that enhances agent diversity within a single po
ioritizing safety or efficiency depending on the current observation. The visual attention map is used to provide spatial attention
d step addresses the constraint violation by projecting the policy back onto the constraint set. Such a projection-based optimizat
igh-dimensional or complex environments, where DRL methods are prone to local optima and fail to explore regions of the stat
is the scalability problem. As the number of objects increases, the computation time of traditional offline-search-based methods
ssue. Large-capacity models, e.g. decision transformers (DT), have been proven to be competent in offline policy learning. How
The learning of a control policy is performed purely on the physical system using model-based reinforcement learning, where th
f nodes and edges with abstract but essential information and are less influenced by the scene structures. However, most exis
just LiDAR measurements to align with the underlying support surface. The SPF is trained in a semi-self-supervised manner an
o learn effective locomotion policies. Inspired by Bernstein’s neurophysiological theory of animal motor learning, we develop
erence motion. Then, we refine these policies using deep reinforcement learning. Our approach enables the acquisition of com
sed approach also enables efficient generalization to different control tasks while also reducing the number of evaluations requi
cies. By reusing the same composer policy, our approach enables adding new experts without affecting existing ones, enabling
e and front limb movements while balancing on two hind feet. The policy is commanded by a high-level motion generator that g
2real gap. We develop LAnguage-Guided mOtion cONtrol (LAGOON), a multi-phase reinforcement learning (RL) method to ge
m joint masking strategy for simulating impaired joint scenarios, 2) a joint state estimator to predict an implicit status of current jo
of serious physical injuries. This paper presents a hierarchical fault-tolerant control scheme employing a multi-task training arc
els of pre-learned skills by using multiplicative skill composition and residual action learning. We also introduce a goal synthesis
duce an explicit recovery stage where tracking rewards are given irrespective of the motions generated by the control policy. T
e MPC using imitation learning to make the controller learnable. Finally, we leverage deep reinforcement learning with limited e
MPC's strength in predictive capabilities and RL's adeptness in drawing from past experiences.Unlike traditional locomotion con
without significantly changing their underlying biology. In this paper, we take a similar approach to developing robot parkour on
ution to account for uncertainty in the robot's interaction with the environment. The value distribution is consumed by a risk metr
consider a novel risk-sensitive perspective to enhance the robustness of legged locomotion. Specifically, we employ a distribut
adopt the Extrinsic-Intrinsic Policy Optimization (EIPO) framework. The key idea of EIPO is to establish a constrained optimiza
onstraints. The cyclic gait patterns resulted from existing RL-based methods to learn parameterized locomotion skills character
ation, making them susceptible to failures due to inappropriate recovery strategies or insufficient robustness. To address this is
e seamless application of these strategies improving push disturbance rejection performance. The proposed controller is valida
nthesizing nominal walking gaits using a simulation-in-the-loop learning approach. Further, we leverage discrete-time barrier fun
gth control as an alternative to center of mass height control, and rolling contact for heel-to-toe motion of the stance foot. We the
able center-of-mass (CoM) trajectories for running and jumping with arbitrary contact sequences and time parametrizations. Th
ic state selections or the use of template models for gait planning. The results demonstrate that the learned latent variables are
1kHz. This ensures temporal continuity of the estimated state and makes it practical for control purposes of biped robots. We a
d by the new leg design and handle unmodeled payloads up to 90kg while operating at high efficiency. The approach’s effic
e robot is capable of hopping with TWR
es policies that exhibit emergent gait preferences, while using a reduced set of simple rewards, and without prescribing a specif

nfluenced by the distance of the CoM to the support area formed by diagonal footholds. In this context, lateral spinal flexion, wh
nce trajectory online using a general-purpose model-predictive controller on robot hardware. Traditional motion imitation for leg
negotiate obstacles under static and quasi-static locomotion and regular joint torque demands. The design of the mechanism is
e control the leg motion for reorientation using Deep Reinforcement Learning policies. In a simulation of Ceres' gravity (0.029g)
mplex learning frameworks, or mass, inertia, and dimension randomization, which leads to prolonged training periods. In our st
ably, our findings confirm that the most commonly observed sequence in both natural and robotic contexts indeed offers optima
we enable a dedicated pedipulation controller that is robust to disturbances, has a large workspace through whole-body behavi
ate the natural behavior of real animals. This approach significantly diminishes the requirement for substantial amounts of real i
ressive Moving Average Vector (ARMAV) model to construct the state error model of the quadruped robot using data. The pred
erconnected SRB dynamics: centralized and distributed MPCs. The distributed MPC assumes two local quadratic programs tha
arned with reliability. Furthermore, by utilizing realistic simulation data that captures whole-body dynamics, RL effectively overc
chieving this is designed based on the method of continuous-time output deadbeat control. Third, a typical steady wheel gait of
ajectory optimization is developed using a genetic algorithm that considers many aspects, including body and limb dynamics, lim
process, without requiring an accurate full-body model. When enhanced by a continuous learning mechanism, the proposed c

along with kinematic-inertial state estimates to build a dense but efficient map of bounded planar surfaces. We present novel al
it difficult for local feedback controllers to effectively recover from disturbances as the system evolves through underactuated d
uations of motion, considering both the robot and skateboard dynamics. By accounting for system physical constraints, an adva
c feasibility of the generated locomotion trajectories, we demonstrate that the experience planning heuristic offers an effective w
resence of actuator failures and unknown disturbances. Our simulated validations of the robot in water demonstrate the effectiv
contrast to the conventional velocity-based controllers, our approach builds upon a position-based formulation of the RL task, w
hieve efficient and stable walking. In this study, we proposed a model formed by tensegrity spine and rimless wheel to represen
ments. Specifically, the annotation can achieve high reprojection accuracy across all surrounding cameras and is spatial-tempo
ate the intrinsic and extrinsic of the sensors while addressing temporal synchronization. During recording, we cover multi-floor
Specifically, we propose a spatial-aware dynamic encoder, which can capture spatial information of the input and generate inp
asets, we proposed a Visual Distribution of Neuron Activations (VDNA) representation to represent datasets of images. This re

s work introduces a fresh perspective by framing place recognition as a regression problem, using camera field-of-view overlap
primarily focused on model-based methods, with limited attention given to data augmentation (DA) methods. Therefore, there is
PR algorithm. The pipeline involves dual-branch feature extraction and shared multi-modal feature fusion based on transformer
pose a novel Multi-modal BEV feature fusion method named MBFusion. Specifically, to solve the semantic misalignment proble
ng itself. To push the state of the art, this paper proposes the first-of-its-kind quantized visual-inertial odometry (QVIO) to offer
cameras' images into 3D semantic occupancy. Addressing the challenges presented by this transformation, we have specifically
nd globally elastic. Locally, the neural feature volume employs a sparse feature Octree and a small MLP to encode the signed d
sed features across challenging scenes and present an efficient method for matching and integrating these features from stereo
fM) procedure to both significantly reduce the computation time and to provide metric scale which is crucial for lidar depth loss.
and lightweight parameterization method to establish improved data association and achieve accurate localization at centimete
reases additional time costs. To tackle the problem, this paper proposes a novel method for online dynamic objects removal fo
on multi-view geometry, to simultaneously localize objects and implicitly learn their dense geometry. We create separate implici
-RT, using a hybrid of off-the-shelf ray-tracing GPUs and CPUs to substantially improve CPU-based PVM. OctoMap-RT employ
A curvature suppression cross- entropy(CSCE) loss is proposed to reduce the ambiguity of prediction on the boundary between
nce. Our method employs contrastive representation learning using both human driving data and instance-based segmentation
ent two strategies for estimating traversability, using a guide filter network (GFN) and footprint supervision module (FSM). The f
infer navigational patterns in real-world environments from partial observations only. We explain how geometric data augmenta
se a high-performance teacher and lightweight student distillation framework called TSCM. It exploits our devised cross-metric
eriments in simulated and real environments, we demonstrated that the 3D-BBS enabled accurate global localization with only a
ance Belief Update module for reverting false positives. We quantitatively evaluate our approach performance on SemanticKIT
nspired plug-and-play method to learn a semantic- agnostic and spatial-aware representation for generalizable visual-audio nav
field in a self-supervised manner, requiring complicated training procedures and loss function design, 3DIML leverages a two-p
ur preference for SOD over semantic segmentation sidesteps the intricacies of classifying a myriad of non-standardized urban
 ock maps and the corresponding switching strategies, ensuring that the robot can estimate the state in large-scale environmen
rent locations in cluttered conditions. Specifically, image processing techniques that involve background segmentation, energy
d approaches. Researchers have made attempts to utilize concurrent LiDAR and camera measurements in pursuit of precise s
oss different scenes. We effectively leverage intensity as an additional modality, using our new feature selection scheme that d
shing and iteratively updates the neighbor list of each particle over time. The neighbor list is then used to propagate the posteri
nd defective regions in a map. We also propose an initial sensor state estimation algorithm that robustly estimates the gravity di
(VO). The method extends recent advances in lidar descriptors and utilizes a novel double (similar and opposing) distance matr
s fine points are screened and furtherly their feature descriptors are enriched from 2D to 3D, which improves the match perform
for online outlier rejection that reduces the necessity for scene-specific fine-tuning while simultaneously improving the overall S
orks, using satellite images as the references for ground-level localization. However, most of these methods merely provide cro
in difficult regions and detrimentally affects scale alignment. To solve these problems, we introduces a feature-based visual od
rieved using a global representation. Our experiments, particularly on sequences captured on a planetary-analogous environme
even when dealing with outlier measurements, through graph optimization based on Graduated Non-Convexity (GNC). Addition
semantic-focused patch tokenizer and a multi-branch Mixer. To mitigate the inference from place-unrelated objects, the seman
essing, the keyframe point-cloud map is built using the accumulated point clouds to construct the frame-to-frame data associatio
s problem, VioLA makes use of a pre-trained text-to-image inpainting model paired with a depth completion model for filling in th
m perspective cameras. We introduce two different imagination modules which are inspired by how humans can comprehend an
e, we design Transformer based spatial and temporal fusion modules respectively to realize sufficient feature interaction among
ound salient image points for the second stage. While existing methods use independent networks to extract these features, on
only contain minimal information gain, which can hinder the regression accuracy or bring unnecessary computational costs wit
tion to this problem, we propose a novel Joint Image and Sequence Training protocol (JIST) that leverages large uncurated set
CCL, to tackle the catastrophic forgetting problem and generally improve the robustness of LPR approaches. Our CCL constru
trol outputs from convex Model Predictive Control optimization. The estimation is further refined through Gated Recurrent Units
SLAM. Leveraging intra- and inter-attention and graph neural network, P-GAT relates point clouds captured in nearby locations
erm changes --such as different colonoscopies of the same patient-- feature-based matching fails. To address this, we train on
alization onto actuation fields generated by an electromagnetic navigation system. The resulting magnetic field is measured usin
ded network to generate attention-guided descriptors. We then employ a fast affinity-based candidate selection process to iden
maps. Firstly, feature encoding is carried out on the original LiDAR point cloud map by generating offline heat point clouds, by w
ironment or rely on normals. Inspired by the recently proposed NeRF representations, we supervise the network by points sam
ditions. By further using sequential infor- mation, our LGPR system can observe richer underground scene contexts, and the as
s framework, a coarse (spatial correlation bound check) to fine (visual odometry consistency check) method is designed to sele
ency and inter-channel communication. Our introduced sequential group attention (SGA) module enhances computational effici
part of the registration, including semantic voxel downsampling, data association, adaptive local map, and dynamic vehicle rem
el APR-agnostic framework, HR-APR, that formulates uncertainty estimation as cosine similarity estimation between the query a
the minimal availability of supervised training data, which is just the corresponding 6 DoF poses of the images. In this paper, w
ph that represents the topology of key texts in the space and becomes the basis for discerning repetitiveness and symmetry. N
equence. In the offline phase, VOLoc compresses the Lidar maps using a emph{Geometry-Preserving Compressor} (GPC), in w
. Our proposed system, VPRTempo, overcomes slow training and inference times using an abstracted SNN that trades biologic
e findings, we propose an improved 17-point algorithm, which avoids using overlaps and derives a simple way to solve normal e
structured environments (urban, outdoors, indoors, aerial, underwater, and subterranean environments) without any re-training
uction method based on locality preserving projections (LPP) that can accommodate faster-to-compute binary features, and an
 thesize novel views, the proposed NeRF-VINS overcomes the limitations of traditional keyframe-based maps (with limited views
with a novel data association and matching scheme using object detection methods. We leverage the Invariant Extended Kalm
nonlinear observability analysis. We also evaluate the proposed algorithm extensively on diverse platforms, including flying UAV
ot increase the cost and time of deployment. To overcome the difficulty of localizing in sparse maps, the proposed approach fu
chronize these sensors to ensure that all data is recorded simultaneously. A camera, speaker, and transmit antenna are placed
observability analysis, for the observability under general motion and the unobservable directions arising from various special m
s on a LiDAR-enhanced visual prior map. Experiments show that the proposed algorithm can provide centimeter-level global po
method of fusing sensor data from a Doppler Velocity Log (DVL) and IMU using an adaptive nonlinear estimator to provide dead
luttered environments. The second FLS estimates smooth corrections to VIO to generate pose estimates at a high rate for onlin
dentify and highlight the most common real-world challenges in INS state estimation, including sensor self-calibration, robustne
ned from a low cost Inertial Measurement Unit (IMU). For this purpose, we introduce a novel asymmetric noise model crafted s
singly accessible. These radio maps not only provide WiFi fingerprint-location pairs but encode movement regularities akin to th
rowth of correspondence candidates. In this paper, we propose labeling landmarks with natural language descriptions and extr
eaks leads to all particle weights becoming zero. To overcome this problem, in this study, a straightforward and effective approa
ough its hierarchy and provides semantic information that helps with localization, like poses of doors and glass. In contrast to th
s ball localization. Additionally, we estimate hidden states like velocity and spin for trajectory prediction. Furthermore, to enhanc
e the viability of our single-sensor method compared to mounting an additional IMU. Our method provides a more reliable transl
int clouds, foregoing the need for scan registration which is error-prone in feature-sparse field environments. We compare seve
ndence generation in pose estimation, leaving them brittle to substructure ambiguities. To this end, we propose a hierarchical o
d to invert the high-dimensional information (Hessian) matrix. In our recent work [1], the First-Estimates Jacobian (FEJ) method
strained learning task with different optimization techniques is performed to determine an appropriate learning approach for a h
multiple challenges in meeting these demands. In this work, we introduce a novel robot navigation approach termed NaviForm
rnel, and the Gaussian model’s online inferential capability and computational efficiency are strengthened by updating a so
f object deformations and variations in lighting conditions. To model the complex dynamics, we employ a recurrent state-space
oposed approach leverages the short-term predictive optimization capabilities of MPC with the exploratory and end-to-end train
everage learned anchor embeddings to predict multiple trajectories, finding that these anchor embeddings can parameterize dis
s, we augment the policy by conditioning it on deformable object parameters and training it with a diverse range of simulated de
nteractions with the box while also valuing balance and gait quality. We combine the learned skills into a full system for box loc
d observations, removing the need and associated errors of state estimation, while embedding Hamiltonian structure in the dyna
trategy. This is due to model quality, myopic behavior from short planning horizons, and approximations due to computational c
-optimal for any particular system dynamics. We introduce a novel model-based reinforcement learning approach that aims to b
ften just as important (if not moreso) for performance as the choice of algorithm that is actually used. We posit that a significan
ntrastive learning to obtain a distance pseudometric that uncovers the sparsity pattern in the dynamics, and use it to reduce the
fers observation-only experience without the action or reward annotations needed for RL methods. In this paper, we develop a
pology and future systems will need to maneuver on flowable surfaces without getting trapped, allowing them to further expand
s equivalent to distinguishing dead-end states, indicating the maximum extent to which safe exploration is guaranteed, and thu
enoting the problem rearrangement with multiple manipulation primitives (REMP), we develop two algorithms, hierarchical best-
g method, ReorientDiff, that utilizes a diffusion model-based approach. The proposed method employs both visual inputs from t
manipulation by real robots. Forces are critical to manipulation, yet have typically been relegated to low-level execution in these
ning (RL) agent with a hybrid action space is then proposed to acquire the optimal policy for manipulating the objects without pr
ered the topology of the DOO in isolation, ignoring the arms that are manipulating it. Searching over possible grasps to accomp
e models cannot directly apply to the noisy point cloud in the real-world. To tackle this challenge, we leverage the property of re
s (MAMO). In this paper we make several improvements to E-M4M - we introduce the use of prehensile or pick-and-place rearra
mits and extreme time constraints (
verage priors in SD maps for the lane-topology prediction task. This enhancement consistently and significantly boosts (by up t
dtrees, which represents the environment using learnable features stored in three planar quadtree projections. Secondly, we co
hod employs a hybrid representation, encoding shapes with implicit features via a multi-resolution voxel map and decoding sign
nd shadow conditions to improve camera relocalization. We implement our scene representation upon a hash-encoded NeRF w
n quadrics mathematical formulation and geometric attributes, including the type, scale and pose, are insightfully integrated for e
escriptor, which combines height difference and height layer information to encode point cloud. Subsequently, we propose the
s, describing the scene at different levels of detail. This representation allows for fast scene geometry initialization and makes s
riors and to formulate a probabilistic, uncertainty-aware optimization framework for 3D reconstruction. We derive a probabilistic
ous works based on pixel-level features. Unlike 3D scene graphs, we create a purely topological graph with segments as nodes
patial information at the metric level. We introduce the Scene Action Map (SAM), a behavioural topological graph, and propose
D egocentric local map. A dual-structure subspace division paradigm, composed of a voxel subspace division and a novel pyra
s to extract spatio-temporal features from scan and local map and segment all 3D points into moving and non-moving ones. Ad
normals in high curvature areas. Mechanical LiDARs rotate a set of rigidly mounted lasers. One firing of such a set of lasers pro
efficient algorithm to colorize point clouds using an independent 360-degree camera. Given a LiDAR-based point cloud and a
map surfaces using Gaussian Processes (GPs) are proposed to describe the map uncertainty in a probabilistic way. To reduce
at each update by leveraging the low false alarm rate of LiDAR sensors. To support our design, we provide theoretical analyses
amic object removal (DOR). First, a two-layer 3D region-wise hash map structure of map management is employed for effective
sents a novel BA photometric strategy that accounts for both RGB-D and LiDAR in the same way. Our work can be used on to
chmarks for centimeter-scale µAUV advancements.
o swim at different energy barrier that needs to be overcome to realize the bistable motion. The theoretical models are derived t
troduce a new concept for a lightweight actuation system that can be used in combination with a traditional diving robot and a s
or testing/validation is resource intensive and expensive. In this paper, we present a physical testbed for underwater manipulati
rwater performance without adding excessive actuators, this design provides a feasible idea for the miniaturization of amphibiou
-dimensional maneuverability. A novel cyclic-differential method is layered on the mechanism. It integrates propulsion and yaws
, is given. Second, its hydrodynamic model is formulated by using Kirchhoff’s equation as well as Lagrangian method, and
m sampled sediment, which include the sampling depth, the sampled volume per area, and the possibility to collect replicas with
g supervised learning to obtain the properties from reliable proprioceptive sensory data. First, we propose a new expression cal
novel locally-optimal motion planning framework for robust real-time autonomous underwater navigation in 3D cluttered enviro
sing an efficient encoder-decoder backbone and modern lightweight transformer optimization stage to encode global context. Th
pose estimates from RGB input. In this work, we propose an approach that leverages 2D object detection to reliably compute 6
ew meters of often featureless expanses.This makes camera-based systems suboptimal in most open water application scenari
sian neural networks or ensemble methods. In this paper, we present a Conditional Variational Autoencoder-based framework (
representation is based on a neural reflectance field model, which learns albedos, normals, and volume densities of the underw
hrough comprehensive benchmark analyses on cave systems in USA, Mexico, and Spain locations, we demonstrate that robus
of the acoustic activity of a coral reef. Simultaneously, the AUV uses unsupervised topic modeling to identify different benthic h
esign the cross-attention model to capture the hierarchical features between different octrees, of which each octree incorporate
he other end is moved by the underwater robot. The obtained results help to determine which cables and which dynamics are c
order to model interactions with the free surface. We also report a preliminary evaluation of the hybrid model in numerical simula
 mentum gradient search, heading server, and initial conditions. Numerical simulations and experiments in regular and irregular
changes in buoyancy and center of mass, in combination with a control surface to move around. They are extremely power eff
g challenge for MPC. It must consistently keep the system within a predefined safe region (such as a reference trajectory) even
d, and swim with multi degrees of freedom in the water (contains only four driving elements). We analyzed and optimized prope
r, we present an N-Sphere (NS) and Shifted N-Sphere (S-NS) camera models, tailored to these cameras and lenses mounted i
ark the performance of 2 classical and 4 learning-based methods. The experimental results show that the learning-based metho
es, but fall short of providing a complete UVMS software framework. We address this gap by creating Angler: a software framew
an be understood intuitively. The approach is entirely decentralized and the communication cost scales by the number of group

conditions. Therefore, a robust mission planning framework is proposed using a centralized Evolutionary Algorithm (EA) and a d
idance policy based on Distributional Reinforcement Learning, which considers the interactions among ASVs as well as with st
ge-scale environments. This work proposes a computationally efficient planning under uncertainty framework suitable for large-s
rne mode, and submarine mode, which enable the platform flexible mobility, high-speed and high-load capacities, and superior
his challenge by using a graph formulation for the adaptive survey task. We then use Graph Neural Networks (GNNs) to both c
eighs 8.91 kg and has a maximum payload of 9 kg. It has a rapid-switching interface that supports convenient mechanical and e
ation (SSF) where single MSS beams or scan-lines are embedded into a pose-graph. The rendering of scans is not only based
ultaneous Localization and Mapping problem. We present a set of boundary factors that aid the inference process by seamlessl
vision-based approach to determine the water clearance. A single calibrated camera together with a semantic segmentation ne
utilizes adaptive color LED markers and dynamic color filtering to maximize landmark visibility in varying water conditions. Both
raphy. A line-detection algorithm is also introduced to improve the reliability of detecting navigable regions from LiDAR measure
y correct for distortion. Through experimental studies using an underwater robot integrating cameras and inertial sensing, the m
iron effects are calibrated using an ellipsoid fitting algorithm. The introduction of magnetometer data results in significant reduc
g window factor graph framework. Furthermore, we propose an acoustic-aided initialization method to exhibit an accurate initial
exploration of the free-space frontier. Special attention is given to acoustic image preparation and segmentation. Experimental
nd-effector being enhanced by merely locking its immediate connected links; this allows us to break down the complex planning
wo key enabling technologies: a gripper mechanism for secure attachment to target surfaces and a tethered drill configuration
tating the realization of load lifting and trajectory tracking. The simulation results have tested the proposed strategy’s accur
as a cubic exterior and can freely place cargo at any point on the flat top surface. However, the center-of-mass (CoM) position
ased tactile sensors on fully actuated UAVs in dynamic aerial manipulation tasks. We introduce a pipeline utilizing tactile feedba
h, there is no FWAV capable of jumping take-off to this day. In this paper, we present a process to realize the jumping take-off
ht (approximately 1 kg), fully actuated tiltrotor that can hover at 90-degree pitch angle is first developed. We design a perching/u
manifold configuration spaces. In addition, the complexity escalates for systems with cable-suspended load due to the hybrid dy
UAV in Gazebo. A control method is developed to automate the perching and take-off manoeuvre. The method is tested in simu
UAVs to execute tasks in mountainous and plateau regions. To broaden the task capability constrained by the maximum desce
udy their stability. Hence, we show the fundamental role played in the robustness of the load attitude control by the internal forc
m allowing 360° tether rotation, and a streamlined pod with two ducted propellers connected via a tether. Winding and unwind
nipulator to successfully execute millimeter-level flying pick and peg-in-hole task. Firstly, the error amplification effect of the floa
ed method on flight data collected over specially designed trajectories. The identified model allows us to predict the aerodynam
ght envelope of a tail-sitting vehicle, which, while commonplace for fixed-wing aircraft in performance textbooks, is a novel figur
e towards longer-duration indoor operation of ion-propelled fliers. This work presents the first empirical study of ground proximit
d to model an Unmanned Aerial Vehicle (UAV) with a penalty associated with a turn due to the required additional thrust to main
 misation. A number of future cost variants are investigated based upon prior work for a realistic urban-rural mix logistics case s
posed a real-time trajectory planning algorithm. The time optimal aggressive trajectory is efficiently generated through multimod
f flying with either side of the wing facing up and this feature determines the craft’s direction of rotation about its body Z axi
effector tip are in contact with the work surface. With a standard full-pose controller, attitude errors often occur due to perturbat
orward 5-DOF geometric tracking control strategy that generates 2D force and 3D torque. A decomposition approach is designe
nal resources than the standard exponential-map-based method that keeps state estimates on the manifold. Furthermore, our m
speed quadcopter control using end-to-end RL (E2E) that gives direct motor commands. To bridge the reality gap, we incorpora
d to comprehensively validate the hybrid locomotion. The robot leveraged the superior maneuverability in the rolling mode to tak
drone’s conceptual design that includes topology, actuation, morphing strategy, and controller parameters. The method co
ejection control (ADRC) as well as geometric control for a quadrotor with bidirectional thrust capabilities. Moreover, to strengthe
ose sign is dynamically switched in real time to select, between two options, the torque associated with the lesser cost of rotatio

n operations. The blind utilization of these platforms for such tasks can lead to instability and accidents, creating unsafe operati
grasp the load, we developed a two-part grasping mechanism, creating a near-rigid connection between the UAV and the load.
esent a learning-based control algorithm that allows online trajectory optimization and tracking to accomplish various aerial inter
acitated generalization of vehicle routing problems (VRP), which may also be of separate interest. In contrast to standard VRP a
dard quadrotor, we propose a modular aerial system, IdentiQuad, that combines only homogeneous quadrotor-based modules
An optimal state feedback controller and a PD+ controller are designed in order to dampen the oscillations of the platform in the
a risk-aware cost function that integrates pre-computed cost maps, the new concepts of damage and insertion costs, and an ad
ty, and transparency when compared to conventional contact-based transmissions. Furthermore, the magnetic coupling effect p
ng capstan effect can produce clutches with the highest specific shear stress in the literature. Our system generated 31.3 N/cm
In the experiments of the vehicle equipped with foldable wings, the lift and attitude torque of bionic foldable wings are measure
upply that can be steered and downscaled. In experiments, we verified our concept with 3D-printed bristle-bot units equipped w
it more compact and robust. The new concept, which we refer to as the Compact Modifier of Active Coils (C-MAC) mechanism
d. We demonstrate module radial expansion and axial elongation in a straight and bent configuration (up to ∼4◦). This is th
perimental method to determine speed and torque bandwidth of the actuator is presented. A rigorous method to determine back
ation of the exoskeleton are analyzed. By changing the working mode of the gravity compensation device, the control experime
ability, positional accuracy, and adaptability to any guide surface through magnetic coupling. The experiments showed satisfact
the finger and the suction gripper, and the function of the supporting finger. The gripper shows higher grasping stability and pra
el description of each module, we can construct a system that allows the robot to attach and detach modules during task opera
cal prototype was developed, and the performance of this origami-inspired continuum robot was evaluated by comparing the m
een any two modes by re-arranging the foreleg wheels. When standing with its forelegs raised, the robot can carry objects and
OF ankle joint, and one-DOF toe joint, is developed. All actuators are located at the proximal part of the thigh frame to minimiz
ects. A system of linear arrangement and connection of two autonomous sailboats is considered as an appropriate solution for t
peed at maximum wiping performance was verified to be consistent with the experiment. The LiDAR distortion experiment resu
ses for effective human–robot interactions. Extensive sets of experiments, testing both the low-level torque controller and the
attern. The hinge design employs rolling contact elements to mitigate collision among the panels, particularly in regions with mu
cally revealed. It is shown that using the synthetic wave input constituting two sine waves based on the resonance frequencies
devices, we propose a nested bi-cavity-based growing soft robot (BIBOT). The new design consists of two coaxially nested ca
he robot to change its mode of locomotion by transforming into a rover, quad-rotor, mobile inverted pendulum (MIP), and other m
SKTs with multiple strips have been developed. The results show that under a specific volume (250 μL), the optimized strip str
 nd patients during a StS rehabilitative task. This approach offers several advantages, including improved therapist-patient intera
erived based on the interaction Jacobian matrix and impedance modeling. Meanwhile, we implement a tracking method of in-pla
pins to generate a repulsive field to push the object without physical contact. Under such a formulation, the main challenge is th
m regulating the cable tension, the overall stiffness is adjusted to adapt to various surroundings. The mapping between actuatio
ocedures since the robot control is often optimized for navigation inside the eye. This challenges the integration of robotic syste
hesis via a bypass adapter are compared for three control schemes: benchmark passive controller, PVIC, and PVI-HVC. The d
(CNNs), which provide semantic scene information from iOCT images. Although the robot performs a surgical task autonomous
reat importance for the active colonic navigation of the robots. To address these challenges, we develop a track-based stereos
Research Kits" (dVRKs), which have been distributed to research institutions around the world to support both training and rese
e introduce a novel approach based on variable impedance control, allowing the real-time optimisation of compliant controller p
tructural stiffness, and a design optimization for dexterity under anatomical constraints. Experimental validations demonstrate th
ble access to 3D data via robotic tracking and incoherent compounding of ultrasound frames, which results in improved interpre
setup, this paper presents an image-based pose estimation method to obtain real-time instrument positions in a standard 12mm
sed non-rigid registration to adapt patient-specific properties using subcutaneous bone surface features rather than the skin sur
an exhaustive scan of the chest wall region is performed. The position of the probe that allows the mitral valve to be centrally p
he DopUS-Net combines the Doppler images with B-mode images to increase the segmentation accuracy and robustness of sm
pon this, a CMF osteotomy robotic system with compliant human-robot interaction and osteotomy trajectory planning was devis
strong vibration coupling of low-frequency force information, leading to low identification accuracy and slow response speed in
ots, to peel these delicate membranes successfully without damaging the healthy tissue. Minimal intra-operative motion errorsw
ven model-free adaptive control method based on partial form dynamic linearization is used to control the vertical force. Second
designing such a controller is associated with the size requirement. For conventional microsurgery, surgeons perform surgical p
s collaborative system, enriched with haptic assistive modes to enforce virtual fixtures, holds the potential to significantly enhan
device. To address the above challenges, we propose a vision-based method for accurately detecting high-touch areas in the s
egrating augmented reality and virtual fixtures to address these limitations. Augmented reality intra-operatively visualizes the hi
ments, respectively. The system features an integrated tactile handle, enabling the reception of human-robot interaction forces
the absence of a complicated static model. This paper aims to solve this issue. The simulation in ANSYS is carried out, and the
rgeries on the leader side. It can capture the interventionists' delivery displacement, twisting angle, clamping force, and provide
izing physical contact risks with patients or other surgical tools owing to the simplified layout near the surgical field. However, co
ed friction in the actuation mechanism. Castigliano’s theorem is used to build a mathematical model of the proposed mecha
netic forces and torques, resulting in a lack of controllable degrees of freedom. Here, we introduce a reprogrammable magnetic
s fixed with a radial magnetized O-type permanent magnet. It not only realizes flexible movement in the fluid environment, but a
e motion accuracy. Furthermore, the application of model predictive control with constraints significantly bolsters the flexible rob
is under-constrained and may be ill-posed for complex shapes, falling into local minima. In this work, we introduce a novel met
apping to mitigate branching methods. We derive a lower triangular transformation matrix to disentangle the inequalities and pr
erate vibrations and impacts on the main body of the capsule. To optimise its motion, we provided a numerical solution for calcu
tro closed-loop control. However, in-vivo procedures will require different feedback modalities due to the opacity of the human b
ntly reduce the cognitive load on the surgeon while speeding up the surgical process. In this paper, a hybrid robotic endoscope
ssification using Multi-Layer Perceptron (MLP) and Stacking Ensemble networks (SE) was carried out alongside unsupervised c
ontrast images. However, ultrasound images are hard to interpret, and it is difficult to localise vessels, catheters, and guidewires
parameters of the robot to satisfy different working scenes. According to the experiment, the accuracy of the robot is related to
e and reliability of medical systems that require micron-level precision, especially in real-life scenarios. To address this challeng
 (SHER) with cooperative control mode can filter out surgeon's hand tremor, yet another important safety feature, that is, minim
e and may potentially lead to non-optimal solutions. This paper presents a hybrid offline reinforcement learning and damping co
olume constraints of capsule endoscopes whilst augmenting their therapeutic functionalities. The FCE’s distinctive flexibility
the trocar and maintain strong coupling with the external actuator. The operation principle and modeling were established to in
on of a magnetically-guided origami robot, incorporating flexible printed circuit units for sustained and stable phototherapy of H
yzed. Secondly, the simulation experiments based on FTG models with three different sizes are implemented. When the cathet
otic surgery. In this paper, we propose a novel data-driven paradigm for simultaneous estimation of shape and force along high
eudotumor with millimeter accuracy. ASTR consists of a dual-camera vision system, an electrosurgical tool, a vacuum grasping
rgeon by leveraging an online data-driven learning and adaptive control method coupled with a reduced-order trajectory plannin
ts. Our system was implemented using a commercial Kinova robot and a dedicated modular end-effector that uses standard m
int, and the initial pose and position of the robot. By combining the mobilization of the eyeball and adjusting the pose and positi
s with surgeons, and analysis of technical and clinical requirements of stent implantation procedures. Results from a human-in
based modulation model for robot-assisted MRg-FUS treatments. By introducing a rapid phase estimator to provide an efficient
ution is tendon-mortise shaped slits along the axial direction, enhancing the overall stiffness. The bending of CAST is actuated
ting the guidewire within the vascular environment was constructed to guide the robot in precise guidewire manipulation. Exper
uces a new soft micro-robotics catheter, with diameter of only 0.8 mm, equipped with a hollow channel. With this new design, t
mework, we have conducted various simulation studies and experiments
t can account for the effect of nonlinear strain-dependent behavior of biological soft tissues under compression. Real-time finite
mplement a symbolic Python implementation of the Recursive Newton-Euler Algorithm (RNEA), the approach aids in dynamical
nt these components through an external magnetic field. The automated assembly has the advantages of high accuracy, high
A comprehensive analytical model is also studied, which was then validated by comparing to the experimental results. A launch
rorobots. The air bubble-based structure enables multiple functionalities including cargo delivery, multimode locomotion, microm
ping motion under specific magnetic field conditions. Magnet rotation allows precise control over gripper actions, returning to a
c steering (MS). The BFC produces an open-circuit potential (OCP) between a bioanode and a biocathode by redox reactions. T
al microfabrication techniques for further miniaturization. The flat geometry aims to provide superior maneuverability and opens
ematical model for multihelical motor that can be selectively controlled. Based on this model, we developed a micromotor that c
deep learning-based image-processing techniques to perform robotic micromanipulation on biological specimens. Using two ex
of this with robot control could provide crucial guidance in the design of feature extraction and motion control algorithms. In this
mini-incision under stereo microscope, featured by an additional rotation degree of freedom mounted on each arm. With this se
an object (e.g., a cell). Despite its high accuracy, only sparse depth data can be obtained due to its low efficiency. This paper a
, we have innovatively developed a dynamic adaptive imaging system. By incorporating peak-finding and in situ Gaussian blur
the limited field of view under microscopy. Then, the transformation matrix between the camera and coordinate system of the r
erical joint allows rotations of 140° around the x-axis approximately, 140° around the y-axis approximately, and 20° around
worm equipped with a novel 2-DoF origami transmission mechanism. The S2worm is fabricated through SCM and holds a top s
l framework from governing equations to calculate the forces and torques exerted by magnetic miniature robots in their respect
onal neural networks (CNNs) has offered promising advancements in microscopic image analysis. However, previous CNN met
he limited depth of field and pattern requirements of the microscope-camera system, and determines its intrinsic and hand-eye
econdly employed hybrid use of the image-based manipulation and the quartz crystal resonator (QCR) force sensor. The imag
mage of a TGX1 calibration grid and a 3.4- billion-year-old organic-walled microfossil are reconstructed by automatically merging
s developed for the task fail to comply with the standard clinical setup. Instead of using a motorized micromanipulator as in exis
ng during ROI Align. Moreover, morphology measurement of sperm tail is demanding because of the long and curved shape an
 led sampling with micro-scale accuracy with due consideration of physical constraints can pave the way for using PAM for robo
work area, and potential damage to cellular structure. Here we develop a hydrodynamic manipulation method to rotate and trans
thin a fluid-filled lumen. This teleoperation capability is facilitated by a scalable X-ray-guided robotic platform, extendable to ma
and dynamic physical simulations to create fast and accurate suction grasp pose annotations. We introduce Sim-Suction-Pointn
ng objects. This models sensors that are range-limited and line-of-sight—objects occluded or outside sensor range are unobs
el environment-adaptive path searching method, which can generate paths in real-time in various environments by adaptively a
red environments. We present HIMOS, a hierarchical reinforcement learning approach that learns to compose exploration, navi
objects and avoids collisions. Furthermore, we propose planning using the minimum statically-feasible friction coefficients, whic
t tuning or extensive training, making it very difficult to introduce novel or diverse tasks. In addition, the adjustment of several, in
are online organized to eliminate the redundant, outlier grasping poses, which can be encoded as a grasping pose observation
the robotic arm on the base and bring it into the legged robot's dynamics, and then use a nonlinear model predictive controller
formative toward manipulation tasks, such as grasping in unknown, cluttered scenes. Our proposed approach, ActPerMoMa, ge
g from current robot interaction with the unknown environment. These two ingredients enable a highly sample-efficient and ada
point clouds as the state representation and leverages pre-trained point cloud reconstruction Transformer to learn a latent dyn
ictive Control (LMPC) to address possible model mismatches of the first principles model in high-speed racing. To this end, a M
nd exteroceptive information. The proposed model generates reliable predictions of 6-degree-of-freedom motion and can even
anics. Drawing inspiration from data-driven modeling, geometric mechanics (or gauge theory), and adaptive control, we employ
e benefits of geometry-aware convex regularisation using the log-determinant divergence of the pseudo-inertia matrix. Also, the
which the measurements (output) are the endpoint poses of the DLON, which are subject to unknown dynamics during manipu
orks to learn an end-to-end model of the multirotor-slung-load system and, at a given time, estimate a sequence of the future sy
ccurate at predicting future robot poses than conventional kinematic motion models. By efficiently solving a convex optimization
work, we introduce a unified human-likeness metric built from a hierarchically weighted sum of individual metrics. The proposed
T is its silent actuation, which we achieve through pneumatic artificial muscles in conjunction with a controller loop to respond to
Based on the proposed dataset, a flexible feature encoder for ESP is introduced to various prediction methods as a seamless p
SynthAct can be applied to arbitrary 2D or 3D video action recordings, making it applicable for uncontrolled in-the-field recordi
data, we summarized three characteristics observed from human demonstrations: varying-speed actions, grasping effect read f
Activities are simulated through a two-step optimization process, involving a quadratic-programming based whole-body controll
el with 90 body segments, 206 joints, and 700 muscle-tendon units, allowing simulation of full-body dynamics and interaction wit
we propose a keypoints-guided lightweight network with an encoding-decoding framework. As the input is an image, a lightwei
not viable for real-time estimation until recently. To achieve muscle forces estimation in real time, a moving horizon estimator (M
at is integrated with the LuGre frictional model for LJ-based continuum robots. Then, we take a comprehensive theoretical analy
mulations are experimentally validated on the MAMMOBOT system, which is a miniature steerable soft growing robot for early b
ing gains within a traditional PID controller corresponding with the predicted motion of the reference trajectory. Through compa
of capturing non-constant curvature deformations of the soft segments. We theoretically demonstrate that this constitutes a mo
nitialization or multiple re-starts, existing algorithms often fail to find a solution. We present CIDGIKc (Convex Iteration for Dista
e using data-driven MPC to control an underwater soft robot, and employ the AutoMPC method to automatically tune the hyper
ormable objects in realtime using scattered and unordered key points. By utilizing a robust probability-based labeling algorithm,
ol for the soft robot, but also achieve accurate learning of the robot's dynamics during the online control process. The proposed
the torque action by leveraging both the knowledge of the model and data obtained from past experience. After presenting a 3D
ctively exerting force on the finger against tremor and meanwhile following the voluntary motion. The proposed pressure control
especially in confined spaces. This can be avoided when the sensor is located at the robot’s base. This paper studies the s
odel of the task. Then, a reinforcement learning method is used to predict the actuation pattern given the goal position. The prop
on sensitivity analysis. Additionally, we show that our general control pipeline can be fused with techniques from 3D character
ese models, whose dimension is often very large, make learning and optimization methods very costly. In this paper, we propos
architecture is proposed, wherein a model-based controller is employed for rapid convergence to a target position, followed by
reconstruction methods to acoustic signals from the microphones for soft robot shape proprioception. To improve the proprioce
ed behaviors. By deriving these connected models from general formulations of dissipative Lagrangian systems with symmetry,
. We introduce several simple models for elastic coupling, typical of those seen in these systems. We then propose a controller
d both muscle forces and grip strength. Our results reveal a nonlinear relationship between tendon forces and grip strength, wh
C (AR-IOC) algorithm. Given the walking speed range examined (0.6-1.5m/s), the converged cost function weights suggest that
asping while reducing mass and complexity. Forces are distributed among the robot's grippers using an optimization-based con
the construction and experimental testing of the quadrupedal climbing robot MST-Q with a high payload carrying capacity of up
h a stable pose for precise tasking, whether the cause be due to underactuation, modeling inaccuracy, suboptimal control para
tation of locomotion by generalizing various nonholonomic constraints into local metrics. By combining these two approaches, w
ansparently combine both modes of locomotion has not been extensively explored. In this work, we present a hierarchical onlin
ifications via hand-crafted or STL-inspired rewards, but might encounter unexpected behaviors due to ambiguity and sparsity in
ought. This paper shows that the piecewise-linear form of the cost function employed in previously work can be used to build a
lanning in environments with narrow passages. Graph-based representation called roadmap is a common representation for pa
aussian. To overcome this limitation, our method aims to identify a target mode of the optimal distribution and guide the solutio
we use a linear first-order hyperbolic partial differential equation to transport the visibility quantity in all directions. In order to acc
ation using the Gauss-Legendre quadrature; and 2) a data-efficient decoder for interpolating clothoid splines that leverages the
ltiple Ackerman (car-like) robots in lifelong settings, yielding high-performance centralized and decentralized planners. Our plan
re battery-powered, it is physically not possible to continuously explore on a single battery charge. Our paper tackles this challe
e a feasible solution is found. In this paper, to enhance the convergence performance of optimal solutions, we present Reconst
vial. In this paper, we propose to perform quasi-static path planning on an implicit manifold. We model elastic obstacles as exte
s. The second method helps to reduce the amount of time robots spend waiting for paths to be cleared, and the overall distance
and infer the states close to the optimal path to guide planning. We propose Neural Informed RRT* to combine the strengths fr
generality-preserving algorithmic improvements to any such planner and significantly accelerating critical subroutines, such as
g approach and an improved method utilizing noisy gradient descent. Simulation results demonstrate the efficacy of our metho
mization challenges for robot manipulators proves difficult due to inefficient, high-dimensional trajectory representations and a l
anteed to find a winding angle-constrained path. The effectiveness and performance of the proposed algorithm are studied in s
on can reproduce scores from traditional discrete representations of boustrophedon path plans with high fidelity. Finally, optimiz
to compute data structures that enable online planning provably guarantee the ability to generate motion plans, potentially sub-
c properties of the surrounding obstacles such as height, density, and solidity/stiffness. The fully-trained policy's critic network i
s within a certain time limit. While deep-learning-based algorithms tremendously improve success rates, they are much harder t
and generates a clique cover of this graph to find clusters of samples that have mutual line of sight. These clusters are then infla
hm utilizes a trio of hierarchical graph data structures that each capture certain elements of the problem such as connectivity (v
cessing algorithm for A* and other graph-search-based planners.The A* post-processing algorithm, called APP, is developed ba
pting Coxeter triangulation's manifold tracing and cell construction procedures to concurrently triangulate the configuration spac
lso propose a frontiers-based exploration strategy able to autonomously and safely build a complete 2D map of the environmen
in these missions, we propose a safe multi-goal (SMUG) planner that generates an optimal motion path to visit those targets. T
bject detectors and the efficient point cloud manipulation using OctoMap, it is feasible for our approach to deploy on low-weight
can faithfully model gate constraints with various configurations and thereby generate the most time-efficient trajectory while co
ut also on the environment. Given a new environment, it is difficult to decide a priori how one should design a motion planner. In
ying shape and non-linear dynamics of the system, ensuring whole-body safety and dynamic feasibility. Additionally, an adaptiv
d implementing cubic interpolation at the time resolution to calculate the continuous gradient for optimization. Facing the challen
ry of the ground target throughout the predicted time horizon of the NMPC, is required. Towards this direction, Bézier curves
lanner to predict future system states. The state reward value relevant to the particular water surface cell increases over time a
mprised of a rigid object lifted by a human and a quadcopter from its end is considered. Custom build sensor systems, namely
rategically distribute the envisioned 3D model to be printed into small, manageable chunks suitable for distributed 3D printing.
ver-actuated systems, we first discover an internal disturbance phenomenon named servo-conflict and consider it in the controll

m. The disturbance estimate is used directly for the passive FTC with rotor failure. At the same time, a modified controller struct
g the end-effector's motion through a virtual quadrotor, which is inherently slosh-free and differentially flat, thereby allowing us t
motion control algorithms from other modules such as perception and path planning. This paper closes this gap by proposing a
d is updated online using real-time data. Then we propose an Online-Learning-based Distributionally Robust Nonlinear Model P
e over time and have to be compensated by a proper adaptation of the robot control. To move forward in this direction, this wor
een successfully applied to traditional cable robots, particularly in handling model uncertainty. The paper presents a novel ILC c
(F/T sensor), implemented through a three-layer control architecture, consisting of motion control layer, Admittance Control lay
reachable set to ensure safety-critical coordination control, i.e., guarantee the safe coordination of the robots during locomotion
s: 1) Trajectory generation for seamless transitions between columns, 2) Foothold re-planning in regions deemed unsafe, 3) Sa
cyclic motion (ARC$^{2}$M) control scheme, which takes into account the cyclic motion index and the physical limitations of rob
safe controller tuning in the presence of heteroscedastic noise, combining safe learning with risk-averse Bayesian optimization.
nd real-world prototype of our retractable worm robot, called RW-Robot, and then we introduce footstep phases to express the t
up the convergence of tracking errors and strengthen anti-perturbation ability. The former aims to enhance feedback portions t
uccessfully utilized in many applications, which recently sparked their use in high-acceleration motion found in wafer scanners,
timization allows us to impose constraints, by which a controller acquires additional properties. As a particular use case of the C
P) controller and a master hand velocity (MV) controller related to the surgeon's manipulation, and a planned trajectory tracking
or in the free (motion) task directions and explicit force tracking in the constrained directions. Advantageously, the involved forc
e-art passivity-based control and Deep Reinforcement Learning (DRL). The concept of energy tank is used to provide stability g
ques at the end-effector, particularly in applications like electric vehicle charging and narrow-space exploration. It also enables
ks. For instance, we can initially train a neural network (NN) policy using velocity commands as inputs. However, to handle new
used by the truncation in derivative order. The identified finite dimensional Koopman bilinear model would serve as predictive m
g (MBRL) method for robotic grinding. Our method employs a cutting-surface-aware model as the object's shape transition mod
e hybrid multi-contact model and continuous improvement of the control performance for contact-rich tasks. This framework incl
novel learned slip approach outperforming similar acceleration learning approaches. Our contributions are validated through an
using projectors. We develop an approach to redundancy resolution and decoupling on position level by determining subspace
r of actions to solve a puzzle. At the core of our approach lies a new path defragmentation method, which rearranges and optim
gurations found when connecting new nodes rather than rejecting the sampled nodes outright, allowing for more productive exp
at we traverse in a depth-first-search approach by repeatedly applying local plans. We analyze how structure size and shape a
of joint positions, collision regions, and motion parameters during the path following. Subsequently, the sliding control point algo
zing the time and space pair along with dynamic obstacles. Qualitative and quantitative analysis in Carla is reported to discuss
ecution. The occurrence of reconfigurations should be minimized because they increase time and energy usage. In this paper,
 Recently, physics-informed motion planning methods have emerged that directly solve the Eikonal equation through neural net
s synthesized by which an ERG is constructed to modify the UAV set-points. The method can handle polyhedral obstacles and
anges to the physical environment so as to make planners fail to find a solution. Blindspot attacks exploit occlusions and senso
astically across space and time. Moreover, current coverage methods dealing with battery capacity constraints either assume s
es. We specifically focus on scenarios resembling 3D puzzles, intentionally designed to be complex for humans by requiring intr
ent errors that cause robot pose deviate from the pre-defined path. The trajectory deformation and tracking is challenging becau
for long-horizon motion planning, by using CBF-induced neural controller (CBF-INC) to generate control signals that steer the s
ffector is required to traverse a continuous task-space trajectory (time-parametrized path), such as in mobile printing or sprayin
eural network output to move sampling regions closer to the constraint manifold. Our experiments show how VQ-MPT improves
g horizon to satisfy all dynamics constraints. The proposed planner can compute plans for scenarios that would be difficult for t
rain risk mapping method that can be used for both UAV and UGV is developed. This method efficiently combines quadtree an
or robots to navigate populated indoor spaces in a passive manner. For demonstration videos, see our project page: https://site
on a novel dataset show that the inclusion of gaze cues significantly improves the classifier performance (AUROC increases fro
". The goal of the robot is to persuade the participants of the game into changing their answers whilst avoiding negative feeling
o the sensor occlusions caused by large objects. To advance the state of the art toward a human-human level of handover flue
is work, we propose an LLM-driven interactive multimodal multitask robot navigation framework, termed LIM2N, to solve the ab
e magnetic films, soft silicone elastomer, a Hall sensor and actuator array, and a microcontroller unit. The e-skin captures the m

ion during dual-hand interactions. Simultaneously, current single-view methods independently reconstruct the left and right han
king 3D information, our approach yields comparable/superior performances and better robustness against missing keypoints t
ugment the intention recognition capabilities of simple binary tactile sensors by adding a robot-mounted camera for human ana
on. Benefiting from a unified formulation of visual dialog and grounding, our method can be trained on a joint of extensive public
ed goal objects while engaging in conversations with users. To solve ZIPON, we propose a new framework termed Open-woRld
task scenario, an intention-oriented utterance (e.g., "I am thirsty") is initially given to the robot. The robot should then identify th
rocessing with an attention-based network to overcome EIT’s limitations. The preprocessing network applies a linearized o
h situational aware context in a zero-shot manner. For ambiguous commands, we disambiguate the command by interacting wi
on of users in the discussion with verbal behaviors; an autonomous gazing behavior module trained via deep reinforcement lea
up system. This article presents the assessment and benchmarking of the novel XoLab Natural Language Interface, a voice use
urements, including 6 DoF and 5 DoF pose measurements from hand gesture detection, as well as 3 DoF position measuremen
blem. In this paper, we introduce a distributed multi-aircraft visual-inertial-range odometry system that analyzes the position, ve
dictable performance in real-world environments. In this paper, we propose a data association algorithm based on neighbor con
er-robot loop closure prioritization technique that reduces communication and accelerates convergence. We evaluated our RO
al consistent map. In this paper, we propose a real-time dense reconstruction system called AutoFusion for multiple UAVs, whic
otent central server, thereby alleviating the computational burden on individual robots for local odometry calculations. The serve
uted manner. Moreover, our approach relies only on on-board sensors and requires only one ultra-Wideband tag per robot, elim
or transformation using continuous time (CT) inertial measurement unit (IMU) modeling and derive associated ambiguity as a po
provides robust place recognition performance despite differences in both translation and viewpoint, (ii) is capable of identifying
he robots and the payload. Theoretically, the convergence of these observers is analyzed using Lyapunov functions under mild
h learning approach that formulates heterogeneous collaborative scheduling as a bipartite matching problem that maximizes a r
ctions while neglecting the dynamic constraints of the robots and unknown behaviors of the targets. This work proposes an onl
er of robots) as too many robots share a resource (e.g., space, communication channel). We provide additional evidence based
of individual robots, asynchronous to tolerate communication delays and outages, and general purpose to handle any CSLAM p
DO) algorithm, a population-based training-based framework that combines reward and diversity training to different stages, rat
subsequent waypoints. For the high-level policy training, unlike other RL-based navigation approaches, we incorporate awaren
nd prompt the LLM agents to improve their plan and waypoints in-context. For evaluation, we introduce RoCoBench, a 6-task b
RL policies. In this work, we propose an end-to-end DRL approach to control quadrotor swarms in environments with obstacles
ot robust enough for most existing DRL works to make new robots fast adapt to current team policies, causing performance deg
ree search (MCTS) provably converges to the true state value estimates; however, the exponentially large search space often
he race. Specifically, the competition information among competitors is introduced as the additional auxiliary signal to learn adv
sting research on multi-agent reinforcement learning. Due to limited communication, when OWC is used, overall performance is
on from a camera with a limited angular field-of-view. Our solution utilizes a particle filter to analyze uncertainty in the location o
ize a heterogeneous coverage control approach to partition the space according to robot capabilities and the estimated probab
cation protocol which can be implemented by the publishing and reading of web pages or other asynchronous communication te
small clusters, consequently resulting in the failure of cohesive flocking behaviors. Instead, our approach leverages spatiotemp
offset estimation with our previous bearing-based mutual localization by a novel error representation. Based on the error model,
tions. MOCHA is based on a gossip communication protocol that allows robots to interact opportunistically whenever communic
method to address this NP-Hard planning problem. Our approach involves initially determining coverage paths without factorin
onal motion of one cell about the edge of another, and increases the number of neighbors each cell can communicate with and
rement for exact knowledge of vehicle dynamics. The simulation illustrates the effectiveness of the proposed controller in mana
end, we propose and implement a system that allows a user to specify missions using Behavior (BTs), which can then, at runtim
ext, we consider the multi-robot navigation and human-robot interaction, respectively, as a potential game and a two-player gam
to compute trajectories that avoid collisions. An integral aspect of our approach is the integration of a human-in-the-loop (HIL)
this decomposition with a hypergraph where vertices are decomposed elements of the planning spaces and hyperarcs are tran
onstration of a method for multi-agent, relative pose estimation that is robust to A) sensor blinding and B) dynamic uncertainty.
namic Targeting framework that comprises a physics-based slew model, a dynamic model of observation utility, and an algorith
ent vector extracted from this process for efficient place recognition tasks. This approach not only achieves comparable place r
proach are: (i) an MPC-based motion controller in each individual robot can be (re-)configured, at runtime, with the locally and t
ing advantage of sensors with different capabilities. Second, performing the aforementioned tasks when considering uncertaint
thms to efficiently tackle static and dynamic obstacles with arbitrary shapes. Collision avoidance between communicating robot
onflict with the intentions of other robots. These policies are concisely expressed as behaviour trees via algebraic logic simplific
ity of this method suffers due to the large dimensionality of beliefs which the iLQG must propagate. We examine the utility of pa
follow the targets in their fields of view. Multi-agent tracking algorithms additionally assume inter-agent synchronization of obse
te the effectiveness of our framework, we conduct a comparative analysis with state-of-the-art algorithms. The results of our ex
xecuted by a coalition synchronously. We define a new problem for heterogeneous multi-robot task allocation and formulate it a
s are too complex. This raises the question: can these heuristics be learned? To this end, this paper particularly develops a Gra
hrough focused discussion and short summaries of common open-pit mining operations. The principal activities that take place
ive path for each vehicle; 3) the Bottom Layer (or Roadmap Layer) defines the final routes to be followed by each vehicle and c
ability and adaptability of our algorithm through relevant experiments. Furthermore, considering the strict communication limita
control scheme that allows maintaining a prescribed minimum PE level so as to ensure filter convergence. At the same time, t
er which otherwise rigid formations become flexible. This paper models the sensor and robot constraints in bearing formations
oposing an efficient iterative algorithm, namely Entropy Maximized Patroller (EM-Patroller), with the per-iteration performance im
 able overtaking rule, a guiding vector field, and a control barrier function with an adaptive radius to facilitate smooth overtakes.
ite tiled surface section of size 1m x 1m in simulation. We first describe the model parameters that characterize our simulated e
placed after the blocks below it, and a block that needs scaffolding cannot be placed until after the scaffolding is. Prior works f
e FFCGS is implemented by using a signed distance field (SDF) as input. Firstly, a network that samples feasible 6D wrist pose
ce the system's inertial mass while enhancing energy efficiency and overall performance. Leveraging 3D printing and carbon fib
tors as well as components 3D printed from PETG, TPU, and conductive TPU. These materials allow the robot hand to have a
y optimization for the dexterous hand while maintaining smooth and natural properties of the grasping motion. Specifically, we f
dom or structures. Furthermore, it fails to accurately capture force feedback information between the hand and objects during th
Hand. It is covered with multi-layored rubber skin to realize human-like soft and pleasant feel. Force sensors on each finger link
driven humanoid hand with a soft and sensorized tactile skin. Multi-material 3D printing allows us to iteratively approach a cast
vention when trained on expert datasets. Notably, BC outperformed even the most advanced offline reinforcement learning (RL
h the proposed method as well as the versatility of the hand for several different parts.
rative landing ability mapping for the cross domain UAV-USV fleets (CDUUFs). Then, by incorporating suitable allocation with U
he-art, we are able to apply distributed CBFs to the problem of global connectivity maintenance in presence of communication
order holonomic mobile robots, the leader has to be controlled as a first (second) order non-holonomic mobile robot; 2) In a tea
ormulate the whole problem as a single factor graph with multiple connected layers representing each aspect. We use Gaussia
manned aerial vehicles (UAVs) and how their operation can be improved by the possibility to dynamically switch between QoS
hput remote-controlled experimentation and data collection for robotic manipulation. Furthermore, CloudGripper is intended to f
We evaluate FogROS2-Config on three robotics application tasks: visual SLAM, grasp planning. and motion planning. FogROS
ead positions communication support robots to directly maximize the probability of packet delivery by the underlying opportunis
n system to identify passivity excesses. This is further used to avoid over-dissipation of energy in the passivity controller by aug
h pixel-level view synthesis. The camera pose can then be estimated as the solution to a Least Squares problem by selecting c
AR. Based on the non-repetitive scanning pattern and point-level temporal nature, we propose a two-stage module to enhance
he denoised occupancy map also plays key role in generating polygon represented free space in the scene. As a result, our ela
se a cascade framework consisting of two depth-aware learning paradigms. First, a depth estimation (DE) scheme leverages re
ween clear and adverse weather can be decomposed into two factors with distinct characteristics: a style gap and a weather ga
edefined keypoints for all objects in the scene simultaneously. Based on the keypoints and an instance semantic segmentation,
ncertainty estimation within a multimodal framework. Through a multivariate Gaussian product of the latent variables in a Variat
e leverage a pre-trained monocular estimator to extract local geometric information, mainly facilitating the search for inlier 2D-3
rate pose estimation caused by the inherent ambiguity of RGB images, we design a coarse-to-fine approach that first harnesse
, such as uncertainty estimation and perceptual difference from reconstructed images, make it possible to identify pixels of unk
better point-to-voxel aggregation function. Our design respects the permutation invariance of sparse 3D points while being mo
e gap between these two modalities. We achieve this by using an object-centric slot attention mechanism that can iteratively de
bjects are the majority. This work proposes two novel attention operations as generic hierarchical designs for point-based trans
ain a fused frame with merged boxes. To ``forward'' frames, we use vehicle motion models to estimate the future pose of the b
t-based, largely due to their coarse representation and simplistic architecture design. While most recent research has aimed to
gh-resolution image into its low-resolution counterpart and axial details. The aim is to well preserve the image content through
this, we introduce an intriguing Cross-Cluster Shifting operation to unleash the representation capacity of the point-based detec
omplementary cues between semantic and instance heads through the gated mechanism. In addition, a novel edge-aware cons
. The feature alignment step further bridges the intrinsic attribute gap between the sensing modalities and stabilizes the training
citly addressing class distribution shift to improve pseudo-label reliability in self-training. Our approach uses the domain invaria
comings, an optimized view transformation is proposed, which can be easily modulated into the LSS-based methods. The prop
s, thus explicitly guiding to learn objects' temporal cues. To this end, we propose a model called DAP (Detection After Predictio
adding an additional prediction head for small objects, the higher-resolution feature maps could be directly used to detect sma
rception. To generate high-quality segmentation predictions from referring expressions, we propose CLIPUNetr - a new CLIP-d
ents, and execute real-time processing. While current methods attempt to address these issues by integrating multi-scale featu
LiDAR data. Specifically, at inference time we assume that the camera-based detectors have access to multiple unlabeled LiDA
data in real-time. Additionally, our approach incorporates a novel method of generating synthetic OOD data in context with in-d
orrect inaccurate segmentation, such as under-segmentation, of static image-based UOIS masks by using robot interaction and
t pipeline. The uncertainty-aware planner propagates uncertainty associated with the future motion of detected obstacles, and b
pre-trained 3D models through knowledge distillation to enhance feature extraction and semantic understanding in 3D point clo
are not as accurate as those labeled by humans. In this paper, we present the first approach that addresses the inherent ambig
mages and segmentation masks along with physics priors to mitigate penetration and relative-sliding across frames. Furthermor
ptations to the state computation, inverse sensor model, and field-of-view computation tailored to the specifics of radar measur
or completing single-scan point clouds. However, these methods rely on high positioning accuracy or the wholeness of object
conservative free space estimate enables robust detection of moving objects without making any assumptions on the appearan
and VIP500-D, an accompanying set of 20 RGB-D point clouds for the object classes and shapes in VIP500. We then use the
stillation (DCKD) strategy is designed to remedy the adverse effect of RV projection of RGB images. Moreover, a context-based
ociation and enhance segmentation performance by including additional motion cues. We implement attention-based tracking f
els of freedom on motion, i.e., the rotation and the moving shift of the whole BEV, and the random movements of objects (e.g.,

ork’s limitations at run time and act accordingly. To this end, we test our proposed method on a number of test domains inc
y, different vision tasks such as classification, segmentation, and depth estimation can be realized in a unified manner, leading
e local semantics and context dependencies of individual features. To enhance both accuracy and robustness, this paper introd
raw point cloud observations. By randomly selecting very few points for manual labeling, a CID-based point cloud instance seg
orithm, called FocoTrack, which maintains tracking ability in four situations, one of which is when objects are overlapped by eac
e the DA performance in a cost-effective way, we propose a general weak labels guided self-training framework, WLST, designe
data may be missing due to hardware malfunction or weather conditions. This work presents a novel fusion step that addresse
the reliance on highly detailed maps. The efficacy of TerrainSense was rigorously assessed on six diverse datasets, evaluating
r Guided BEV Encoder which transforms camera features into precise BEV representations using the guidance of radar Birdâ€
asses but struggle with the minority classes, negatively impacting segmentation performance. To tackle this challenge, we prop
adar effectively and achieves notable improvement over existing methods.
hat drastically improve annotation times. We therefore propose a discwise approach (DiAL), where in each iteration, we query th
available, they train their initial model from randomly selected data samples, leading to low performance. This situation is referre
nhance robustness, we present an Adaptive Trajectory Prediction (ATP) algorithm based on prediction anchors. In this method
ngle can turn with a small radius. This paper presents a controller capable of selecting and switching between the two aforemen
inclusion of third-order constraints. Approximating inequality constraints can be particularly challenging because the resulting s
nvalues of the target matrix. To improve the performance of these approaches, we present a new parallel-friendly symmetric sta
e NMPC solver that leverages an accelerated preconditioned conjugate gradient (PCG) linear system solver at its core. We sho
al time. This paper proposes a novel approach to computing time-optimal trajectories, by fixing the nodes with waypoint constra
vector coordinates. Methods and software are provided for robustly calculating the invariant descriptors from noisy measureme
to balance multiple objectives or when bang-bang control is unsuitable. In this paper, we present a novel algorithm based on fa
 sue however, is related to the lack of continuity when their stack of tasks is changed. Existing works that address this issue clea
ficient and general-purpose algorithm for risk-averse trajectory optimization. We prove that the method is asymptotically optima
n the drive train. This gives rise to a new class of robots, so-called clutched-elastic robots (CER), with which it is possible to pre
own to be difficult to control and the use of DDP-based algorithms to control them is not yet addressed. We propose an efficien
orce measurements, combined with high-frequency nonlinear model-predictive control. We report an experimental validation on
the modelling of cost functions for optimal control can be done uniformly across different geometric primitives leading to a low s
he system, we then propose to estimate the state-dependent uncertainty with Gaussian Process regression, which improves th
constraints. In this work, we first identify challenges that NCBFs face during training. Next, to address these challenges, we pro
eloped filter minimally modifies the teleoperation input in order to ensure that the robotic system remains stable, safe, and pass
ne systems, use these approximations to generate approximate optimal controllers, and perform regional analysis on the closed
rk via a spring contact model. Moreover, we achieved real-time control for a 7 degrees of freedom robot without any simplifying
ontrol of a rigid body-driven granular material system whose dynamics is learned by a GNN model trained by reduced data gene
hen dealing with continuous state and control spaces, the RL training process may be excessively long and strongly dependen
10th scale hardware, and deployed the proposed LMPC on a full-scale autonomous race car used in the Indy Autonomous Cha
alization of the noise a priori. We are motivated by the future of autonomy where robots will autonomously perform complex tas
Based on this norm-based criterion, we aim at making the maximum energy of external forces admissible for balancing the bip
driven control techniques, such as adaptive dynamic programming (ADP) and robust ADP, encounter a challenge in learning th
nner optimization layer (RTC) simultaneously performs direct transcription (DIRTRAN) to find a nominal trajectory while compu
o multiple kinematically constrained agents. Then, we generate control actions by calculating a consensus among the agents. T
erformance of ReLU-QP across three model-predictive control (MPC) benchmarks: stabilizing random linear dynamical systems
all CODAK, achieves the coordination by relying on minimum set of interfaces that are commonly available on most industrial an
late the guided trajectory planning problem through the lens of a dynamic Stackelberg game. We then leverage Koopman oper
ge) than if it were unimpeded. A support vehicle, which we refer to as a service vehicle, is simultaneously deployed to assist the
of the problem of allocating heterogeneous agents to different search regions in the frequency domain. We relate the spectral c
ng Dec-POSMDP and multi-agent reinforcement learning. Furthermore, we introduce SAMARL: a novel benchmark for coopera
stics of optimizing the crossing order of CAVs at signal-free intersections. We devise an optimal and complete algorithm, Order
cessitate prior environment knowledge. We design a novel decoder that leverages compressed information to estimate the unkn
called non-exclusive reservations specifically for handling wind fields. We train a deep learning model to predict the deviation o
reover, our approach can simultaneously optimize the informative paths while accounting for routing constraints, such as a dista
With the goal of studying and further improving multi-robot planning capabilities in those settings, we propose a method to deve
rediction (CAP) network, we anticipate potential conflict zones, allowing for low-level path planners to explore conflict-free paths
ed approach that does not require building a conservative free-space approximation. Instead, both static and dynamic collision
oller is insufficient, and results in unexpected collisions in tight navigation scenarios under realistic execution. Compared to othe
at allows bounded discontinuities between precomputed motion primitives. The second level identifies inter-robot collisions and
wever, this often makes policies short-sighted, since agents lack the ability to perceive and plan for obstacles/agents beyond th
bots that have traversed previously computed paths entirely (thus on-demand) and reuses the remaining paths for the other rob
k, we extend an existing topology-guided single-robot motion planning method to the multi-robot domain to leverage the improv
rage time that is provably at most four times the optimal. Moreover, we propose two suboptimal yet effective heuristics that redu
collision-free throughout the receding horizon. The N-MPC achieves real time control of a UAV with a control frequency of 100
th feedback of the cable's state – within a much shorter time period compared to frame-based algorithms. Using a point cloud
battery recharging. The system is evaluated in an active outdoor three-phase powerline environment. We demonstrate multiple
 stimation of the quadrotor directly. To train the proposed LPNet, a primitive-based expert policy with privileged information abo
at combines frontier-based exploration with a collector strategy that achieves global exploration and complete map creation. In
mpensate for the limited availability of real-world slant-angle training data. A synthetic dataset is generated using Unreal Engine
tegy for object detection based on a Gaussian YOLO network structure. To solve the low computational resource problem and
(ii) A novel Pyramidal Culling and Aggregation (PyCA) technique that yields robust features while suppressing redundant ones
action framework, for fully autonomous flight in 3D environments to improve flight safety. Our framework only utilizes the onboa
en, we convolve Phi with an attitude-dependent sensor FOV quality function to produce the objective function, Gamma, which q
onents using information-theoretic techniques, these approaches either operate on individual sensor observations, making them
nts neural processors. The DL-based map predictor provides a prediction of the occupancy of the unseen environment while the
ecomposition (SSD). Additionally, theskeleton guides us to effortlessly determine free space. Weutilize the skeleton to efficient
e propose (1) a geometric cue extraction method that processes 3D point cloud data and detects the locations of potential cues
stitute an open challenge, especially when the environment is cluttered and exhibits large and frequent occlusions due to high o
s compressed encoding is combined with an estimate of the robot's odometry and the desired target location to train a deep rei
e-tilted and narrow gaps, and an effective Sim2Real method is needed to successfully conduct real-world experiments. To this e
rrects teleoperated inputs when they are deemed unsafe. We define a point in 3D space as unsafe if it satisfies either of two co
hiskers), we alter the path of airflow as it moves through the array. We have named this technique “flow shadowing” bec
mall onboard computers. To address this issue, we propose a lightweight 3D dynamic obstacle detection and tracking (DODT) m
tory that reduces the error of visual-based estimator while satisfying the constraints on smoothness, safety, agility and the quad
ell environment around a stationary dual arm manipulator, the system can plan and execute procedures to control every plant's
detection of individual updrafts. Instead, a sequence of sensor measurements from a time window of interest can be directly fed
c of the ResNet and RepVGG network that are usually utilized to extract features, a feature fusion module named RepNeck is d
g keypoint-based alignment, rapid encoding, and error correction via ensembles of image patches of natural patterns. We deplo
limits their efficacy in real-time grasping tasks. To address these limitations, this paper introduces a novel two-stage pipeline fo
current fabrication methods face significant challenges in realizing their full potential. In this paper, we introduce thermoforming
ptimal and inefficient robot movements. In more dire circumstances, such negligent handling may precipitate detachment of the
ne features of RGB and depth images as well as a local one specifically working for grasp features in the two modalities. In part
h high accuracy. NeRF-based depth reconstruction can still struggle with especially challenging transparent objects and lighting
This is to address the harsh geometric nature of wiry objects (e.g., severe self-occlusion) and also to avoid a typically expensive
properties, we predict the liquid behavior with fixed squeezing motions in a data-driven way and calculate the required drawing s
alizable implicit function, HandNeRF, that explicitly encodes the correlation of the 3D hand shape features and 2D object feature
termine a suitable grasping location on the stem. To this end, we have introduced a grasp pose ranking algorithm with online le

detection model (VERGNet) to improve the robustness of robotic grasping in low-light conditions. Firstly, a simultaneous grasp
on the Kalman filter. This framework enables the reconstruction of 2-D HR contact surface shapes using collected LR tactile seq
work, we propose to use the best of both worlds by introducing an online estimation method that merges vision-based affordanc
o find this information, tactile sensors measure the deformation field of an artificial skin in both the normal and tangential directio
spondence Network (SCN) which predicts these sets and their correspondences, and an alignment module to compute their re
notated RGB-D data facilitated by an Iterative Annotation Resolver. Furthermore, a self-supervised pose refinement method ba
ensors. A semantic approach to define environmental constraints is introduced and incorporated into a graph-based model of t
al-temporal domain. Additional awareness of bjects' center-of-mass is incorporated into the learning process to help improve gr
e DBScan clustering algorithm. In pose estimation, we employ a deep learning model to predict seven semantic keypoints on th
ereo images, we perform unknown object instance segmentation and generate a 3D reconstruction of the object of interest. In s
ical handling in laboratories and various manufacturing sectors such as pharmaceuticals or beverages. In this work, we presen
work, we leverage interactive probes to visually evaluate grasps in clutter without fully executing picks, a capability we refer to
ect information regarding the geometry of a grasped object during in-hand manipulation with an underactuated hand. By solely
single wrist-mounted camera, which can scan a scene from scratch, detect and estimate the 6-Degrees of Freedom (DoF) pose
t from a viewpoint similar to that in the goal image. Existing image-goal rearrangement systems, due to their reliance on a fixed
otion-blur which are problematic in traditional computer vision solutions. EDOPT uses a strong object prior, leading to a novel s
, we leverage an oracle policy, executed in simulation, which uses the topological description of a mesh of points for representi
nsists of pose data from motion capture systems, we also propose an approach to compute the ground truth linear velocity bas
datasets, and seven popular offline reinforcement learning algorithms in three realistic driving scenarios. We also provide a uni
planning, perception, and control to give information on each aspect and the overall problem. The proposed work gives an over
l-based and model-free 6D robotic grasping on the benchmark, where representative algorithms are evaluated for object perce
st a number of complementary metrics designed to measure the admissibility and diversity of trajectories, however, they suffer
AP and LET-3D-APL, allow longitudinal localization errors of the prediction boxes up to a given tolerance. To evaluate the prop
HuNavSim includes, as a novelty, a rich set of individual and varied human navigation behaviors and a comprehensive set of m
c measure that observes internal system states with minimal interference. Our benchmarking framework provides ready-to-use
h objects, and (2) to study how these properties influence robotic manipulation tasks. Our preliminary results validate the frame
ystem can exert, in order to guarantee the overall safety. These bounds are difficult to estimate a priori, might vary over time an
ng appropriate interaction schemes. We present the first ML-based HRI platform for physical robot to human skill coaching in re
he use of Robo-Coach (robotic coach system). We first utilize a Task-Parameterized Gaussian Mixture Model (TP-GMM) to lea
encoding of a general notion of constancy constraints, which allows to consider dependencies between consecutive or far apart
signs subsystem-based adaptive impedance control to achieve a stable pHREI for a 7 degrees of freedom haptic exoskeleton.
e a model predictive controller that uses an Attention-based Relational Graph Convolutional Network (ARGCN) for modeling the
stment (DDA) conditions in which the speed of the moving objects and air pressure of the PGM vary with respect to the knee sh
lower limb responses for a similar intervention with different exoskeletons, implying that the exoskeleton's external dynamics af
while a low granularity may gloss over the fine nuances of human feelings. Additionally, there’s a notable risk of receiving ca
er a time window in the future. We validated our approach by testing the network on directions of motions that were not provide
work that can generate plausible human grasping motions suitable for training the robot. To achieve this, we propose a hand-ob
or deferring to the partner's lead. In this paper, we propose a hierarchical robot control framework that emulates human behavio
icipate in our lives and help us in various scenarios. Roboticists need tools and data from human modeling to build next-genera
odification in the architecture allows us to obtain a second output from the model with a velocity prediction, which allows us to im
ullfil this current research gap, we carried out an investigation on the above-mentioned problem by altering a cobot’s base r
exoskeletons. This letter presents a novel method to estimate interface power from experimental data in a wearable hand devic
nidirectional platform and a robust aluminum structure mounted on top of it, which provides partial body weight support. A comfo
he-art place recognition methods use sequence matching for more robust data association. However, naively applying these se
e assumption in many applications such as autonomous driving. The prediction regions are valid, i.e., they hold with a user-defi
Then, we integrate DRSCCs into the trajectory optimization framework using Bernstein basis polynomials. Theoretically, we rigo
mulations used by MPPI require significant onboard computational resources, making the algorithm inaccessible to robots witho
y efficient safety filtering solution to reduce the collision risk of ego robot motion plans using multiple samples of obstacle trajec
entially. As part of this work, we present a novel algorithm, Hybrid Belief Monte Carlo Planning (HB-MCP) that utilizes the Monte
component corresponds to a different data association hypothesis. However, this method can lead to an exponential growth in
 e consider safety requirements represented as almost-sure reach-avoid specifications (i.e., the probability to reach a set of goa
g the optimal set of edges to add to the roadmap without considering an exponential number of traversability scenarios is challe
stems in complex environments, without exact knowledge of environment geometry. We combine a novel implicit neural model
ce Constrained Options Belief Tree Search (COBeTS) to leverage this hierarchy and scale online search-based CPOMDP plan
on the learned incomplete scene geometry, which explicitly identifies these unseen regions in the scene. By considering the acc
ng-based motion planning approach that uses an estimate of the model error and online observations to correct the planning str
nt in the agent's favor. The proposed planner, Dynamic Autonomous Exploration Planner (DAEP), extends AEP to explicitly pla
aversability constraints and an optimization formulation used for the coverage. We highlight the benefits of our approach over st
ominal trajectory graph. Finally, we grow a belief tree by searching the graph using the proposed heuristic. IBBT interleaves ba
ce-constrained model predictive control (DRCC-MPC) which: (i) adopts a probability of collision as a pre-specified, interpretable
riggered that leverages the same GP model to find a safe state from which the robot may continue towards the goal. The propo
eal world may be different than in the determinized world. This causes collisions if the determinized world optimistically ignores
earance between nearby viewpoints. To address this, we propose a new active perception framework incorporating an arbitrary
Processes (TV-POMDP). We propose a two-pronged approach to accurately estimate and plan within the TV-POMDP: 1) Mem
generalization presents a significant challenge. In this paper, we propose running several computationally cheap SLAM front e
ntaining high computational efficiency. Building upon the MI approximation, we develop the Adaptive Particle Filter Tree (APFT)
the task. We present a preprocessing-based planning framework for high-precision insertion in repetitive and time-critical settin
aining adequate datasets. To address these limitations, we propose to employ ensembles of deep semantic segmentation netw
checking, which trade off computation with accuracy. Our algorithm generates motion plans driving each robot from its initial to
s. The planning algorithm involves new metrics that quantify heteroscedastic uncertainties in the rover traversability prediction m
signed to optimize the time progress of the trajectories temporally, serving as a post-processing tool to enhance trajectory time
ge graphs and teams. To address this curse of dimensionality, we propose to use RL to enable agents to learn such graph trave
orcement learning-based algorithm. To effectively handle precedence relations and global information, we propose a policy net
e treatment of topology and geometry: Topological invariants are singularly utilised for distinguishing non-homotopic property a
pair of vertices in the set without passing through any additional vertices of the set. Identifying an LWCS has many potential hig
fferent locations (e.g., prefabricated rooms, crates of material/equipment), where the weight of each payload defines the require
y by proposing a hierarchical approach that decouples the problem into two layers. We use a low-level optimization formulation
ncludes a traffic flow prediction layer, a sector-level planning layer, and a road-level coordination layer. In traffic flow prediction,
We extended the MILP formulation to account for complexities with legged robot teams. Our results demonstrate that this appro
and weight. In this paper, a powered transtibial prosthesis using a Parallel Elastic Actuator (PEA) is proposed in order to genera
ive. In this study, we proposed a novel deep-learning structure to improve the accuracy of bone tracking by an A-mode ultrasou
ong-Short-Term Memory (CNN-LSTM) models to generate exoskeleton joint trajectories. Our methodology involves collecting d
y F1-score, recall (true positive rate), and precision, which are 96.3% on average with a standard deviation (std) of 1.9%, 96.6%
or the different parts of the stride. In contrast, this paper presents a controller that uses smooth impedance parameter trajectorie
ant crank slider mechanism is also developed to power-assist AFP mechanism to improve the adaptive compliant contact betwe
mptomatic users. The force fields mimic common perturbations encountered in exoskeletons, such as residual friction, over/unde
ntroduces neural information of motor neurons in conjunction with the combination mechanism of muscle contraction. The inter
king principle can accommodate the angular difference between the hip joints and provide the same assistive torque at both hip
ROM. In addition, we designed the artificial muscles using the properties of human muscles, such as moment arm and displace
me while weighing 1.8 kg. Furthermore, a cable routing method is proposed to actuate the plantarflexion of the ankle. The kine
ation analysis recommends placing the CoM in front of the user. We address historical constraints on front-enveloping designs a
 machine learning models, we effectively distinguish motion patterns and we identify transition moments between locomotion m
mpedance controller, promoting cooperation between the exoskeleton and the user. Moreover, our framework is based on the
ked with the profile optimized to them, the two profiles optimized to the other two participants, and the average of the three torqu
ontrolling the prosthesis based on these system-level, or task-space, variables would further represent a new way of coordinati
esents ongoing challenges. The integration of computer vision techniques, particularly object recognition algorithms, can greatly
es in a set of model variants. We find that by adjusting the candidates during robot deployment, our geometry-informed distanc
ap in, to allow an algorithm designed with a projection model to accept data of a camera calibrated with another model. So, a pr
usly reported, existing methods cannot be used on most UAVs, which have limited computational power and energy; thus, exis
registration approaches. To mitigate this, we introduce an object-matching network based on self- and cross-correlation that ca
ption of local geometric models, an adaptive adjustment of voxel segmentation according to the point cloud distribution and den
d on-tree down-sampling. The i-Octree is built upon a leaf-based octree and has two key features: a local spatially continuous s
pth sensors error while preserving geometric and map consistency details. Despite the effort, depth correction of lidar measure
to solve the terrain estimation and point cloud ground segmentation problems. We evaluate the ground segmentation and terra
rom-silhouette model to reconstruct complete object shapes. We use the proposed method to exploit strong geometric and perc
e policies, the proposed navigation system facilitates precise, low-latency (36ms), and extremely efficient obstacle avoidance w
re developed and tested with simulated and physical experiments: a wall-follower, a navigation algorithm based on Theta*, and
ses should be set to large. However, the large noises yield non-smooth control input sequences, resulting in non-smooth paths
ured points. Furthermore, we capture and encode robot performance through capability metrics and improve the same using a
ergent time and the local goal state, we present a high-level planning node encompassing terrain-aware kinodynamic search an
ork using mixed reality technologies to ease the use of RAMS. The proposed framework achieves real-time planning and execu
r, we propose the keypoint-based visual imitation learning (K-VIL) approach that automatically extracts sparse, object-centric, a
gh measurement delays and inaccuracies. Our extension enables the definition and effective execution of complex tasks in soc
d to select a safety-aware subgoal that leads the robot to its final destination. The algorithm is designed to run in real-time and i
ving the performance and usability of the UHEx design, which was initially optimized using a naive bruteforce approach, by integ
of assistance on post-stroke gait during overground, community-based walking. We first test our device with stroke survivors an
ntaining interaction forces at target values, even with motion disturbances. To this end, we developed a shoulder actuation testb
hat could be incompatible with these controllers. This paper addresses a new way of generating compliant trajectories for passi
ning while respecting a set of equality and inequality constraints designed to minimize the risk of stumbling. The used approach
practical RL setup to control FES for arm movements. We also present a flexible method, called co-kriging adjustment (CKA), w
covery. In this paper, we present an adaptive hybrid robot-FES controller to do this and enable the triadic collaboration between
ing stretching exercise games for older adults with PD. SWS consists of a compact and lightweight mobile manipulator (Hello R
se of training efficiency. In this paper, we study the problem of constrained offline GCRL and propose a new method called Rec
hat are transformed into safe trajectories by a trajectory optimizer, thereby effectively ensuring safety and increasing training sta
dens and diminished stochastic behavior. In our work, we present a novel DRL framework that leverages the Gaussian mixture
ed by the CBF configuration, highlighting the necessity for a responsive and dynamic approach to CBF design. We present AS
uses them to initialize the agent in the environment in order to guide it toward more informative states. We validate our approac
y shaped using a safety critic. The advantage of the textit{Safety Optimized RL (SORL)} algorithm compared to the traditional S
ion of field robots that will need to be able to learn on the edge in order to adapt to their environments. In this paper, we begin t
a common policy with dense task-specific corrections through task-specific layers to better express shared and variable task in
sequence. It recurrently utilizes detected lanes within historical frames as prior information to achieve robust lane detection. Add
gradually being deployed. However, when vehicles encounter uneven roads, vertical vibrations lead to mismatches between ex
hardly contribute to policy improvement, resulting in slow or failed learning. We propose to improve action selection in this mode
gorithms, has not received adequate consideration in research. This paper aims to address this gap by proposing an attack on
s and representative scenarios. Moreover, we employ point clouds to represent our exploration scenarios, leading to high-fideli
ploitation dilemma and the dynamic gap between inaccurate simulation and the real environment. To address this, we propose a
locity actions as countermeasures to stabilize the drone. REFORMA consists of two components: a base policy trained comple
high-dimensional pattern for robot actions. It also integrates two HD-based lightweight symbolic learning paradigms: HDC-based
ation approach consisting of a material property (e.g., flexibility) estimation and policy learning for DLO insertion with reinforcem
event detector, necessary for the task's progress, may increase the uncertainty further. To cope with those issues, we propose
ied multi-sensory object property representations. While it is undeniable that visual information plays a prominent role, we ackn
t retrieval framework that efficiently performs rearrangement planning of unknown, arbitrary objects in confined spaces to retrie
g., the scenes in the LLFF dataset), where the input cameras are all observing in similar directions with small relative translatio
ons parameterized by their start and goal configurations. This allows to quickly compute signed distances for any point in the tas
s implicit representations. In this paper, we present Reg-NF, a neural fields-based registration that optimises for the relative 6-D
ocal and global geometric features apart. To address this problem, we present a novel and efficient self-supervised point cloud r
arious entities in the scene, capturing "where" information. To each such segmented entity, we apply other pre-trained models t
modate the shape variations of supporting items. To this end, we propose Semantic Keypoint Trajectory (SKT), an object-agnos
aper, we propose a transformation from spatial measurements to a coordinate-invariant feature space, based on the pairwise in
ed object discovery and uses a feature transport mechanism to maintain temporal alignment in object-centric representations. T
of their trends in the real world, 2) the use of PVRs enables a first-of-its-kind result with indoor ImageNav (zero-shot transfer to a
he robot’s transformations through an external signal (inertial sensing, for example) and give attention to image space that
Cloud World Model (PCWM) and point cloud based control policies. Our experiments show that policies that explicitly encode po
continuously. However, they are hard to update incrementally as neural networks tend to forget previously observed parts of th
odels that bridges this gap without any further training. Rather than construct representations out of only the final layer activatio

xtraction strategy in this paper that balances the accuracy in boundaries and computational efficiency. Our method begins with
e lifting. Specifically, we leverage the generalization capability of large visual language models (Dalle-2) to inpaint the missing a
ey insight is that the correlations between each model point and scene point provide essential information for learning point-pai
ce map by propagating sparse hints using a learned deformable window. These maps then guide cost aggregation in stereo ma
ock that combines a GRU recurrent unit with transformation synchronization, blending historical and multi-view data. Our appro
en overlooked but laden with substantial risks. MMAUD stands out by combining diverse sensory inputs, including stereo vision
c manner. We introduce a zero-shot split for Tabletop Objects Dataset (TOD-Z) to enable this study and present a method that
), a new transformer architecture that simulates the von Mises-Fisher (vMF) mean shift clustering algorithm, allowing for the joi
me, we introduce Gaussian-decay Depth-prior Projection (GDP). This module projects the 2D features into the 3D volume along
y adapt the model from clear to relative adverse, and then to adverse weather scenes. It encourages the model to gradually gra
ego vehicle in accident-prone scenarios. Our approach formulates collaborative decision-making as a classification problem. W
g a static scene. Moreover, it provides a mathematically grounded formulation for relevant robotics applications, e.g., where a d
gment a DLO in 2D images and post-process masks to eliminate false positive segmentation, reconstruct the DLO in 3D space
e transient histogram directly to enable recovery of planar geometry more accurately than is possible using only proprietary dis
grating VLM into the domain of 4D point clouds remains an unresolved problem. In this work, we propose the Vision-Language
s between entities in the environment, which are useful for downstream planning. In this work, we propose ConceptGraphs, an
e application of Motion Coherence Theory to impute the spatial velocity of occluded nodes, the use of the topological geodesic d
 ta re-sampling, loss re-weighting, and transfer learning. In contrast, our work proposes to effectively utilize network classifier w
ed compliant joint (T-C joint) design, which utilizes the torsion deformation of materials. The potential of the T-C joint is demonst
ckers are commonly used in large solar farms, they are rarely used on rooftops, an area where solar panels are commonly inst
ieved through actions that deposit or remove bundles of atomic building blocks to form higher-level nonparametric macrostructu
gned as the actuator to control the transition between the two stable states. To facilitate gripper design, a simplified model of th
ingularities, which is essential for its optimal use. However, for various reasons, robot arms with unorthodox architectures are o
optimal parameters that maximized speed and minimized the cost of transport (COT). The maximum speed recorded was 0.20
econd) and a minimum cost of transport (COT) of 2.9. The robot has two degrees of freedom (DoFs) and is capable of perform
orientational workspace by exploiting the redundant DOFs. Moreover, the orientational workspace can be further increased by i
s. The main contribution of this research is the design of frictionless pulley-guided nodes. To validate the proposed concept, we
an we instead train ``generalist’’ cross-robot policy that can be adapted efficiently to new robots, tasks, and environme
inear-attention counterparts maintaining high quality. We demonstrate the effectiveness of SARA-RT by speeding up: (a) the c
grasping with two soft fingertips. Through comprehensive evaluations, the DenseTact-Mini demonstrates a lifting success rate
this paper, we address this gap by using contact microphones as an alternative tactile sensor. Our key insight is that contact m
n `plan'), followed by `translating’ this plan to the robot’s embodiment. Specifically, we learn a human `plan predictorâ€
kinematics problem to parametrize the configuration space, resulting in a lower-dimensional representation where the set of va
loyed in benign environments, making them ineffective for scenarios where robotic systems experience sensor faults and attac
bots to stay in predefined safety regions with desired confidence under these stochastic uncertainties. BCBFs are general and
o move from any state in the set to any other state in the set within a given time horizon without having to leave the set. We cal
ling factor, with an offset, to systematically define a CBF for obstacle avoidance tasks. We provide theoretical analyses of the c
ontrol policies directly. However, these methods require large training sets that must be collected and curated. In this regard, w
mization-based moving horizon estimator (MHE) to characterize the disturbance affecting the system, which is incorporated into
paper aims to bridge the gap by proposing the use of the linear impactor, a well-established tool in other standardized testing di
nd mitigate such closed-loop, system-level failures. Specifically, we leverage a reachability-based framework to stress-test the v
namics and the actual system. This learned disturbance model can be used in conjunction with probabilistic safety methods suc
obability. We apply the proposed approach to study a vision pipeline deployed on a quadcopter. With the proposed approach, w
at the robots reach their exploration targets while avoiding any obstacles discovered and each other. We employ a scalable ap
ystems, such a set cannot be exactly computed in general. Numerical algorithms exist for computing approximate control-invar
roposed in the literature, but they suffer from the curse of dimensionality. This paper presents a new approach for numerically a
structed dense map, along with object-level semantic and consistency estimates into a control barrier function (CBF) to regulate
e main contribution of this paper is to provide a quantitative assessment of how MR actuators can contribute to reducing the imp
onstruct safe sets in the underlying belief space. We design a controller that provably keeps the robot's belief state within this sa
they come with provable statistical guarantees of having low risk without any assumptions on the world model whatsoever; the
control invariant sets via a feasible quadratic program (QP). However, BCBF performance largely depends on the design and c
ance to rotation, translation, and scaling to estimate the transformation without relying on point correspondences. Based on gra
alizing flow estimates the probability density of features under the training distribution. The label distribution and density estima
ultrafast and numerically-stable square-root filter (SRF)-based VINS algorithm (i.e., SR-VINS). The numerical stability of the pr
tinuous frames. Unlike previous methods that primarily adopted three strategies: matching two consecutive point clouds, predic
elds and performing a quick adjustment to the loop closure. The fast convergence and rapid response to loop closure make our
propose Universal Visual Decomposer (UVD), an off-the-shelf task decomposition method for visual long-horizon manipulation
branch dynamic system. We performed 756 apple picks on the proxy with varying branch stiffness, stem strength and gripper p
 y can save time and labor costs. Common methods include methods based on feature size and learning-based methods, with
ed system for grain appearance inspection, called AV4GAInsp, which consists of a device and an analysis framework. The devi
an unmanned aerial vehicle over a wheat crop field. The dataset is registered using photogrammetry aided by fiducial ground co
vines and demonstrate initial steps towards estimating pruning weight in high cordon procumbent (HC) vines such as Concord.
which is the problem we address in this paper. To this end, we exploit recent advancements in 3D deep learning to build a con
odeled tree crop as a graph. This representation allows us to employ graph networks to learn both a forward model for predictin
gs by using multiple simultaneous localization and mapping systems and navigation systems. The results revealed that the path
well. Obtaining such labels is challenging as it often requires background knowledge on the side of the labelers.This paper add
nsuming and difficult to perform tasks, by collecting multiple leaves and automating some parts of the overall SWP analysis pro
The perception of the tomato is done by a fixed depth camera and a camera without depth on the gripper. The proposed dual-c
ations in fruit size, shape, color, and occlusion. In this paper, we develop a few-shot semantic segmentation framework for infie
ges. In the first stage we develop a robust skeleton extraction method from unstructured plant point cloud data by adopting Hau
estimation of phenotypic traits such as leaf area, leaf angle, and plant height. We validate our system on a real field, where we
e analysis using the concept of hybrid dynamical systems is required. To demonstrate the dynamic characteristics of the object-
ative navigation (VACNA). We propose a Cross-Entropy Method (CEM) based approach for solving the trajectory optimization u
ing CBF (DCBF) respectively, eliminating the need for an external planner. The control scheme has a backstepping approach f
environments such as agricultural environment has been under-explored. A key challenge here is that concepts and structure o
tion is generally not supported by currently available tools. In this paper, we present a methodology which allows the efficient re
ts electrical control, mainly composed of a servo-electric two-finger parallel clamping module, a swing-cutting module, and a fru
s it and plant seedlings of cabbage accurately. Path generation is performed considering vehicle tilt in the roll direction dependin
ethod is suitable to be implemented on edge devices necessary for the application: we are using a mobile robot with the NVIDIA
, planar orchard architectures with simple pruning rules that combines various subsystems from our previous work on perceptio
frared [RGN] camera) and 3D LiDAR point clouds, via a customized tree landmark association and parameter estimation algori
e strategically select viewpoints that target the small fruits' volume. Additionally, we integrate a dual-map representation of the e
of perception. Our formulation focuses on quickly improving the perception accuracy of a single target node to maximise its cha
ion from a farmer to extract the motion constraints associated with the tasks, namely, transplanting and harvesting, and can the
mination for self-supervised representation learning. INoD interleaves feature maps from two disjoint datasets during their conv
ce of machine vision systems for new fields and different robotic platforms without additional manual labeling. In an unsupervis
ne for the log loading operation with a fixed-base manipulator (a.k.a., the crane), starting with perception of logs around the ma
region, we define the intersection area of features' enlarged radius and desired operating area as negative and positive virtual
ct an inventory, we first extract tree instances from point clouds. Then, we process each instance to extract forestry inventory in
nes transfer learning and active learning through a Multi-Task Gaussian Process and an information-based objective function. T
her neighboring robots. Once the encirclement is complete, the robots transition to drive the animals toward a designated goal b
ng control using proximal magnetic sensors. Next, we propose a framework for dynamic reconfiguration and task-space motion
teraction adequately while allowing execution orders of magnitude faster than real-time. We evaluate PLASTR in simulation and
le into reconfigurable robots via a reversible solder joint. The robot assembles higher performance structures using a set of con
cription, and correspondence grouping using learned descriptors to keep track of an open set of objects. Object correspondenc
uster at a time. We employ movable visual fiducial markers as infrastructure and an autonomous underwater vehicle (AUV) for m
ader in the real world and demonstrated that it's possible to fine-tune a neural network controller that was originally trained with
-learning methods, the domain parameters are often uniformly sampled; this implicitly assumes that the difficulty of the task doe
ers. This paper presents a robotic inspection system that uses vision-based positioning and tags NDE measurement with pose
ent state. The robot then zooms to the fine-resolution inference part to perform the concrete action corresponding to the action
encounter in our daily lives. As a consequence, a promising solution to the limited representation in previous grasp datasets is
les. Our framework leverages the generic knowledge of LLMs through a small number of prompts to perform high-level task an
hods using Large Language Models (LLMs) and those using Deep Reinforcement Learning. Second, we introduce a method we
updated state estimates. In this paper, we present an interactive planning technique for partially observable tasks using LLMs. I
to enable optimal hierarchical LTL planning with LLM guidance over scene graphs. To achieve efficiency, we construct a hierar
nditions. Our approach enables embodied agents to execute more tasks than baseline methods while maintaining semantic cor
e GraspGPT, a large language model (LLM) based TOG framework that leverages the open-end semantic knowledge from an L
ponents that uniquely differentiate DOS from existing robotic systems: (i) A cloud orchestrator that build and schedule pipeline
the system due to the reduction of its control parameters while providing practical finite-time convergence, robustness against
avoid the two-subsystems control paradigm generally adopted by conventional UAV controllers; (ii) all control inputs to be defin
c convergence of tracking, estimation, and identification errors is proved in the nominal case and the stability conditions of the
predicting the successor state, the controller can solve the OCP one time step ahead of time thus avoiding the delay of the nex
nstrumentation and provides a stable online parameter estimate, enabling robust model-based autonomy.
fline model identification methods or rely on initial parameters within a narrow range (typically within 15-20% of the actual value
output-feedback Model Predictive Control (MPC). By taking a pessimistic estimate of the value function resulting from the MPC
derivative function of the low-pass filtered switching control is appended in GTSMC, by virtue of the accuracy of integral and th
ose relative to joint-dynamics, flexibilities, external perturbations, etc. Then, we propose a robust QP control formulation based
me safety framework during physical interactions with the surrounding environment, including humans. To validate the effectiven
ng DS research often assumes demonstrations solely in Euclidean space, overlooking the crucial aspect of orientation in variou
mputation resulting from conservative model simplification. To attain true optimality, we require a method that can compute dista
for the robot that ensures a handover that avoids physical contact with the person, especially in scenarios when hands and arm
g robots that can safely and seamlessly interact with humans, we propose a general robot trajectory planning framework for syn
Networks (LSTM) based inference modules to capture both driving styles and rationality from the observed motion of human dri
an-Robot Collaboration safety predicated upon basic perception metrics: Critical Collision Probability (CCP) and Average Collis
easible. This is achieved by constraining a dynamical system based impedance control law with a relaxed hierarchical control b
e expensive to acquire. In this work, we try to balance the need for quality data with the goal to minimize cost by selecting the m
and those that are possible in the underwater domain. A necessary component to support UHRI is establishing a system for sa
ng adjacent conductors to probe line components with micro-Ohmmeter, installing and retrieving custom build sensors for mult
ized for the Dragon Firefighter, a jet-actuated flying continuum robot on a mobile base, for rapid and safe fire extinguishing usin
ployment of cellular-enabled collaborative robots in SAR missions is their energy budget, primarily supplied by batteries, which
ocations as frontiers, based on the direct pointcloud visibility, allowing fast graph building and path planning. Additionally, the g
ves $frac{1}{2}(1-frac{1}{e})overline{OPT}$, where $overline{OPT}$ is the approximate optimum. The experiment results show t
methods in multi-robot search.
ere specific deployment patterns allow for guaranteed k-connectivity, making the resulting network redundant while keeping cha
with a path-following vector field and a state-of-the-art SLAM algorithm. The methodology allows the robot to explore its surroun
ection, update and dependency of variables. Then, we present a first implementation of NaivPhys4RP that demonstrates the le
on of the knowledge of the agent. To enable region identification, the method uses a vision-to-language model to provide scene
semantic constituents and employs a visual grounding tool, such as OpenScene or LERF, to identify objects in a 3D scene. Th
ersability in new circumstances. To this end, we first generate self-supervised traversability labels from past driving trajectories
ed point distribution by increasing the volume of voxels along the radial distance to include fewer near-side points in a smaller v
 o enable few-shot panoptic segmentation by presenting Segmenting Panoptic Information with Nearly 0 labels (SPINO). In deta
al closed-loop bipartite matching algorithm to establish a total loss function consisting of the unsupervised illumination enhance
Fusion (DefFusion) framework consisting mainly of a Deformable Representation Fusion Transformer and Dynamic Representa
ely remap. Human users can query inventory by providing natural language queries and receiving a 3D heatmap of potential ob
authoring, visualization and simulation of complex environments. We investigate the conversion of USD-based environment mo
projected feature maps from previous frames with current-frame features greatly improves image segmentation quality, compare
uch as distance and angle to the frontier, similar to existing frameworks, with the key difference being the additional utilization o
uncated Signed Distance Function (TSDF) for swift 3D scene reconstruction. By leveraging the VLFM, we extract region-based
dicts semantic instances and their temporal associations without relying on any hand-engineered non-learned association strate
the scans consistently over whole sequences. Previous approaches for 4D panoptic segmentation rely on post-processing ste
ese methods are plagued by repeated exploration, effective-knowledge sparsity, and wrong decisions. To alleviate these issue
on of Expectation based on Correlation Information (CECI), to reasonably approximate real Belief/Expectation, by learning histo
LP) to the vision domain as a way to embed visual scientific knowledge into the estimation process. This modification and addit
here the sensor for 3-D information acquisition is unavailable, or just when multi-sensor calibration falls short of the necessary p
proach to robustly segment 3D point cloud scenes. Specifically, we first build a multi-prototype graph and then suppress noisy
o enhance scene interpretation for safety-critical tasks. Our Radar Instance Transformer enriches the current radar scan with te
re presented and compared on the ScanNet dataset. We show that the most widely used Bayesian fusion strategy is among th
construction through the use of satellite maps to complement onboard sensors. We initially generate the satellite map tiles for
mplementary random masking strategy of RGB-T images and 2) self-distillation loss between clean and masked input modalitie
utomated driving. CURB-SG leverages panoptic LiDAR data from multiple agents to build large-scale maps using an effective g
rsity in the entire BEV space. Toward that end, we devise Point-scattering that scatters LiDAR BEV distribution to camera dept
ework designed to search for safe and energy-saving air-ground hybrid paths. AGRNav contains a lightweight semantic scene
era and an inertial measurement unit (IMU), enabling crowdsourcing training data from commercial vehicles. Bringing the visual-
-based methods. To achieve this, we propose a unified information-theoretic surrogate metric for camera and LiDAR evaluation
In this paper, we propose EOLO, a novel object detection framework that achieves robust and efficient all-day detection by fusi
ose a Bidirectional Fusion Network with Cross-Modality Knowledge Distillation (CMDFusion) in this work. Our method has two
when only IR images are available. To address this issue, this paper proposes a novel contrastive learning-based spectral know
. The objective of this module is to improve the point cloud quality and is independent of the detection network used for inferen
HD maps in both short ranges, i.e., within 30 m, and also predicting long-range HD maps up to 90 m, which is required by down
ed validation of the quaternion spline fitting method, and evaluate the proposed system, SFUISE (spline fusion-based ultra-wid
d to different precision, robustness, and usage dependencies. To clarify the capacity of GNSS algorithms and accelerate the de
bject has a complete columnar structure in the point cloud composed of motion-compensated events along the timestamp. Mot
ce long-distance detection even amidst data discrepancies. AYDIV consists of the Global Contextual Fusion Alignment Transfo
We propose a Radar-Camera framework for highly accurate and fine-detailed dense depth estimation with four stages, includin
s for outdoor environments, especially in the context of autonomous driving, there is still a notable lack of datasets specifically d
on exchange between sensor modalities (RGB and depth). Our approach incorporates self- and cross-attention layers at differe
alibration parameters. Consequently, the extrinsic parameters are optimized for cost minimization based on the intensity varian
ecially in regions with marginal light changes. Therefore, direct fusion methods, e.g., RAMNet, often ignore the contribution of th
e extended Kalman filter and an interacting multiple model (IMM) filter are tested using both simulated and experimental datase
s, which are the first sensors to become inoperative. While the structural integrity of the robot is at risk, the robustness of the S
racking work that incorporates point clouds together with RGB videos to achieve information complementarity. Moreover, PVTra
 omorphic computing framework LAVA for offline training and evaluation on an embedded system. For the evaluation three open
een examined. However, a seamless transition from impact reduction to contact impedance control has yet to be demonstrated
chnique enables safe navigation through unstructured and dynamically changing environments and is shown to be significantly
aw producing a decoupled exponential decrease of the sensor errors in perfect conditions, while being more robust to estimatio
method is designed to adjust image exposures within the brackets online. Gaussian process regression fits the relationship betw
technique, which simplifies the manufacturing process while ensuring repeatability of production. Compared to traditional visio
g. The first training step enables our model to learn from multiple datasets covering various real-world situations, allowing it to r
a is collected at 1, 3, and 5-meter distances. The model performs well even at a 5-meter distance with a Word Error Rate of 40.
write: a coarse-to-fine approach for robot arms to discover and recover plausible writing orders from diverse calligraphy images
are slow but accurate, or by encoding such knowledge into the weights of a neural network, leading to faster but often less acc
r images to orbital maps using a modified census transform to achieve sub-meter accurate, near-human localization performan
and unstructured environments. By employing a two-layer hierarchical policy structure, we exploit the capabilities of a highly rob
three types of failure in more realistic GPSR application settings: insufficient information, incorrect plan generation, and plan ex
nces resulted from an infant's limbs laid on part of the quilt. Building upon prior research, the EM*D deep learning model is emp
specific scenes but face issues when generalizing to novel environments. In this paper, we encode keypoints and corresponden
hods and the large convergence of learning-based VS methods. NUVS estimates the calibration embedding from past observat
ultitasking medical service robot designed for telemedicine aspects and to facilitate remote medical education activities such as
ator with a spherical parallel mechanism as the continuous segments. Due to the spherical parallel mechanisms’ character
maintain recognition distance and accuracy. The proposed marker achieved a measurement accuracy of 0.1% position error an
ds on the requirements of the task and the capabilities of the robot's hardware. In this paper, we propose RoboKudo, an open-s
be used for mass customizing the morphology of a robot to improve area coverage performance in an environment of interest.
ation. At its core lies a novel replanning strategy that handles physically grounded, logical, and semantic errors in the generated
e large time steps, and specialized sparse algebra, our method runs robustly at interactive rates. We provide validation results
ralize effectively. In this paper, we introduce a data-driven crowd simulation method called SocialGAIL, which leverages Gener
arios as often needed in multi-robot scenarios, where the different environments have varying numbers of agents. In this paper
ly used in robotics either cannot couple deformable and rigid bodies in the same framework, lack important systems such as cl
amic interactions among schooling fish. We analysed two scenarios: one driven by real fish kinematics in spatiotemporal forma
mulator. This simulation provides fast procedural environment generation, multi-robot capabilities, along with synthetic data pip
eneControl learns an expressive diffusion model from data. Then, using guided sampling, we can flexibly control the sampling p
We derive the gradient calculation to ensure the whole simulation process is differentiable under the backtracking mechanism. W
n. This paper addresses these challenges through a case study on creating a model for a highly dynamic logistics robot. The stu
on loss fully adapts the keypoint detector to any jointly learned descriptors and benefits the self-supervised learning of descripto
e for sensor fusion approaches. Furthermore, alternative environmental representations such as semantic Bird's Eye View (BE
ts with high intensity gradients. Moreover, DISO is implemented to incorporate a multi-sensor window optimization technique, a
3D maps with a continuously varying density but also efficiently reduce map storage space by using CURL's spherical harmonic
y leveraging a complementary perception modality while preserving the accuracy of LiDAR in well-conditioned environments. T
ntage of the overlapping fields of view (FoV) by extracting cross-matched features across cameras in the rig. This limits the line
ensure successful initialization, we propose an efficient strategy that comprises three different methods: stationary, visual, and d
e benefits of neural implicit field and feature-metric optimization. This hybrid method optimizes a multi-resolution implicit field an
ect level to improve localization accuracy and enhance reconstruction details. Moreover, our approach detects and eliminates d
formed in the Markov blankets of affected landmarks. It transforms pose-landmark constraints into pose-pose constraints while
 of local maps generated from locally consistent odometry estimates. Several recent approaches compute the maximum elevati
LiDAR scan and propose a ring-based fast approximate least squares (Ring FALS) method. With the Ring structural informatio
arios where the sensor is being blocked by large dynamic obstacles. It exploits a constant velocity filter and the measured Dopp
ble of handling large numbers of outliers are yet to come. This paper presents IPC, acronym for Incremental Probabilistic Conse
ate-of-the-art (SoTA) plug-and-play correspondence matching approach. First, we eliminate the pre-defined threshold employed
or ellipsoids, are less accurate than feature points. In this paper, we propose a Visual Object Odometry and Mapping framewo
ssue or exploiting a too heavy architecture. In response to this challenge, we propose a groundbreaking solution within the texti
polants. We demonstrate the suitability of rational quadratics for pose prediction by applying our approach to real data from the
ed up the mapping operation, we propose an efficient strategy of edge-pixel sampling according to the local dynamics of events
1) a keyframes layer with robot pose estimates, (2) a walls layer representing wall surfaces, (3) a rooms layer encompassing s
tion of the VPR problem, a general-purpose algorithmic pipeline, an evaluation methodology for VPR approaches, and the majo
radar data does not present clear geometric features. Moreover, disparities in radar data resolution and field of view(FOV) can
ights into the underlying map regions. Experiments on public large-scale datasets demonstrate that our technique surpasses th
ctor data association module. We show that the proposed network dynamically learns and adjusts the confidence maps of both
point cloud into uniform grid cells whose contents are subsequently modeled by normal distributions. We show that the propos
d builds the spot uncertainty model to quantify the ranging error. Generally, the number of variables in continuous-time estimati
vel information-theoretic loss function that accounts for the fact that different regions of the map may be learned to varying deg
er, we propose LIO-EKF, a tightly-coupled LiDAR-inertial odometry system based on point-to-point registration and the classica
tilizing a novel parallel strategy for state updates and a voxelized map format, Multi-LIO optimizes computational efficiency. Fur
formulations for Dynamic SLAM, representing dynamic object points in either the world or object coordinate frame. While expre
ements with covariance of a larger plane (father plane). Thus, we design a plane merging module based on union-find which ca
features can be analytically solved, drastically reducing the dimension of the numerical optimization. To represent and solve th
oarse-to-fine offline framework that exploits global 4D spatial-temporal LiDAR information to achieve clean static point cloud ma
ing module takes the registered LiDAR scan for incrementally reconstructing the triangle mesh on the fly. Finally, the real-time
e sensors. In this work, we conduct an information-theoretic investigation of this overlooked element of robotic perception in the
mplementary attributes in both key-point identification and descriptor expressiveness. To this purpose, we design a pre-process
he image plane. To avoid convergence to a local minimum in nonlinear optimization, we separate the pose into translation and
ant, we propose a trade-off optimization algorithm between maximization of the Fiedler value and minimization of the largest ei
ed covariance matrices by formulating the learning process as a constrained bilevel optimization problem over factor graphs. W
own to be used in anchor-tag SLAM systems. One or more anchors are installed in the environment and the tags are attached t
tforms and sensor types. Our quantitative results demonstrate that the proposed radar-inertial odometry system outperforms al
sing four-point LiDAR; subsequently, we build a reliable global Manhattan world (MW) map while simultaneously estimating 6-D
oarse-to-fine candidate selection from the database. And a revisit can be identified following geometric verification. In addition,
tion. The approach first builds a topological SLAM graph of the environment (using visual-inertial odometry) and embeds Contra
mic element - any rigid object free to move in the scene - whose pose can be reliably inferred from a single image. Specifically, w
rmulation of LiDAR residuals is introduced, expressed only in terms of the occupancy fields and its respective gradients. These
es how active visual localization can be used to overcome such challenges of viewpoint changes. Specifically, we focus on the
icit surface uncertainty-based reconstruction tasks to achieve high-quality reconstruction. and 2) introduce a method to achieve
n a pose graph, which considers both how much pose graph uncertainty would be reduced by getting a loop closure there, and

at one common time. This grouping simplifies the estimation problem but, absent additional sensors, sacrifices the inherent tem
alman filter (IEKF) approaches that simply tack IMU bias and other states onto the SE2(3) group, our filter builds upon a symme
e line. Then, we utilize multi-frame data of the lane lines and the curvature characteristics of adjacent lane lines to optimize the
ay cause a collision, in various crowd environments, which reduces the reliability and the safety of the robot. In this paper, we p
g LLMs and actively gather information from the environment that is required to make the right decision. For instance, after dete
atings and decision-making depending on the employed strategy.
t's personality. Moreover, they account for different contextual, environmental, and spatial characteristics. We present the resul
nd foresighted behaviors for robot navigation. In this paper, a novel deep graph learning architecture based on attention mecha
lts affirm the viability of the proposed approach, allowing to perform vision-based conversations effectively, despite processing
response latency interacting with the robot narrator while articulating shorter utterances compared to the human condition whe
tivity with an interdisciplinary team, resulting in the extraction of the design, robot roles, and technical requirements. Based on t
man society, and exhibits robust proficiency in the comprehension and acquisition of natural language. Therefore, in this paper
s tunability and precise control. By manipulating the continuum robot structure, we achieve real-time reconfigurable responses
s variant, RRT*, have been employed. This paper introduces a unique RRT*-based motion control method tailored for continuum
e working range of object speed of the proposed gripper could be reprogrammed by controlling the sensitivity of the structures.
t the spherical acrylic beads into form strings instead of discrete beads.The beads could be pulled to the initial position by the th
on samplers, that capitalises the advantages of soft material actuators. At first glance, soft material actuators may not appear s
or to via powerful overlapping belt friction. The device enables passive, high-strength, and reversible fastening, and can easily re
ptic and proprioceptive signals. Moreover, previous works do not incorporate potential domain symmetry and thus must search
ool-center-point. The force estimation is probabilistic and relies on normal distributions considering model parameters and devia
ld testing has been extremely limited. In this paper we developed the first closed structure soft gripper for the harvest of blackb
ators and two adaptive electroadhesion (EA) pads, which enable it to flip forward, transition between two walls, turn in two direc
oints by rigid nodes. By extending or contracting the bellows, Tetraflex is capable of large size and shape change, and rolling,
straining high- power optimized actuators with densely spaced lateral supporting plates, we could significantly enhance structur
evelopment of robotic functionality. Here, we designed a flexible soft sensing Wire with the piezoresistive Eutectogel packed in
This solution effectively leverages the advantages of ECR collision buffering and adaptable maneuverability within confined spa
, this actuator is capable of replicating wrigglelike movements. Through experimental analysis, we have acquired the constitutiv
y and four identical multi-spoked wheel-legs with passive unidirectional forward rotation as limbs. The synergy of the body actua
r details the mechanical design, methodology, analysis, and experimental results of a compliant robotic gripper created for this
a ray. The robotic fish is driven by dielectric elastomer actuators (DEA), which controls the double wing structure on both sides b
f our approach on a set of examples, including two fabricated prototypes, illustrating applications for grasping and locomotion ta
ngular accuracy for long periods of time. Here, we present a nonmetal, light-controlled, multi-position inchworm actuator array t
corporate a high-resolution tactile sensing system inspired by the GelSight sensors. Concurrently, we design RObotic Modular E

ting above 60â—¦C makes the sponge soft and deformable to allow it to penetrate within surface asperities via a pressure norm
ution Strategies (ES). Additionally, the software can automatically produce 3D models of origami designs. We then used a dual-
nd, we utilize a redundant magnetic actuation system composed of five electromagnets and implement 4 degrees of freedom (4
ehavior is challenging due to the complexity of the constituent materials and the highly nonlinear nature of the problem. In this
city. We experimentally validate our modeling approach and investigate the effects of geometric patterns and input loading on th
tant curvature assumption. The soft robot consists of a soft bellow actuator and inextensible rigid skeletons, which exhibits a hig
ssure system. The thumb utilizes rigid joints and the palm features internal rigid structure and soft skin. The design can be direc
allenges. The glove enables adequate finger extension, thumb abduction, and extension for high MAS score patients. Modelling
 It can be fabricated through simple processes with low-cost materials. For a sample SFTTA with the size of a human finger, it c
action mechanisms that maintain these advantages tend to be relatively slow and rely on heavy driving fluids. This study introdu
y limited in their ability to simultaneously steer into tight curvatures and apply substantial forces to the environment. Here, we p
by pressure. Design, experimental characterization, and closed-loop control of the module are reported. In addition, a multifunc
ffness. By embedding the ability to independently control the stiffness of distinct sections of the structure, the manipulator can r
s, high sensitivity, a broad sensing range, and scalability. Then, for the challenging task of contact point localization, we propos
rom a single build material, simplifying the fabrication process. The flexibility of this approach offers significant control over sen
s and to take into account the presence of the actuator pneumatic supply. It is shown a CNN model can interestingly be used in
ated rigid-soft hybrid robotic glove, which combines the advantages of rigid robotic gloves and soft robotic gloves, and has high
pneumatic actuator (SPA), the SoftER. This novel actuator has the ability to rapidly unwind when pressurized to deliver impact
structure and enables modulation of stiffness. An analytical model is presented to investigate the variations in stiffness resulting
ace upon actuation. We demonstrate the individual OSM signal response to forces and folding. We further characterize the OSM
bending angle of 102â—¦, having high accuracy tracking capabilities, resulting in an average tracking error of 3.08 mm, that is 7
scle deformation information from the human arm for continuous control of the prosthetic grip force. The optical waveguide sen
es in resistance at multiple segments throughout the length of the sensor, providing improved resolution of local deformations i
data of different motion primitives (tapping, rotation and translation), an LSTM-based model can be used to replicate motion fro
EAs by taking both nonlinear electromechanical coupling, mechanical vibration and rate-dependent viscoelasticity into conside
s applied to these systems using current technologies. Adding additional sensors is challenging without compromising the softn
e structure and materials, realizing a compact form factor of the entire structure. The stiffness of the actuator can be increased
nt a vine-inspired robotic device with material-level responses embedded in its skin and capable of “growing” and steerin
atic soft actuator or robot that achieves the desired deformation behavior. Instead of constraining the robot's geometry within a
rlying reactions in order to achieve more complex behaviors for soft robots. In this paper, we present a thermally-activated reac
ssure and contact area of press events is shown. The sensor relies on spatio-temporal resistance changes driven by a pulsating
ly selectable number of COTS (commercial off-the-shelf) electrical drives and sensors. On the software side, this is matched wi
obotic applications in order to accelerate development cycles, loosen functional dependence, and improve resiliency and elastic
ows Plug'n Play behavior, which facilitates incremental and modular system design. Extensive experiments on real and simulat
de tasks such as machine tending and welding in synthetic environments and 3D scans of real-world machine shops. All bench
of obstacle avoidance. In this paper, we propose GSL-Bench, a benchmarking tool that can evaluate the performance of existin
h to model the temporal nature of cooking recipes. Leveraging a pretrained Large Language Model (LLM), we present Cook2LT
can be harnessed and applied to real-world challenges in the manufacturing industry. We present a novel system that uses Ch
ecution time depending on sensor data (e.g., pick and place with inspection, assembly) as it is able to reduce the total executio
s by LiDAR sensors, and then force the robot to take longer paths to avoid them, or abort navigation altogether. To overcome t
w a novel single-image Super-Resolution (SR) application - ANUBIS, Adversarial Network for Uncertainty Based Image Super-
pproach leverages PPO’s reinforcement learning power and MPC’s precision to navigate the complex control dynamic
s this problem as a hierarchically decentralized implementation of robotic trajectory planning and control and demonstrates the
nscented Kalman filter where the NN is used in the loop as a measurement module. Specifically, the filter tracks the target's rela
hnique to enable the gripper to proficiently approach a free-floating moving object, ensuring optimal pre-grasp success. For effe
Adaptation techniques to mitigate this issue. In the search for viable solutions, event sensing has been explored in the past an
scenarios with ultra short baseline and highly oblique viewpoints close to the ground. Our methodology surpasses traditional win
dmark locations are then combined with IMU data and a dynamics model in a batch nonlinear least-squares state estimator to
surable states are eliminated, while in Exp.2 the unmeasurable sloshing states are eliminated, and a novel estimator is used fo
 ots. Firstly, a novel standard interface is introduced, which features multiple mating modes and a lightweight design. Additionall

rings benefits of a passive, compliant contact surface that can affect its performance. However, the added softness of its shell a
erface (EMI) from the TransTerrA project, which has been proven in several field tests and reached TRL 4. Docking under load
t in microgravity, physical contact with the surface induces reaction forces, causing the robot to quickly bounce away from the s
all, so it is hard to climb obstacles such as steps and doorsill. Therefore, we propose a new solution: the variable wheel diamete
controller attractive due to its small computational effort is the Linear Parameter Varying (LPV) gain-scheduled controller. Howe

robotic manipulation with 3D surface feedback to solve these issues. First, we produce a deformation model of the manipulated
formance is further optimized by introducing robust acceleration control, with disturbance observers. Experiments performed on
rform an end-to-end assessment of the proposed architecture using the JHU-ISI Gesture and Skill Assessment Working Set (J
convenient boundary points for partitioning complex tasks into simpler subunits. Our method has successfully learnt a policy fo
We train agents in simulation with a reward formulation that encourages efficient and safe learning, domain randomization that a
of large amounts of annotated data. In household and industrial settings, synthetic data, generated with 3D computer graphics
reduce complication rates. Deep reinforcement learning methodologies offer scalable solutions for surgical automation, but the
r, we propose a new learning-based framework by leveraging the strong reasoning capability of the GPT-based architecture to
e segment can be lengthened or shortened. The advantages of the extensible segment are discussed in two main aspects throu
ansform and joint angle measurement errors for surgical robotic tools using an image based insertion-shaft detection algorithm
ecifically, we first implement an inter-frame matching mechanism to extract a coarse region of interest based on optical flow fro
ation with simple initialization from cameras may not fully capture. To use the simulation techniques in real surgical tasks, the r
stic 3D occupancy map. Using the 3D occupancy map, we plan sensing paths on the surface of the anatomy via a graph searc

cclusions and lens contamination. We leverage an aggregated dataset of six dry-lab surgical tasks from a total of 28 subjects to
od (FEM) that mimics the tearing behavior of soft tissue during CCC. Then, we train and evaluate RL models in simulation, dem
on subtasks in surgical training. ORBIT-Surgical leverages GPU parallelization to train reinforcement learning and imitation learn
ve Perspective-n-Point (PnP) problem and propose using the Levenberg-Marquardt solver on the Lie manifold. Then, the recen
mation make physical modeling of CTRs challenging, complicating in turn kinematics and control. Deep reinforcement learning (R
heir variable-curvature 3D shape when actuated, including the effects of torsion and external loads. To accomplish this, we em
and geometric modeling of the fiber using optical frequency domain reflectometry (OFDR), without requiring prior knowledge of
s using MSRCs, this letter introduces a new steerable robotic catheter system that integrates magnetic location sensing and ma
from the manipulator body and tip to a haptic interface. Second, we demonstrate how hybrid forward/inverse mechanics togeth
ing the shortcomings of existing puncture robots. To deal with the challenge of cross-modal and cross-dimensional registration
in terms of inter-agent distances and the distance to the target location. In DDC, each agent continuously learns and updates a
implemented with a set of lights turned on and off with a time-varying pattern. Additionally, we introduce also a minimalistic fully
gging activities. However, ARK is limited in mostly enabling experimentation with a single swarm. In this paper, we introduce M-
rovide physical support and strong power, meanwhile, these needle-like motors also enhance the physical interactions among r
source, and to sense contacting neighbors using vibrations. We validate this robot through full-robot tests demonstrating photo
use a general form of the swarming oscillator, swarmalator, model to study reciprocal and non-reciprocal interactions among a
bots that populate their environment. We frame our research within robot shepherding: the problem of using a small group of ro
ulator interface that we develop runs an instance of the Crazyflie flight stack firmware for each individual drone on a commercia
that explores efficiently with the help of weighted frontiers. Our hierarchical approach uses a combination of motion planning a
Streams for Task and Motion Planning (COAST), a probabilistically-complete, sampling-based TAMP algorithm that combines s
ension that is based around a cooperative pointpair. This extension keeps all the benefits of the original formulation that is base
earch and global optimization for efficient hybrid planning. Through empirical evaluation on three benchmarks, we demonstrate
o avoid missing trolleys during exploration; room layouts for semantic space segmentation; and polygonal obstacles containing
n trees as a problem formalization, where before we can expand or evaluate a node, we have to solve a preemptible computatio
e required for a single measurement but also hinders sampling during robot motion, thus exacerbating the scarcity of available
chability graph to enable solving optimal TAMP on high-DoF mobile manipulators. The proposed reachability graph can incorpo
- driven forward subproblem search with an optimization based Task-and-Motion Planning solver. To guide the search, we intro
executable formalism. The framework is supported by a set of open source tools and its functionality is shown with a realistic a
information across the modalities. To this end, we propose History-Aware Planning based on Fused Information(HAPFI), effec
carry out a garbage collection task in an open-air environment by operating a mobile robot with manipulator arm, and we demon
to sequential heuristics and implicit time and roadmap representations, and (2) maximized feasible solution space obtained by
a general planning framework for TPO specifications, a MILP formulation accommodating time windows and precedent constrai

that operates on both the geometric space and the symbolic space. The RRT is guided by the task plan and it is enhanced with
rates concentrations. In this paper, we propose an online, distributed multi-robot sampling algorithm based on Monte Carlo Tree
planning while satisfying task constraints, our algorithm achieves the best balance between the task success rate and energy us
from materials that are readily obtainable. The part design and detailed instructions are publicly available. In order to broaden c
cannot be completely ruled out in real pHRI applications. For example, to allow edged or pointed tools such as screwdrivers ne
egies by introducing a novel formalism of planning in a strategy-augmented domain. We also introduce and formulate the notion
loiting object-level knowledge as a FOON for task planning and execution. Our approach automatically transforms FOON into P
uncertainty and unreliable features of LLM outputs using task tree retrieval. We combine multiple LLM task tree outputs into a
hieves the goals of multi-agent task planning by solving an initial problem using a minimal number of agents. Subsequently, tas
the existing methods implement frontier point detection based on the Rapidly-Exploring Random Tree (RRT) and use greedy a
mote robot equivalent to a "Marionette" control, but whose feedback was limited to only visual feedback on the human side. In t
s the Hidden Semi-Markov Model (HSMM) and its task-parameterized version (TP-HSMM) to their confidence-weighted version
ments of the operator's arms to obtain the desired joint references, sent to the follower dual arm (FDA) to reproduce in a natural
DoF to be controlled by the operator and introduces extra mental demand. On the other hand, the communication perspective
tic and tactile rendering, could further amplify the intuitiveness and efficiency of MR-based immersive teleoperation. In this stud
ployed with or without previous knowledge of the environment. An online estimation of the OBM is proposed using a JND and a
al feedback quality for the attached eye-to-hand camera with motion constraints in consideration. The effectiveness is demons
pulation (i.e. pivoting, sweeping). Besides, haptic interface, which is used to render interaction results efficiently, still limits perfo
s -- from demonstrated end-effector trajectories. We use a probabilistic model, Kernelized Movement Primitives, which additiona
ELESIM's modular design, it is possible to control the digital twin using any device that outputs a 3D pose, such as a virtual rea
otion model on the HRP-4 humanoid robot in forward, lateral and backward walks with concurrent upper limbs motions experime
real-to-sim and the semantic reasoning capability of large language models. In other words, given place context information (e
. Diminished reality techniques are frequently employed in these cases to mitigate the robot intrusion into the environment, alth
k performance with iRoCo and gold-standard control systems in teleoperation tasks. Additionally, iRoCo users complete drone p
trol. This framework preserves the fluency of our avatar interface by minimizing obstructions to the operator's view and using th
cal information in neural networks. Technically, we employ hierarchical dependency loss to boost overall accuracy. Furthermore
a dynamic locomotion mapping, utilizing human-robot reduced order models, and a kinematic retargeting strategy for manipula
ntelligence (AI) to predict just-in-time the user's intended motion and then, if the user accepts, completes it autonomously in 3D.
orse for point clouds. To overcome this dilemma, we revisit the official Point-MAE implementation and find the critical role of ag
wise 2D supervision signals. In this work, we propose Multi-modal Prior Aided (MoPA) domain adaptation to improve the perfor
r practical implications, i.e. transfer the policy from robot A to robot B. Unlike prior methods that rely on paired data, we propose
instead of adopting the full fine-tuning paradigm to substantially improve the parameter efficiency. Finally, a lightweight PointAd
etection. Since BEV perception approaches are complicated and contain several components, the domain shift accumulation on
pre-trained NeRF. Though trained exclusively in a simulated environment with simplified objects and 4-DoF top-down grasps, o
as the first attempt to address this difficulty. Motivated by the idea of multi-grid solver, the control policy on a virtual agent over
Nonetheless, the ”reality gap” frequently results in performance degradation when transferring policies trained in simulato
including individuals with disabilities, able-bodied individuals, children, and adults, with the assistance of AI robots. The core co
ned to prevent external obstacles from entering the sole of the foot, while also being designed for shape adaptation to uneven g
ovement. Based on this module, a modular and reconfigurable supernumerary robotic limbs robot has been developed, which
ance suction units, each of which integrates a three-degree-of-freedom active compliance actuator and a flexible suction chamb
an reduce the number of dimensions of motion planning for the plate entering the groove from six to three and adopt the A-star
enderizing meat with a mallet, are made possible due to the backdrivability provided by the actuators' low gear ratios. The syst
not generalize well to a new environment. Moreover, they do NOT allow a person to be away from the Line-of-sight between de
r and combat environmental degradation. Once deployed, they connect to our Collaborative Robot Interface for Telepresence (
nvolvement of multiple individuals. Similarly, handling and transporting heavy objects for machine assembly, and part fixation po
se estimation, and most rely on slow online optimization for stability. Unlike conventional 2D or 3D keypoint estimation, 3D hum
auditory inputs to improve active collaboration with human users. To this end, the cobot was equipped with object classification
meras) and learned from scratch in only 5h on a single GPU -- as well as grasp unknown objects based on a single depth imag
k transparency, and it is difficult to revise the existing knowledge in response to previously unseen changes. In a departure from
ctive localization methods via on-board sensors such as Visual-SLAM and LiDAR- SLAM is also limited by the sensor property.
he camera mounted on the front of the robot. The robot is equipped with four thrusters, which allow it to move freely underwate
oving object for marine animal surveys, employing the tilt control of MBS on AUVs. YOLOv3, the object detection algorithm, de
nism using a string of beads that can be composed of cut-resistant and incombustible metals, arrange the mechanism to form
system that uses a social robot to perform music intervention sessions for groups of dementia patients. This system, named the
pen chest surgery. However, these life-saving procedures are accessible in a few clinics with skilled and experienced operators
create a service robot application, in which the robot autonomously navigates to a table, picks up a bottle and delivers it to the
arameters. Our contribution is a systematic method that extracts dynamic features from human demonstrations to auto-tune the
g methodologies. The first method capitalizes on foot-ground friction, employing repeated steps with one foot while using the ot
efore, these robots must possess the autonomy to make economic decisions aligned with the interests of their respective paren
ndent of human input. In our novel automated grasping controller, the slipping motion of held objects is detected by an inexpens
es a pointcloud, which is formulated with a Trimble X7 scanner of the physical room. The Unity environment works as a server a
mbers tend to be heavy, slow, complex, and expensive due to their ground-based designs, with excessive actuators and structu
developing sensitive hands, a multi-modal robot head, and a Lightweight Exoskeleton to fulfil the aforementioned goals. While t
ned deep-sea submersible, Deep Sea Warrior, serves as the operating platform for the nitinol-embedded soft robotic gripper. In
integrity), IMUs, a microcontroller, and a Bluetooth module. This robot achieves speeds exceeding 10 body lengths per minute
wo epoch-making devices are introduced based on the above principle. Grow-hose-I, shown in the first half, is the world's first d
and thus altering buoyancy. The regulator is approximately 5cm in height and 3cm in diameter. A single regulator weighs only
 ental test (see the video), which indicates that the quick switch mechanism makes the idea of the Series-Parallel-Reconfigurab
and the physical load experienced by its user. If the forces generated are inappropriate, its user's natural movement trajectory
c arm system. Taking the typical multi-hand task in aircraft manufacturing as an example, We built a task model and interaction
mation as well as the embedded nodes of the produced plans to define similarity measures. These are then utilized to select th
g show the new methods significantly improves tracking duration and robustness, both for perspective and fisheye cameras. Ou
available from many sources, including structured lighting, homography decomposition of infinitesimal planes and shape priors
al solution: both simple and cost-effective. However, the challenge in this process is in tracking the robot under larger environme
ion-based visual feature module to fuse the hierarchical visual representations. The proposed method achieves state-of-the-art
ctive when most targets are static objects with the same appearance, as in the agricultural case. To this end, on the trail of SOR
o be calibrated in advance or the UWB anchor position estimation method used is not robust enough; 2) these approaches do n
o the accumulation of translation errors. To address this, we first independently estimate the gyroscope bias and use it to form
ation of the Euclidean Signed Distance Field (ESDF). We introduce a library, nvblox, that aims to fill this gap, by GPU-accelerati
manner. We configure the exploitation module to take over in the deterministic final steps of navigation i.e. when the goal becom
ding-mode control that effectively handles uncertainties due to sensor noises, parametric uncertainties, and wheel-terrain intera
nd deploy it in both simulated and real-world environments for zero-shot object navigation. We show that our proposed method
ed-loop control of mobile furniture. We propose a perception pipeline that addresses these challenges. A machine learning mod
vate the performance our model. Addressing uncertainty within the system, we train an ensemble of predictive models and estim
two challenges: (1) novel and realistic visual observations; (2) error and misalignment on coarse maps. To tackle novel visual
s in pixel brightness but faces limitations in recognizing stationary 2D markers. Although compact blinker markers using blinking
y estimate high-quality optical flow in embedded devices. Our proposed RAPIDFlow model combines efficient NeXt1D convoluti
stopping at the wrong place. In contrast, humans intuitively frequently use identifiable objects or store names as reference land
and generic computer vision components. We utilize visual place recognition for the subgoal selection of the topological navigat
urate and coherent scene descriptions and address this misalignment issue, we propose the Aligning Knowledge Graph with Vi
of both end-to-end reinforcement learning methods, which suffer from memory decay over long-horizon tasks, and traditional m
igation skill (PixNav), which bridges the gap between the foundation models and the embodied navigation task. It is straightforw
ation head to generate pseudo-labels for re-training on unseen environments in a self-supervised manner. GeoAdapt uses geo
ccurrences, such as stairs. To overcome these issues, we introduce ViPlanner, a learned local path planning approach that gen
n localization. UIVNav consists of the following phases: (1) generating an intermediate representation (IR) and (2) training the n
reward functions that may not align with operator preferences. In this work, we posit that operator preferences for visually nov
oy a Transformer backbone to explore the global clues of the image. To compensate for the inadequate local detail optimization
ment. The best-prototype pairs (BPPs) are searched for in support and query images based on appearance similarity, to yield i
ccupancy models only using 2D labels. Specifically, we extract a NeRF-style 3D volume representation from multi-view images,
mizer alternately updates the depth and camera poses through iterations to minimize a feature-metric cost, and two gated recurr
pixel in the other image. Doduo uses flow-based warping to acquire supervisory signals for the training. Incorporating semantic p
abilities of MLLMs enables the creation of a simplified framework, thus mitigating the need for task-specific encoders. Specifica
sentations. Intra-modal and cross-modal contrastive learning techniques are employed to facilitate effective comprehension of
ichannel output representation for stable training of depth prediction in driving scenarios, (2) an optical-flow-based mask design
n shift from real events, the absence of quantified validation, and layering problems within the time axis. In this paper, we prese
s are better suited to leverage the co-dependent nature of the tasks for improving the accuracy of both tasks. To this end, we p
mal-based depth estimation by leveraging pre-trained visual models initially designed for RGB data. In detail, we design a novel
gical Activity Recognition (OSSAR) framework. Our solution leverages the hyperspherical reciprocal point strategy to enhance t
object categories, and slow surface extraction during the training of implicit reconstruction models; thus hindering both the spee
chieve more robust keypoint detection and tracking. Specifically, we propose a novel keypoint detection network that fuses the
ntinuous space. Specifically, our framework leverages the ability of Neural Ordinary Differential Equations (Neural ODEs) to mo
at combine thermal and visible properties. This paper introduces an end-to-end fusion network that leverages generative adver
the original 3D information. In this work, we propose a learning-based 3D box adaptation approach that automatically adjusts m
e prediction transformers and adapters. We demonstrate the effectiveness of our proposed method by running evaluations on s
omes, adept at producing diverse, high-quality, and condition-aligned results. Nonetheless, the training of these models deman
entation that effectively utilizes the geometry prior from LiDAR in autonomous driving scenarios. Lightning NeRF significantly im
econstruct 3D scenes without additional inputs. Moreover, we present a novel density-guided patch-based sampling strategy fo
DAR provide precise depth estimation, the prohibitive cost of high-resolution LiDAR systems restricts their widespread adoption
ge of the scene. To address this problem, we present a simple yet effective solution called NeRF-Enhanced Outpainting (NEO)
due to complex deep network designs. We in this paper propose the DL-PoseNet to address these issues. Specifically, we pres
ntation. We study the scenario where the model is jointly optimized using the standard diffusion loss alongside an auxiliary obje
arning (K-VIL) to bimanual manipulation tasks. The proposed Bi-KVIL jointly extracts so-called Hybrid Master-Slave Relationshi
e scene representation. We employ the technique of spiral sampling and feature rendering to enhance rendering quality while s
sh view planning for active reconstruction; the former approach poses a challenge for complex or newly identified targets while
l fiducial marker called CopperTag, which incorporates circular and square elements to create a robust occlusion-resistant patte
than relying on external hardwares, this work proposes a novel approach: aligning by recognizing the inherent geometric patter
rops and then train a Neural Radiance Fields to recover the 3D scene implicitly. By leveraging the strong scene representation

his problem by marrying image matching with NeRF. With 2D matches and depth rendered by NeRF, we directly solve the pose
an occluded part detector to identify occluded body parts within a human skeleton. It is based on an autoencoder trained on a
mes and a novel distillation scheme between the teacher and student networks in the multi-frame self-supervised depth estimati
addition, they neglect overall image quality improvement. To partially address these challenges, we introduce an innovative app
sent a novel semi-supervised framework for visual BEV semantic segmentation to boost performance by exploiting unlabeled im
o-labeling system that can automatically generate 2D masks, 3D masks, and 3D bounding box annotations for vision and point
ge in requiring to retrain the whole model on variant datasets. Nonetheless, segmentation models developed on SAM remain co
pe space. We demonstrate improved performance in feature detection, camera pose estimation and reconstruction compared t
th estimation, leveraging its annotations and spatial information. The inherent differences in modalities and data distribution ma
ouds are encoded using a pillar grid structure to generate pillar features. Then, OPP determines the active pillars used for point
guarantee an optimal field of view (FoV) coverage of retro-reflective targets during the whole production horizon. Especially in a
s paper, we introduce the Multiple Morphology-Aware Network (MMA-Net), a novel framework that improves Cobb angle meas
ates. On the other hand, the provision and use of synthetic data introduce individual challenges, such as potential domain gaps
orrespondences. To address this problem, in this paper, we propose a novel deep learning based two-stage method for pose e
tor and demonstrated acoustic field modulation and microparticle manipulation by adjusting PZT position and orientation. The P
h images and above 30,000 point clouds of cluttered scenes with different number of the objects, which save time significantly f
or interest. In this paper, we introduce an instance-level object pose estimation method for bin picking, which is trained using on
ormation without requiring object tracking. This results in a highly efficient classification model that can easily be deployed on e
erate rotations often outperform classical approaches. In order to address this shortcoming, in this paper we propose RIDE, a le
bility of old data when updating the model, necessitating an exemplar-free continual learning (CL) setup. However, prior studies
hese include issues of uneven lighting, field-of-view loss, and motion blur. This paper proposed a perception method named 'M
 ts in unstructured environments using high-resolution RGB images instead of utilizing relatively noisy depth data. We conduct s
require complicated calibration. Moreover, if external sensors such as cameras are used, it is necessary to grasp foods without
onstration graphs by splitting existing demonstrations and recombining them. In order to traverse the demonstration graph in the
ond-order derivative of features and the robot's acceleration. The current state-of-the-art techniques mainly focus on image-bas
ew control sequences for the UAV. Real-world experiments validate the scheme, showcasing its robustness in handling comple
exity of the approach from the objects’ geometric complexity. Our approach is based on the As-Rigid-As-Possible (ARAP) d
arkness clue prompts. Without a separate enhancer, DCPT directly encodes anti-dark capabilities into prompts using a darknes
detector, ensuring continuous target visibility, even when faced with motion blur, abrupt light shifts, and occlusions. Compleme
allenging to detect and recognize; (2) drones move and see objects from different angles, causing the unreliability of the predict
r multi-object scenarios, we propose an ensemble of K interconnected OPFs, where each filter predicts plausible object tracks t
e models for open-vocabulary video tracking? In this paper, we combine an open-vocabulary detector, segmenter, and dense o
and carries out feature correlation at the end of each stage based on sparse pillars. More specifically, in each stage, self-atten
yield the opposite trajectory as tracking forwards). In this work, we take on the challenge of improving state-of-the-art supervise
ng capable of running online in real-time. The most probable track associations are identified by evaluating all possible track hy
gers could be put at risk due to erroneous detections that have propagated to downstream tasks, including MOT. While there ar
humanoid robots, targeted for industrial manufacturing, to manipulate and assemble large-scale objects while walking. It is asse
hieves a more favorable trade-off between accuracy and efficiency than the other lightweight trackers. The main innovations of
d accurately separate the quantitative windows between the crop rows. Our system constructs spatial constraints according to
hicle motion under specific control inputs on flat ground using ordinary differential equations. We use a singularity-free and differ
ptive selection between the original Lucas-Kanade (LK) method and the inverse compositional method when constructing photo
se the GNSS pseudorange, Doppler shift, single-differenced pseudorange, and double-differenced carrier phase measurements
hing dual-fisheye images with extrinsic and intrinsic parameters) and sparse depths (generated by the front-end local tracking) a
er, we present BIT-VIO, the first Visual Inertial Odometry (VIO) which utilises SCAMP-5. BIT-VIO is a loosely-coupled iterated E
her through well-designed feature management. To achieve reliable state estimation, we tightly couple the point-based and line-
ating JacobiGPU into ORB-SLAM3's g2o, we enhance the linearization process. Our evaluation, conducted on 12 sequences of
ce under fast rotational motion and extended integration time. Furthermore, we extend conventional front-end tracking and back
ed a lightweight multitask network capable of simultaneously detecting track lines and obstacles. Additionally, due to the limitat
ntly work for the STS assistance. To implement the strategy, this paper utilized the High-Strength & Flexible Mechanism (HSFM
meters of the tail are examined. Through simulations it was shown that by incorporating the delay dynamics, the requirements o
here the device absorbs and can store energy in the form of compressed air, leading to energy savings in active mode. The abs

oposes a new motion planning scheme for upper-limb exoskeleton robots, which drives the robot to provide customized, safe, a
s constrains normal motion. This paper presents a novel solution: a pneumatic back exoskeleton made of structured fabrics tha
s study proposes an alternative method by incorporating FSRs on three inclined planes within the insole. A vector scaling proc
hod that could utilize the human-exoskeleton locomotion interaction experience (HELIE) to assist naive wearers during the exp
h and shin harnesses of the exoskeleton to replicate the natural trajectories of the knee. However, achieving this is a complex
controller exhibits adaptability, making it suitable for various subjects without the need for prior parameter identification. To dem
a limited amount of assistance torque. Funnel-based feedback control is used to mitigate the typical time lag seen when using L

ltiple constraints. In addition, both position and impedance control strategies are investigated. The proposed strategy is experim
of velocity and ground contact state probabilities using encoder data, these probabilities are employed to modify particle weight
 ons of motion for model-based control algorithms - specifically Operational Space Control. The formulation was used to simula
ases, a tractable nonlinear programming problem is obtained. The model explicitly accounts for contact switches and impacts, tr
tor to control the robot by mapping oscillator states to wheel-leg trajectories and implementing differential steering within the os
MR dynamics with the Pfaffian constraints of the no-rollover and no-slip conditions. Similar to the case of Navier’s table prob
t for planar motions, which results in an over constraint of the vehicle. In this work, a two-wheeled robot with an addition structu
sing current methods, our control law satisfactorily addresses the cases where the vehicle's velocity passes through zero. In ge
r the distance to the landing surface nor the flight time. Based on an estimate of the Center of Mass horizontal velocity, the met
on through a hierarchical optimization-based control framework. First, an online manipulation planner computes the manipulatio
obility, most seeking to affect the world would significantly benefit from planners/policies that facilitate self-righting whenever po
standing-up motion of humanoids with variable body structures and conduct evaluation experiments. We have demonstrated th
ke agile whole-body recovery motion. Then, during the rising phase, it reduces these forces to zero and stabilizes the robot in a
oint space, allowing robust distance queries in arbitrary joint configurations. The proposed RDF representation is differentiable a

olicy for legged manipulator control. Force control enables us to realize compliant gripper and whole-body pulling movements t
ontrol methods for wheeled humanoid telelocomotion on a whole-body level has yet to be explored. In this work, we explore ho
 nario. In this regard, we present the community a text spotting validation benchmark called Under-Water Text (UWT) for noisy
ple consistency to encode both local structure and global semantic information for learning invariant point cloud representation
s or partial or total occlusions. We used both a laboratory environment and an HRI scenario where the robot followed a person t
es not require registration. We demonstrate how the optimal inter-cluster score of the correspondence compatibility graph of two
e, our research reveals that all combinations of paired data are unnecessary to train such unified embeddings, and only image-
Ds (PS-VLADs) to aggregate descriptors. A BPU includes a parameter-shared convolution module (SharedConv) that significan
y of response, supplemented by a novel background-aware module to enhance the ability to learn from negative samples. In ad
g 20,000 different objects to reach such generalization capabilities. To overcome this shortcoming we introduce ZS6D, for zero
ings as the transformer layer iterates. The letter proposes a robust model for action recognition that overcomes the limitations o
nism and works between three discrete levels of transmission ratio depending on the user’s initiative. The proposed prototy
nsists of rigid actuation on joint 1 and PEAs on joints 2 and 3. With this configuration, root mean square (RMS) and peak load t
egration of the lever technology into VSA often faces challenges in achieving a compact design. This paper introduces a mech
the lever length, the range of stiffness adjustment can change from 0.3Nm/degree to therotical infinity. Moreover, the continuou
age and release timing. In this work, we introduce an initial proof-of-concept of Bi-Stiffness-Actuation in the form of a 1-DoF phy
sition control experiments, and a 10-hour endurance test. The results show strong performance characteristics, and crucially, s
deling of the pneumatic actuator is presented in detail. Based on the actuator modeling, the multi-objective optimization for para
s characterization. Testing of the dual-mode actuation is finally described. The complementarity between the two motion modes
eflection behavior -- for example, when they cannot meet the force demand despite storing enough energy to perform a cyclic ta
ees (RRT*) with novel sampling and steering techniques to speed-up the computation. This algorithm is able to obtain collision-
omplish this goal, a relay UAV is positioned to enable indirect but constant LOS connectivity between the mission UAV and the
et to learn energy consumption for the UAV and heteroscedastic uncertainty associated with LiDAR measurements, both as fun
d to the dimensions of the flight area and the coverage of the positioning system. Therefore, this papers introduces a new way t
volutional neural network (CNN). Leveraging a real-world vision-based regression task, we thoroughly explore performance-cos
tness property of the global tailsitter flying wing dynamics, which is derived in this work. By performing snap minimization in the
ence in optimal conditions. Several practical considerations, including construction of a training dataset, data noise handling, an

no-drone perception algorithms. This work presents a lightweight obstacle avoidance system based on a novel millimeter form
sketching systems. We have implemented a robotic sketching system featuring an UltraArm robot equipped with a RealSense
es to satisfy joint movement limitations. We use three different agents to validate the planning algorithm with our handmade rob
Safe Active Exploration paradigm based on Gaussian Processes to explore and characterize the audio onset response of feasib
d has sparked a need for robotics courses in higher education. Incorporating real robot platforms into such courses and curricul
of real world flight data, our geometry-aware model outperforms state-of-the-art baseline models trained with more than 15 min
easured or controlled (e.g., nominal model mismatch, air plate effects, and unpredictable wind), which we call ``latent''. Imprecis
ch camera in capturing detailed images of the targeted objects. This integrated approach enhances the overall efficiency and ef
st, we propose a framework that accurately detects and segments open-world object categories in a zero-shot manner, which o
work that generates PDs using state-of-the-art LLM and vision-language models. ViLaIn can refine generated PDs via error me
markably, we show that many of the intermediate results used to obtain the gradient, especially the Kalman filter, can be efficie
utlier features to reconstruct the texture normality, offering a rational segmentation map through decomposition and reconstruc
el task planning, sidelining low-level robotic control. In this work, building on the idea that the kinematic structure of the object d
l framework that improves LLM-based planning through an iterative self-refinement process. The framework operates through t
 on of a convolutional spiking neural network for feature extraction and a Bayesian spiking neural network for odor detection and
nts and predict their dynamics. Our model infers dynamic interaction graphs in a latent space augmented with interpretable edg
r enhance scene flow estimation performance, we formulate a novel loss function that accounts for the data imbalance between
joints vital for motion control. Through extensive experimentation with sophisticated agents in diverse environments, we highlig
troduce the PO-ECAS framework and a MILP-based solution method for approximating an optimal ethically compliant policy. N
he guidance of LLMs. In particular, we utilize the LLM as a strategic planner for applying previously learned skills to novel tasks
roblems over complex terrains. We propose algorithms for learning expert reward functions, and we subsequently analyze the
heeded. Our method is specifically designed for efficient detection even with high dimensional data, and it empirically achieves
aptively accept or discard a candidate action sequence with respect to a probabilistic belief-dependent constraint, before expan
ow that such policies achieve the formal convergence guarantees in finite time. In addition, we thoroughly evaluate their advant
cially vital in high-dimensional scenarios. Theoretically, our proposed filter is shown to offer a near-accurate representation of t
y is insufficient. To address these limitations, we propose an innovative interdisciplinary framework that incorporates human-com
anced telepresence robot was assessed in an unstructured domestic environment by 10 users. The results show an accuracy o
ain a robot to compliantly manipulate a knife, by reducing the contact forces exerted by the food items and by the cutting board.
ervation, imagination, and execution—to adeptly address the task. Initially, objects are discerned and extracted from a cluttere
ainst the limits of context window length. It is therefore critical to find token-efficient LLM planning frameworks that are also able
Multi-modal Large Language Model (MLLM) to weave knowledge of object locations into natural language instruction, thus aidin
ility to industrial robotics and automation setups. To transform our physical Foosball table into a research-friendly system, we a
components: 1) an instruction discriminator to determine which system should be activated based on the current user's instruct
API documentation and detailed descriptions can improve code synthesis capabilities. However, it is not clear how to document
btask policies using a multifidelity simulation pipeline, the framework not only allows for efficient RL training, but also for a refine
n such partially unknown environments still remains an open question. In this work, we first conduct an extensive simulation ex
method that leverages pretrained AI foundation models to extract rich semantic information about specific types of objects (e.g.,
rmines how to reorient it properly to be used. Experiment results show that our method achieves a high success rate on the rea
ositions, and reward functions using large pre-trained generative models of language and vision. We generate 3D assets for si
ework named FLTRNN. FLTRNN employs a language-based RNN structure to integrate task decomposition and memory mana
s driving systems. We introduce a method to extract nuanced spatial features from transformers and the incorporation of latent
ed, inference over planning tasks must be performed jointly with motion plans using traditional task-and-motion planning (TAMP
s or precise object locations, easing real-world deployment. We train ASC in simulated indoor environments, and deploy it zero-
onsolidation in the robot’s LTM only when new information is encountered. The proposed approach combines fast filtering d
essly integrates the physical and virtual worlds, providing a bidirectional egocentric and exocentric environment representations
altitude and speed. Our framework features a bi-directional interface that allows the concrete algorithms used for enforcement to
erarchical control structure, the dimension of the optimization problem is greatly reduced with synthesized wrench commands.
en question, showing that solving for trajectories by minimizing Kalman covariance computed in a noiseless environment is rea
unded curves and reduces the number of gains to tune. The proposed improvements are verified in simulations using the Papar
h can be used without location or velocity sensors. During the landing phase, the tethered lifting-wing multicopter only needs to
light. With this paper, we address this gap by presenting AirFisheye dataset, which is applicable for tasks such as segmentation
itionally, a novel swarm control algorithm is proposed to make it compatible with the new relative localization method. We provi
map Exploration (HIRE) planner for UAVs exploring dynamic environments. The proposed planner adopts an incremental samp

ms as they operate over a state-space discretization. This paper presents Morse Graph-aided discovery of Regions of Attraction
arking against the state-of-the-art solver OSQP, achieving nearly an order of magnitude speed increase, as well as through hard
n addition, a capacitive sensor was used to detect oocytes in the manipulation processes. The results showed that oocytes wer
without the IE condition. In the TDMC mechanism, a filtered regressor is integrated with multiple time intervals to generate a g
t the safety-critical events it sees are virtual. Sim-on-Wheels is safe, interactive, realistic, and easy to use. The experiments dem
h as a variational autoencoder, to learn the likelihood of the generated parameters from the training datasets; It penalizes the ge
se an effective data augmentation approach, which allows the training of a network capable of handling different views of the ta
ation-based and speed-based re-weighting techniques, we achieve improved overall performance, as measured by most-likely F
ances of human preferences on realism and unifying diverse traffic simulation models. To tackle these issues, we propose usin
y, ensuring compliance. Moreover, verifiable safe operation is essential for deployments that satisfy worldwide standards such a
d the large transmission delay. In this paper, we introduce a novel approach to cooperative perception. By fusing both the road
een pre-trained with complete observations, to the student model, which has only few observations. POP achieves comparable
, which captures potential future interactions in an end-to-end manner. FIMP adopts a future decoder that implicitly extracts the
. In addition, the representation is rooted in the Ego vehicle's coordinate system with a logical representation of roads. This des
amework follows a query-response-update mechanism through an online Behavior Tree framework, where agents broadcast qu
assignment phase, each agent builds a squared cost matrix by cloning agents adding pseudotasks when necessary, and applyi
using a synthetic dataset. The proposed controller is shown to be more efficient than a model designed for a single prediction h
navigation results on RoboTHOR with a success rate (SR) improvement of over 27% over the current baseline of the OWL-ViT
and embedded by different autoencoder variants. The robot can select the most promising next exploration goal through the di
ese steps into initial BTs. Subsequently, a BTs Update algorithm is proposed to expand the initial BTs dynamically to control rob
of hypergraph over time but also highlights the vertices and hyperedges with larger impacts. We apply DyHGDAT to a CVAE-b
ector modalities with static LLM representations using vector captioning language data. We also introduce an evaluation metric
ions, we propose a Heterogenous Hypergraph Graph Neural Network (HHGNN) based motion prediction model that leverages
dentify further false positive raw targets in the current measure which are not accessible from the previous state. Moreover, we
erception and path planning. Here, we introduce a novel end-to-end navigation system, the "Thermal Voyager", which leverage
ompounding errors. Furthermore, we highlight an imitation gap that has been overlooked by current learning systems. Finally, i
dopts the crowdsourcing way to collect semantic information about the environment and traffic flows. The topological structure i
ch for predicting both observed agent trajectories and inferring occlusions in a partially observable setting. It uses a transformer
employs stable semantic features instead of low-level texture features. First, semantic maps are constructed offline by detecting
resents the full distribution. Other methods optimise closest-mode evaluations, which test whether one of the predictions is sim
semantic-level ego policy. Then, the generated policy-conditioned scenarios are further formulated into a tree-structured represe
functions of the all demonstrated lane-changing trajectories, and the cost function database is constructed. Then, the scenario
collect datasets of our wheeled robots crawling over previously non-traversable, vertically challenging terrain to facilitate data-
common reaction has been to entirely discard classical navigation systems and start from scratch, building a completely new le
he detectability measure, the robotic motion planning framework based on Model Predictive Control (MPC) is proposed to achi
cision-making based on visual cues. We extensively evaluate Talk2BEV on a large number of scene understanding tasks that r
ciates the latent feature representations derived from different modalities through a learnable geometrically linked positional em
OOD situations without human intervention. Specifically, we develop an image-goal conditioned offline reinforcement learning m
onditions the placement and motion trajectory of every new agent on all existing agents and their trajectories, leading to realistic
m a LiDAR. To solve the above problem, we propose a new deep learning network named as implicit point function (IPF). The b
of the surrounding environment, which enables a semi-supervised learning for the CADNet, and thus desirably avoids annotatin
nsory information. The resulting model captures the inherent uncertainty for future human trajectory prediction and achieves sta
we first learn the N number of human digital drivers from the multi-human demonstration dataset, wherein each driver possesse
h discrete options and continuous variations of the complex driving skills. The final policy treats learned skills as actions and can
ed UniTraT, through self-supervised learning using ground robot travel simulation. Furthermore, it employs Uni-directional to Mu
hieving complete environmental coverage with LiDAR remains prohibitively expensive, particularly in extensive areas. Hence, t
nts. To tackle these complexities, we introduce an integrated framework that merges gas source localization with robot navigati
relation and topological layout into account. We adopt a modular system design with this map representation and graph policy,
enBot-Fleet can learn robust navigation policies that generalize to unseen homes with >80% success rate. OpenBot-Fleet repre
hese modular approaches, design new paradigms that mitigate these limitations, and accelerate the development of end-to-end
ntervals of the prediction values and thus supports further processing in safety-relevant applications. The effectiveness of the m
trained visual backbone for policy learning and then propose SpawnNet, a novel two-stream architecture that learns to fuse pre
d (b) action representations that can extract performant policies with small yet diverse multi-modal datasets without overfitting. I
out needing to collect a training dataset of example arrangements. Results on a series of real-world tasks show that this framew
E, a reset-free fine-tuning system that pre-trains a multi-task manipulation policy from diverse datasets of prior experiences and
lop a generalized framework to distill motion forecasting model ensembles into small student models which retain high perform
ot), resulting in false positive bug reports that can impede robotics developer trust and productivity. In this work, we study how d
periments under different time-averaged thrust-to-weight ratios (0.745-0.876) and launch angles (45°-90°). The results of flig
o characterize the influence of structural parameters on the motion performance. Experimental swimming results show that com
relation between the deflection of the vertical tail and the roll angle. We prove that, even though this deflection causes a yaw m
By changing four parameters extracted from the model, the kinematic model can demonstrate all the aforementioned swimming
er water content. Measurements of mud reaction force revealed two distinct failure mechanisms. At high water content, the red
e main bottleneck has been the difficulty in collecting detailed sensory information, corresponding locomotory responses and
drift of the IMU sensors through a fused Kalman Filter. Furthermore, we propose the Anti-Shake Estimation (ASE) algorithm to
m and determine the optimal tunable stiffness interval. The results show that tunable stiffness can significantly improve the resp
mics of neural activity and the combination of attractive and repulsive forces. In addition, a neurodynamics-based self-adaptive m
cillator couplings between oscillators, and can such coupling improve sim-to-real transfer for navigation robustness? 2) What a
ronment using lateral undulation.
to the robot by modifying some of its chassis structural plates to extend to and support the shell. We evaluated the effectivene
dly optimize online. We design foot force profiles parameterized by only a few parameters which we optimize for directly on har
for generating the stabilizing torque. The resulting sail-like tail is substantially lighter than the original, but still improves the perf
ient locomotion. Preliminary results from experiments demonstrate that STRAY can achieve a trotting gait of approximately 0.6
adaptive dynamics model, providing a robust theoretical foundation for subsequent motion planning and high-precision control.
mental results demonstrate a maximum moving speed of 15 cm/s and a maximum jumping height of 8.7 cm. Continuous jumpi
a in order to obtain a more efficient strategy for vertical self-burrowing. This strategy was able to outperform many other standa
ry, this electronics realizes over 5 min of practical operation with a thrust of 1.5 times its own weight, while offering wireless com
n of energy and then high jumping stability (second step) through optimizing the force characteristics. A series of experimental t
ensing and computation triggers an actuator at the jump frequency range that unlatches the spring, launching the microrobot in
mass and maintain contact with the air-water interface. We developed and validated a simple model to characterize the drag fo
speeds as high as 15.1 mm · s−1 (0.33 Bl · s−1) when driven with a heuristically-optimized propulsor. To improve maneuv
rring the home vector orientation from the currently experienced view. Using a compact convolutional neural network, we demo
tation of the flapping plane and imitate the ”force vectoring” mechanism. Besides, referring to the flapping kinematics of
ctural optimization to increase thrust while decreasing weight. For materials, metal-coated polyethylene terephthalate (PET) film
 (success rate > 89%) from the sensory to the motor area in the mBNN was demonstrated. It is achieved in the mBNNs by comb
g the robot's body back and forth. Our prototype robot demonstrated a maximum extension of 450 mm and a minimum bending
eeled modes through foldable structures and variable wheel diameters; then, we present the architecture to control these two m
body mechanics and interactions with the environment, we present an innovative design of a mechanically intelligent limbless ro
tion wheel on the jumping quadruped robot SpaceBok. In this paper, to benefit from both unlimited RoM and efficient MoI-induc
d modules (referred to as Roblets) capable of generating a 2D structure and subsequently self-folding themselves into a 3D co
ss and load capacity were incomplete. In this paper, we calculate the analytic equivalent stiffness of the nematode-inspired cab
mpractical for the ubiquitously aimed robot. To address this predicament, this letter proposes a reduced-order state observer to
technologies: differentiable rendering-based camera pose optimization and consistency-based joint space exploration, which e
e the extrinsic parameters by maximizing the consistency score of the point attributes that fall on each mask. The point cloud a
uce a simple yet effective solution for global optimum extrinsic parameter estimation in the presence of stereo video sequences
me and without memory. Both methods use low-level point-like features, a robust kernel-based loss function, and work with a s
present PeLiCal, a novel line-based calibration approach for RGB-D camera systems exhibiting limited overlap. Our method le
fore, our research introduces a novel calibration method using a uniquely designed acrylic checkerboard which allows the lidar
on. In the final stage, the z-axis (the vertical direction relative to the ground plane) errors caused by the plane motion of the aut
he objects using Delaunay triangulation. Then, we design a novel graph matching algorithm to associate the objects in the two d
directly detected from the events. Such calibration methods can provide accurate intrinsic calibration for both frame- and event-
his paper, we introduce LCCRAFT, an online calibration network designed for LiDAR and camera systems. Leveraging the 4D
involves deriving the state-space dynamic model, embedding the system's dynamics into a recurrent neural network (RNN) wit
wer-dimensional models generalize better over the parameter space, which in turn results in a faster optimization. To speed up
MNet) in this paper. It establishes a comprehensive connection between camera and LiDAR features by calculating a globally-a
ormation compared to traditional hand-eye calibrations. To demonstrate improvements, we apply the approach to two different
ps. A neural network then estimates joint torques from encoder discrepancies. Given multiple samples, we derive the closed-fo
ds usually suffer from issues such as sparse feature maps, unreliable cross-modality association, inaccurate calibration param
method to calibrate the extrinsic parameters between a dyad of an event camera and a LiDAR without the need for a calibratio
nd small viewpoint-difference. (3) Sparse, noisy, and very cluttered 4D radar point cloud. Thus, it is challenging to rapidly and a
ain a heightmap predictor, whose input is the history of depth images and robot states. We demonstrate that with appropriate d
igation policy to reconstruct the environment information in the latent space from corrupted perception and react to perception f
value map. VLFM then uses this map to identify the most promising frontier to explore for finding an instance of a given target o
he former providing the ability to reach a user-specified goal once it has been located. We show that this unified policy results in
Moreover, we investigate the relationship between the geometric prior and the concept of Parameter Sharing in multi-agent rein
ned variance is processed to reduce false negatives (FN). We demonstrate that CollisionGP can use GPU acceleration to proce
as scenarios. During planning, scenarios observed in similar situations are reintroduced (or replayed) as motion predictions. By
er utilizes the prediction outcomes to achieve collision-free trajectory tracking for the mobile robot. The proposed integration of
). Furthermore, these low-dimensional QPs can be solved parallel with GPUs, significantly reducing computational time. High-fi
rld implementations. To overcome these limitations, we propose a Learning framework for robot Navigation in Dynamic environ
ctory planning method for CAMS that combines deep reinforcement learning (DRL) and quintic polynomial (QP) trajectory plan
n terms of the quality of the solutions found and computation speed.
ubgoals that can be obtained in the tree and the graph are utilized for leading the navigation. To determine a subgoal, we deve
ed and data driven path planning algorithms with an average of 87% consistency in producing feasible paths of various length i
eir poor scaling properties, which makes it challenging to employ them for complex problems. In this paper, we explore a simple
 ng and path-following controllers, whose parameters are specified according to the hardware platform. However, two roadblock
ization, paired with distributions with compact support, to reduce conservativeness and ensure numerical tractability. The resul
ing the safety of a twelve degree of freedom motion plan in just over a second. Moreover, the method is capable of discriminati
dule, constructed on the foundation of information distillation networks, is designed to further refine the quality of reconstructed i
o examine and select the generated subgoal candidates. Experiments show that planning from the current robot state to the se
image encoders, which transform obstacle maps and robot footprints into embeddings, which reduce problem dimensionality b
sitions in robot behavior around the constraint boundaries leading to graceful navigation. Further, we present proof for the forwa
faces such as glass walls, bushes, trees, etc.) in indoor and outdoor environments. Further, to handle narrow passages, and na
ersability map based on the SGP representation to guide our planning process. The RRT* planner efficiently generates real-tim
estimators. Secondly, we integrate this model into a collision-aided localization framework and implement it in two widely used
ring algorithm leveraging the differentiable Gumbel-top-K relaxation. Our method accurately infers minimum distances through
cancy for a minimum-sized block. When the adapted Bresenham algorithm reaches a surface of a block, it bypasses grid-by-gr
edictions, and 3) learn an additional safety filter that closely follows the RL controller but avoids the uncertainty intervals. We us
Graph Neural Network (GNN) to predict trajectory success probabilities. We further analyze the number of features in the image
troduced to optimize the sampling process. Finally, the obtained path is enhanced by adding control points and applying cubic p
safety function using the Gaussian process (GP). It is both efficient in data size and eliminates the requirement to store previou
ysObjects, an object-centric dataset of 39.6K crowd-sourced and 417K automated physical concept annotations of common ho
he required number of views to reconstruct an object depends on its complexity. Therefore, we design a deep neural network, n
). The key idea behind our approach is to use implicit representations to obtain the small missing surface areas instead of obse
dal solution for object query initialization for transformer-based 3D object detection models. The proposed initialization method
d objects, identify movable parts, and infer the parameters of their rotating joints. By constructing 3D geometries and textures fr
e decouple the spatial-temporal information by capturing the motion from residual maps and generating semantic features from
ndent and view-independent colors, and leverage two novel consistency constraints to enhance detail reconstruction performan
of challenging appearance variations. We formulate the training objective to account for in- herent uncertainty in individual estim
and single-arm target-oriented manipulations. Experiments demonstrate that our TRTM system can be applied to daily cloths t
g baselines, along with the Amodal Flow Quality metric to quantify the performance in an interpretable manner. Furthermore, we

elation feature extraction method through CV estimation using MLP-Mixer operations; (3) developing an upsampling strategy ba
-grained scenes. Specifically, a multi-attention module CTMA is designed for fusing temporal features from circum-views to fully
such search directions. We introduce IDNet (Iterative Deblurring Network), a lightweight yet high-performing event-based optica
oactive understanding of the visual features' relevance does not require the transmission of the features themselves, enhancing
flectance intensity. Our method is built upon denoising diffusion probabilistic models (DDPMs), which have shown impressive re
of negative transfer — the prevalent issue in multi-task learning — we introduce a task-adaptive attention generator. This gen
at learns attentive features between the 2D dense Bird's Eye View (BEV) and 3D sparse voxel feature maps. Additionally, we p
he contact condition can be identified via the local deformations from the deflection sensors. Prototypes are developed using s
2-d contact model under pivoting, and proposes a rotation measurement method based on the line-features in the stick region.
interactions with unknown objects. While previous works collect only positional information, we include the local normals in a ge
ts such as a hinge. They are parameterized to be differentiable with respect to all of their parameters (rest position, stiffness, co
erated inside the sensor under the external contact interactions. The contours of the contacting objects are outlined by white m
states of the meshes. We also incorporate cross-attention mechanisms to capture the interplay between the objects. By jointly
otion constraint. Then, the pushing planning problem can be formulated as a constrained optimization problem using nonlinear m
 s that satisfy the rolling dynamics and achieve the desired goal; and stabilization of planned rolling trajectories. The theoretical
cant stochastic component. We present a framework for modeling the coefficient of friction as a distribution rather than a single
digms and propose an integrated task planning and skill learning framework named LEAGUE (Learning and Abstraction with Gu
rst presents a robust TTA framework with compound domain knowledge management. Our framework helps the TTA model to
, we propose a novel emph{virtual gradients} based approach for continual representation learning which adapts to each new e
stency distillation method for robot continual reinforcement learning to improve the data efficiency of the experience. Specifical
can utilize previously acquired skills to effectively tackle unexpected novelties. We show that it can adapt faster and generalize
utions : 1) a soft memory module that allows dynamic storage and retrieval of past experiences to serve as context, 2) a self-gu
d framework on multiple tasks in simulation and the real world. Videos are available at: url{https://sites.google.com/view/halp-su
ational Inference by optimizing the uncertainty through backpropagation, which significantly reduces the required model parame
forgetting of previous tasks and adds new skills to exhibit novel behaviors. LOTUS trains a meta-controller that flexibly compos
. In bicopter configuration, the minimum size (equal to ideal passable gap's width) of the system was reduced by 58% through
eshaped action space, we train RL agents that can relocate diverse objects through tactile observations only. Intriguingly, we fin
er lack adequate exploration of the design space or the possibility to adapt to complex tasks. We propose combining a genetic
ed "WiBot" to address these limitations. WiBot is a kinematic chain comprising modular linkages with a prismatic joint and two r
colliding with the obstacles. The controller has been applied in simulations to 1D and 2D deformable objects with varying numbe
ools, namely graph spanners, to design greedy algorithms that efficiently compress 3D Scene Graphs with the aim of enabling
n capabilities in the group. The UAV uses SLAM for its odometry and mapping while sharing this information with UGVs via limit
aviors. The optimal control is obtained by maximizing a posterior distribution of a GRF. A region-based shape control is accomp
cal instruments and retinal tissue. To enable the community to design algorithms taking advantage of such force feedback data
during the labeling process and enabling the training of machine learning models with low effort. The proposed method introduc
pant, such as ignoring the robot or not releasing the object. We also present two baseline methods for handover failure detectio
the camera. We also propose to consider the observation distance in the evaluation process and introduce new metrics to do
ated annotation pipeline leveraging Semantic Segment Anything to efficiently assign semantics. To benchmark progress, we pr
tasks, including object detection, instance segmentation, and embedding sim2real visual perception in industrial robotic graspi
coverage of interactive elements in dynamic scenes. We address this gap by proposing, to the best of our knowledge, the first d
OF) ground truth trajectories. This novel dataset includes six-DOF ground truth trajectories generated using a system of three R
d with RaSim can be directly applied to real-world scenarios without any finetuning and excel at downstream RGB-D perception
current dataset includes 13 larges-scale dynamic scenarios, 90K image frames, 20M 2D/3D bounding boxes of static and dyna
is to exploit a correspondence between human and robot actions that enables transfer learning from human-human to human-
Pittsburgh, and Singapore) and includes high-quality auto-labeled object tracks and traffic light data. We exhaustively mine an
nts and fast trajectory label production. Compared with existing pedestrian data collection methods, our system contains three c
We explore the economics of collection costs and find that for a fixed budget it is beneficial to take advantage of the cheaper hu
oncise representations of regular driving scenes as scene graphs, minimally manipulating their structure and properties. Our mo
e many complex skills, some of which may even require both visual and tactile perception to solve. This paper aims to unlock th
d on the counterfactual perturbation of the human: if the robot had not intruded into the space, would the human have acted in t
ntroduce RiskBench, a large-scale scenario-based benchmark for risk identification. Our benchmark is created using a scenario
pproach and a suitable metric to compare the results on the test data-set. The data-set is publicly available and seeks to make a
xes, 9K important object tracks, and 1-12 objects per video, IDD-X offers comprehensive ego-relative annotations for multiple im
assessing it on another dataset (e.g. KITTI). This setting results in two distinct domain gaps: scenarios and sensors, making it d
images with 6D pose ground truth for ROVs. These synthetic images are generated using BlenderProc and further rendered w
uds, which are utilized to train a range of state-of-the-art segmentation models on both image and point cloud data. We open so
y transit from a single object evaluation to an entire scene evaluation and lower the dimension of the labels to the image space.
R, we propose a semiautomatic annotation approach that computes the positional information of each image. Our method spec
tics platforms, such as UAV and mobile robot platforms carried by vehicles and human operators. In the first release of this dat
pulses, which is then used as an input to the glass probability estimation network. Moreover, the 3D feature similarity estimatio
uncertainty to refine pseudo labels, our method leverages multi-confidence, including transformation consistency, task confiden
ule and a global modality consistency fusion optimization module. Experimental results demonstrate that our two-stage optimiza
performance by emphasizing the dynamic features of articulated objects. Additionally, the paper presents a double voting strate
outliers allowing for robust detection of adverse weather effects. In extensive experiments, we show that our method performs
ues. We perform EdgePoint on various datasets, and show that it surpasses SuperPoint in performance while utilizing only 1% o
near computational complexity. Specifically, the TrT builds coarse-to-dense feature trees, and a novel Tree-based Attention (Tr
its size and diversity, enhancing its potential for generalization. Moreover, we substantially enhance the baseline algorithm by
en amplify the challenge of error accumulation due to noisy pseudo-labels, and pose practical limitations stemming from the he
p detection by allocating more attention to the object-sensor contact area when processing complex 3D shape data. A binocular
nment as input. Using raw point cloud representation allows us to model commonsense knowledge easily and improve accurac
ility than existing approaches which have been dependent on depth images or geometry knowledge. To this end, alongside a c
egion and generates appropriate 6-DoF poses for any unconstrained affordance label. Our method consists of an open-vocabu
ers to make subplane clusters, trained in a self-supervised manner using pseudo-labels generated from RANSAC. Multi-Object
g methods: In the training phase, we attach ArUco markers along the robot’s body and then label the keypoint locations as
ew UAV-based image synthesis, especially from high altitudes, capturing salient scene attributes. Finally, we demonstrate a con
ment. Control policies relying on the embedding are able to operate in unseen environments without the need for finetuning in th
pproaches unsuitable for moving vehicles. In this work we demonstrate the feasibility of visualizing air movement by predicting
ork tackles the challenging relative pose estimation between nano-drones using only a gray-scale low-resolution camera and an
notably enhances the training stability through end-to-end training and improves pseudo-label quality. Specifically, for point-lev
nging horticultural scenario and in doing so demonstrate an end-to-end trainable system that can make use of noisy robot poses
s based on gradient magnitudes.The incorporation of semantic constraints for indoor self-supervised depth estimation improves
on challenge into two manageable subtasks: detecting 2D keypoints and lifting 2D keypoints to 3D. This separation promises en
onfined to a frontal perspective, LiSV-3DLane provides a full 360-degree spatial panorama around the ego vehicle, capturing c
R point cloud sequence semantically similar to the ground truth and thus improves the significance of the deployment for more
ematic feasibility of the motion. In this work, we introduce Neural Navigation for Mobile Manipulation (N2 M2 ), which extends th
frames, and a 2D front-end that predicts the future information embedded in the SOGM representation, potentially capturing th
cking scale-invariant keypoints over time. Our approach can successfully resolve inherent ambiguities in images, enabling keyp
the trajectory going off-road or breaking traffic rules. In this paper, we improve upon goal-based prediction by proposing the Pa
ansitions. In this paper, we propose a deep-learning-based method that leverages a novel long-term memory scheme to achieve
els without centralized training. We also demonstrate the importance of various methodological choices when merging policies,
g the human walking trajectory, which converges to a predicted goal position, and to provide the Transformer model with prior k
stead of the whole team, where Altruistic Q-learning plays an important role in learning the subteam action value. In addition, w
ditions each action on the estimate of the current world state. Despite being conceptually simple, our Statler framework significa
motion prediction benchmark and show that the proposed method achieved state-of-the-art performance, ranking 1st on its leade
ectures, representations of uncertainty, influence of data augmentation, and dataset variability. By evaluating our method on a p
odels, within the same architecture, remains an unexplored area. In this work we aim to determine the impact of using chemical
steps. Firstly, we review state-of-the-art methods and design our lightweight model by removing redundant modules from those
Fields (NeRF) suffer from catastrophic forgetting if trained incrementally. In this paper, we incrementally reconstruct 3D models
es of the two domains using an adversarial procedure. Our evaluation is conducted on a dataset of tactile images obtained from
we reformulate the social motion problem as the refinement of the predicted individual motions based on the surrounding agent
e transmission? To address this issue, we propose a scalable point cloud geometry compression framework (SPCGC) for mach
aches from these two paradigms we get the best of both worlds - efficient approximate solutions from generative AI which are m
ve policy and dynamics models in simulation, predicts environment configurations, and fine-tunes dynamics models with few-sh
he training are obtained from a reference model, solved by the shooting method. Our approach allows for a computation of the l
dering the robot's dynamics. This paper introduces a novel learning-to-plan framework that exploits the concept of constraint m
mines their relocation regions, and pairs them with available robots under kinematic feasibility and task reachability for execution
orward baseline approach. The benchmark problems and associated code utilized in our experiments are made available online
ing as an RL problem that will allow us to adapt both grasp reward prediction and grasp poses. We propose various uncertainty
ed data to enhance learning. In addition, we address the imbalance between labeled and unlabeled data by proposing a contex
d architecture, featuring a bottom-level residual U-block to extract richer contextual information from diverse receptive fields to
ables inspection and task outcome verification capabilities using a large number of positive examples but only a small set of ne
rises four interchangeable neural networks that interact through dependable reasoning components, resulting in a single system
m a target blueprint. We then use a motion planner and optimization to generate part movements and contacts. To train PAST,
e to improve robot-object interactions for various robot end effectors and objects. We further validate our designs in the real wo
when applied to a dual-arm manipulation setup, specifically to the grasp variations. Our quantitative analysis shows that existing
aborative manipulation of rigid objects with deformable objects. Our framework comprises an Action Prediction Network (APN)
le camera viewpoint. To enhance the student policy's robustness against camera location perturbations, it is trained using data
ate that SVFL can significantly outperform state of the art baselines for visual force learning and report several interesting empi
earance. Throughout our extensive simulation and real-world experiments, we find that our approaches perform well in terms of
sual observation to pixel-wise possibilities of picking isolated objects or separating tangled objects and infers the corresponding
s, providing an extremely natural and fast data collection pipeline. Then, using only a handful of such demonstrations, we show
ccess rate while reducing computational load. We present a closed-loop grasp control system based on deep reinforcement lea
eas Critic II evaluates the success rate of the current state-action pairing. To further refine the push-grasping synergy, an activ
formable object goal shapes directly from a small number of human demonstrations. We demonstrate our method’s effectiv
mulation and on a real robot language-conditioned manipulation task to quantify the difficulty of generalization to different (sets
ene in order to obtain high-quality 6-DoF suction poses in real time but also generalizes to novel environments, novel arrangem
this motion across changes in the scene configuration.We show this results in robust robot policies that can solve complex obj
an control the object's state through exploiting interactions between the object, the gripper, and the environment. Designing suc
pulate the object into an arbitrary target shape using image observations. We trained a DRL agent using PPO (Proximal Policy
mpedes their application in broader scenarios. In this paper, we propose a novel framework of Generalizable Articulation Modelin
diverse 6-DoF grasps for parallel-jaw grippers in robotic bin picking. We introduce a parameterized grasp distribution model ba
m and propose a novel approach Contact Energy Based Prioritization~(CEBP) to select the samples from the replay buffer base
h heavily depend on depth restoration networks and the quality of depth maps generated by depth cameras, our system disting
d image with task-relevant features, enabling efficient correction. In particular, the training of the two stages can be performed i
depth and 3D information crucial for robotic manipulation. To mitigate these challenges, we propose an image-only robotic man
icy on real robots, we design a Frame-consistent Uncertainty-aware Sampling (FUS) strategy to get a condensed and hierarchi
mized hardware proxy remains as close as possible to its realistic counterpart, while still enabling task completion. We demons
 ce only. Moreover, the resulting policy improves significantly over the state of the art on a recently proposed real robot manipul
d tasks. Our method, which we dub IDA, enables the efficient discovery of visual affordances for several action primitives, such
g task. In this letter, we propose a novel hybrid architecture for robot grasping that efficiently learns to adapt to different gripper
an encoding and the grid-based strategy are applied to predict grasp heatmaps as guidance to aggregate local points into gras
BVS). HPN-NC can complete 6 degree of freedom visual servoing with large initial offset. Taking advantage of the fully differen
grasping examples, is trained completely in simulation, and is deployed directly to the real world without any fine-tuning. The e
onv-LSTM along with channel-wise and spatial attention dually complemented by a 3D-CNN branch for extracting an enhanced
ents with one based on the vision transformer (ViT) and the other based on the CNN. Our method subtly incorporates the multi
s time-consuming. In this paper, we address these challenges by proposing CrackNex, a framework that utilizes reflectance inf
etwork, where all parts except the modules specific to the task are binary, poses challenges and bottlenecks in quantizing the a
ng our proposed new objective function with supervision from corresponding LiDAR point clouds, our approach efficiently handl
ng nouns and modifiers from images, and propose our method by introducing a dataset and CLIP-like recognition techniques ba
cle-consistency, enabling a simple data collection pipeline and training on unpaired RGB camera views. The key idea is to auton
monly used Replica and TUM RGB-D datasets demonstrate that our approach can compete with widely adopted NeRF-based S
hesizing training data targeted to scenes and objects from the real world. COV-NeRF extracts objects from real images and co
sentations. Next, we construct a reward function using these visual representations through optimal-transport based matching o
nges by leveraging the dependency between the applied and detected contact force on the deformable linear objects. We seam
on experiments in three distinct environments: densely packed building blocks, randomly positioned building blocks, and commo
mputational perspective. It is applied to the balancing control of a rolling bottle on a robotic forearm covered by a custom tactile
he basket. This empowers robots to skillfully reconfigure cluttered scenarios through pushing, creating space for collision-free g
ation using inverse Jacobians that are self-identified by a particle filter-based estimation scheme that leverages the ability of und
ctile data and Hololens2 for capturing visual data. With this system, we collect a dataset of turning bottle caps. Furthermore, we
een a grasped object and its environment using force and tactile observations to estimate the stability of the object during a con
gaits, our force control scheme is derived through backward chaining to leverage the dynamics of the screwdriver and arm/hand
arable to those of human experts with only image information, further validating the efficacy of the proposed method.
ments that involve the execution of tightly constrained peg-in-hole tasks. The experiment results demonstrate a significant impro
asks, DexDLO can perform tasks such as DLO grabbing, DLO pulling, DLO end-tip position controlling, etc. Using the Mujoco p
licate ones like strawberries. Videos are available at the project website: https://sites.google.com/view/everydayfinger.
pressing the target relative position as inequality constraints in the Quadratic Programming (QP) problem. By extending its desc
sical interpretation of the modes enables us to formulate an analytical deformation Jacobian matrix mapping the robot manipula
e of the two methods for simple systems, and shows empirical equivalence on several complex examples. To further alleviate (i
work offers two main advantages. First, it allows efficient global reasoning over high-dimensional complex space created by con
ataset designed for classifying these objects based on tactile sensor output and a new control algorithm for in-hand orientation
eep deterministic policy gradient (DMADDPG) algorithm for autonomous navigation of R-UAV. This algorithm introduces the lea
ices, it dynamically identifies and transitions to the optimal service deployment that fulfills the application's latency requirement,
ation control which, at a minimum, requires exteroceptive sensors to perceive bearing and/or distance to nearby agents. We pre
with ADMM, further reducing computational complexity. Both frameworks are fully decentralized since all computations are par
oalition strategy called K-serial stable coalition algorithm. Different from centralized approaches, it is distributed and complete, m
-time understanding of environments like post-disaster. Based on point cloud data, which is inherently non-convex and unstruct
rential, and (2) o(1/k) convergence when the objective is convex and has a unique solution. Second, we also develop the Incre
s of non-ideal communication factors. Comparison results show that bit error factors cause conflicted allocations. For the pre-a
system (ROS). Then, we establish the hydrodynamic model for the robot by adopting the Fossen approach, thus constructing a
g issues and challenges presented. In the controller, to eliminate the steady-state error, an embedded integrator is adopted by
ristic can notably enhance sample efficiency. In this work, we propose a novel neural-network guided (NN-guided) sampling stra
lly robust for a given threshold. The second objective is to maximize the worst-case spatial robustness value within a bounded
omb friction identification, which also learns motor parameters such as rotor inertia. An approach is proposed for comparing this
y the time integrator. These additional calculations are handled here with a new Newton-Euler recursive inverse dynamics algor
make new connections between the LQR's dual Hessian and the inverse operational space inertia matrix (OSIM), permitting eff
co-simulation scheme was limited to systems consisting of rigid bodies in previous studies. In this work, we introduce the formu
teractions within a compact latent space, capturing robot's traversability. Operating within a reinforcement learning paradigm, R
sists of two main components: the ``Force Planner'' and the ``Gain Tuner''. The Force Planner is responsible for planning both
rted into the hole due to the effect of gravity and wrist compliance if the configuration of the peg is included in the no-escapable
e of initial pose estimation. To refine the initial pose and achieve precise 6D pose estimation, we introduce a novel binocular ed
uence. The search can be guided by either geometric heuristics or graph neural networks trained on data with simulation labels
gent Path Finding (MAPF) simulator that uses a Safe Interval Path Planning (SIPP) algorithm. By analysing the system through
ession maps, which is not always obtainable and accurate enough. The ambient magnetic field has exhibited ubiquity and high
ty to neutralize attacks that can follow an unknown strategy and that, hence, can be viewed as a black box component. We app
ual robots. Starting from these existing solutions, in this paper we introduce a new method bridging these two complementary a
facility. We then employ a Kalman Filter that incorporates local measurements of materials into a global estimation of the mater
possible, even with complete rotor or servo stuck failure, by driving the bi-copter into relaxed hover state through the abandonm
is needed due to the unavoidable joint and environment flexibility. In this way, it becomes possible to compare a desired ideali
earch community. As a typical robot task the dataset includes a pick-and-place application which involves movement, actions o
uch cases to enable the robot to (i) maintain the integrity of both itself and the environment, (ii) recover from failed tasks in the s
ion, our controller identifies the fault implicitly with a neural network-based adaptor. Furthermore, we achieve a unified policy th
bility to enable the printer to operate reliably, even when encountering new corruptions not encountered during training.
o detect LOC in quadrotors. The potential of this metric for LOC detection is demonstrated through both simulated and real qua
related to the AV, and a domain-specific language that enables building propositions over those graphs and composing them th
variety of contingencies in assembly applications and minimizes impact on the assembly completion time.
Line-of-Sight (LOS) features of LiDAR sensor points to enhance exploration efficiency inside large-scale subterranean environ
smission, allowing the UAV to perch indefinitely. The depth-vision system automates the perching and unperching procedures b
bot exploration and dexterous manipulation. State-of-the-art perception systems construct perceptual models via multiple dispa
ribed into a nonlinear programming problem and divided into subproblems to be solved in a receding-horizon planning fashion.
ap to analyze the alignment strength against the principal directions of the optimization as part of its multi-category LiDAR local
greater stability and agile seabed mobility thanks to their legged propulsion system. This paper presents the field demonstratio
ents. We introduce the upstream source proximity (USP) as a natural metric that exploits the joint knowledge represented in the
ents (conducted with OceanVoy as shown in Fig.1). Under experimental circumstances, the Mean Absolute Error (MAE) of the
nvironments without using a prior map. Our approach utilizes a hierarchical planning scheme, decomposing the inspection prob
l efficiency. The proposed design was built to be compact, portable, and lightweight, and it utilises closed-loop control for the in
emonstrate its impact on the object detection accuracy. Then two types of faults are analyzed: fluid leaks, by means of image c
gy consumption. In this paper, we propose a novel pipeline to identify and disaggregate the energy use of mechatronic devices
g UAV pose errors in windy conditions. Compact enough for installation in an economy car, the system emphasizes two novel s
istribution map, followed by high resolution sampling informed by the low resolution map as a prior. The low resolution prior act
 ime and cost. It requires a team of people to go on site and spend typically a day to monitor the deployment. It is vital to not on
tion method that fuses total station prism measurements with IMU and GNSS heading information for high-precise pose determ
n situ particle diagnostics. The flight control uses computer vision to localize and then trace the movements of particle-laden flo
e data recorded from a receding horizon estimator to train a self-supervised neural network for traversability prediction, eliminat
comparative analysis, we also fabricated a ViTac sensor devoid of biomimetic tips, as well as a TacTip sensor with opaque sk
showing high sensitivity and fast response. Based on the piezoresistive sponge, a flexible large-area tactile sensor is designed
ntal laboratory environments, it is key to design systems capable of capturing and processing data in real-time. Within this pape
slip identification. In this work, we introduce the notion of the Tactile Dipole Moment, which we use to estimate tilt torques from
an object to a goal. To achieve real-world performance, we adopt a sim-to-real approach. Our results demonstrate that it is pos
ural network based two-class classification technique to detect contacts even with environments not seen in training. In experim
e three-dimensional elastic displacement of the elastomer structure upon contact. Such displacement is tracked by detecting the
he cameras and motors, we optimize the linkage transmission with a planar linkage mechanism simulator and develop a planar
urations, the sensors can be easily manufactured and require minimal optical tuning to obtain high resolution depth reconstruct
tasks, and learning objectives involved. To address this challenge, we present a unified tactile embedding space capable of pre
ze of a human thumb while with a large contact surface. We show the ability of AllSight to learn and estimate a full contact state
Tact builds upon the optical principles of DTact and is optimized to achieve 3D shape reconstruction with enhanced accuracy a
ection process, the camera rotates to multiple angles to capture the tactile image of the contact area. To stitch together the tact
es photometric stereo to reconstruct the tactile imprint generated by an object during contact. To this end, two cameras were int

s work presents the first attempt to discriminate compliance through soft optical tactile sensing based on a computational mode
ack: bi-pushing, bi-reorienting and bi-gathering. To learn effective policies for challenging tasks in simulation, we contribute sev
ly collects tactile data and reconstructs 3D shapes of the objects as well, which can serve as a representation for higher-level d
ertain robot leg configurations or in the lack of interpretability of the extracted features. In this paper, a terrain attribute extraction
g a slippage estimation network when slippage is detected. The estimated object velocities are integrated into a geometric-kine
% of cases. Given these strong "tactile-only" results, we argue that the fusion of both tactile and visual sensing can bring cloth u
approximation that only requires calibrating 3 variables. Our method is shown to reconstruct normal and shear forces significan
er Bragg Grating-based flexible contact force sensor to capture gripper-vegetable interactions. The sensing module was 3D pri
hear can also be perceived by analyzing the Voronoi diagram. A series of experiments are conducted to verify the performance
ouch sensing accuracy and mitigate the ghosting effects, where false or misleading touches may be registered. Furthermore, o
g its sensing proficiency, and present a user study underscoring its perceived safety and comfort for the assistive task of limb m
cleGAN while including two additional loss components aimed to accurately reconstruct background and contact patterns includ
vision-based tactile sensors are predominantly applied to measurements within a limited area. In this study, we design a novel
method to generate realistic marker-embedded tactile images in GelSight-like vision-based tactile sensors. We trained the prop
pose of the grasped object. Our method is tested on diverse objects and configurations, achieving desired manipulation objectiv
The 3D experiment results demonstrate an average reduction of approximately 34.1% in shape reconstruction errors across al
ive information to leverage the advantages of both primitives without being overly disruptive. In doing so, we achieve a 10-fold i
able for the response of iontronic tactile sensors. A simple denoising method is provided to reduce initial array noise obviously.
arns to perform servoing from raw sensor data. The predictive model is formalized as Conditional Variational Autoencoder with
leverages contrastive learning to integrate vision and touch sensations in a self-supervised fashion. By availing both sensory in
generalizable to a wide variety of robots, enabling them to predict and avoid dangerous collisions and reactively handle acciden
nerate prescribed but tunable robot motions to reveal high-dimensional sensory feedback, which is not obtainable from static im
 provide degenerate motion analysis on general local and global sensor-paired systems, offering insights applicable to a wide ra
ed version optimised for readout linearity which provides reliability in contact state estimation and force tracking. While a tactile
predict the maximum von Mises stress on the flexure element using design parameters. Subsequently, we clarified a way to att
uations, and calculate contact location analytically. In addition, a robot motion control approach based on the contact sensing m
BC), which takes into account the full body dynamics, the actuation limits, the external disturbances, and the interactive constra
urbance Kalman filter (GPADKF) is developed where the variational Bayes technique is employed to achieve online identificatio
stracting the task constraints experienced by the object and transferring them to the robot's operational space frame. Second, u

e-tree in the hybrid space by combining a sampling-based planning method with a novel bandit-based technique to select and i
ts demonstrate the feasibility of the approach.
way of evaluating robustness compared to previous approaches by considering model-based predictions. In particular, we use
mathematical framework to handle such geometric constraints. Nevertheless, their recent adoption in robot learning has been la
on. Our approach combines automata-based and MILP-based methods and leverages the strengths of both approaches, while
zed. We tackle the problem using an iterative approach. In each stage, it selects edges' traversal duration and success probabi
e. This paper develops methods to ensure the safety of robots operated over communication networks with stochastic latency.
solution, we define a Mixed-integer linear programming approach to solving the task planning problem over a Varying Weighted
ing of task instructions during the training and then encoding the context variable via the Transformer again for improved task p
many aspects of human-robot interaction that cannot be fully expressed and many assumptions that must be made in each mod
ogic specifications for local reactive tasks when encountering dynamic events. Specifically, inspired by the causal knowledge gr
rementally to represent the task progress and does not require sophisticated product automaton, which significantly reduces the
rithm that composes task-agnostic skills learned during training to safely satisfy a wide variety of novel LTL task specifications.
or into two-dimensional coordinate values in the distribution function upper plane. To confirm the validity of our model, a paintin
onent motion primitives are constructed according to the criteria we specify, the resulting behaviors are probabilistically verifiabl
realized in a graph-based representation well-suited to algorithmic manipulation and composition, exposing synergies rife for o

asping position detection,which shows good performance in both synthetic and real scenes. Inspired by human grasping, a tact
nsion mechanism similar to what happens with human tendons or another physical phenomenon yielding the same effect. This
mploy a single expert grasp demonstration during training to learn a local implicit surface representation model. At inference tim
predicts a distribution over the unobserved parameters while the decoder predicts action feasibility, which can be used in an u
t models by learning graspable points, associated grasp configurations from observed point cloud in a parallel manner and opti
function-to-function transfer using sphere primitives to establish semantically meaningful categorical correspondences, which a
camera software (to inject a pixel-intensity change on a single pixel). We then use evolutionary optimization to obtain attacks th
e data or coverage for objects in a database. In this work, we propose to perform object identification with multiple reference dat
lation relationship detection algorithm called Graph Sampling Aggregation Network for Visual Manipulation Relationship Detecti
er methods to learn the relation between sound wave modulation, the applied force, and its location, respectively and picked the
n generated with Quality-Diversity (QD) methods on 3 different arms and grippers, including parallel fingers and a dexterous han
o control these grasp characteristics in the synergy space. This paper presents an adaptable representation for kinematic grasp
ncy of our sampling strategy: We dynamically scale the sampling region size and number of samples in it based on past grasp
wo Fractal Fingers, coupled with a closing actuator, can form an adaptive and synergistic Fractal Hand. We articulate a design
arget object. Most grasp detection algorithms directly predict grasp poses in a monolithic fashion, which does not capture the co
andardized state machine interface for control, which allows for integration of most manipulators with minimal effort. In addition
 eemptive feedback, limiting their usability in realistic domains. In this work, we aim to endow a robot with the ability to monitor a
ed models, known for their robustness in noisy environments and their proficiency in recognizing essential features, but also th
-world without the guidance of explicit state estimation and dense rewards can lead to unsafe behavior and safety faults that ar
hanges, enhancing robustness in dynamic scenarios. Riemannian Gaussian Mixture Models are employed to assess reachabili
obot and achieves a success rate of 78.64% after training with only 500 real-world grasp attempts and 87% with 4500 grasp att
parate contact maps for each fingertip on the object point cloud, effectively capturing the specifics of finger-object interactions. I
s, addressing issues arising from uneven surfaces and intricate features that interfere with suction sealing. We categorize PCB
everage human-like object understanding by reconstructing and completing their full geometry from partial observations, and m
ng the results. We then show how to analyse these results, in a statistical framework which accounts for the variability between

otor, the developed gripper is equipped with two novel self-motion switching mechanisms. These mechanisms switch their moti
two-fingered gripper that relies on the deformation sensing of elastic elements in the drivetrain to obtain finger force. By using
h different equivalent mechanical properties. Adaptability tests were conducted to demonstrate the effectiveness of the propose
ers, robot picking of rigid objects can be accelerated. The complete high-speed picking solution is experimentally verified. A tim
nsport. To test our analytical guidelines, we built a vibrating surface using flexures and a voice coil actuator that can accelerate
in the time required for each operation. Through the integration of particle jamming, we have enhanced the overall stiffness an
alysis of the magnetic gripper are undertaken, wherein the relationship between the open angle and the external magnetic field
abnormal vibrations. We detail our design requirements and specific design choices for the mechanical and electronic design a
hat can automatically derive destructive disassembly actions from a symbolic representation of the disassembly state. Viable de
esign, the hand is compliantly stable at every equilibrium condition.Therfore, the Fractal Hand adapts to a very wide range of pl
coop-shaped fingertips that can form a single plate when fingers are flexed. Data obtained from an on-board camera are used t
n. For efficient search, we present the co-design Jacobian that describes how the robot's motion varies with control as well as h
s and pad damping allow for higher impact speeds. Thus, these features should be considered in the design of future agile robo
f objects in all three axes of rotation by imitating human use of chopsticks. HASHI delicately positions and orients food through
erforming everyday dexterous tasks. The D'Manus comes with ReSkin sensing on the entire surface of the palm as well as the
sping performance remains limited, and tasks that involves grasping various objects are yet to be achieved. Here, we develope
s. Gecko-inspired adhesion provides extra lifting capability for objects up to 2.3 kg, using a 7.6 x 12.7 cm adhesive swatch – 2
satisfy all the aforementioned properties within a single framework. We conduct comprehensive evaluations of our approach o
on accuracy by 0.0223, 0.1416, 0.2735 and 0.1592 in these datasets.In order to verify its performances on dataset with rich spa
many scenarios, such as teleoperating a robotic hand with anthropomorphic kinematics but with different size to human. Here we
(with and without the feedback). Standardized questionnaires evaluate users’ satisfaction and workload, enabling a system
actuators, and achieve the precise production of vibration patterns that can focus at any desired position on the board. Specific
at re-directing the motion of the trainee’s hands and wrists toward targets or away from obstacles, and providing a quantita
racterize the performance of Hapstick, showing that the bending stiffness increases linearly with the increase in applied vacuum
al impedance matching and focusing.
only the most informative aspects of an object during data collection. Here we use an active learning approach that uses a data-
sical display. We demonstrate our display’s ability to create 16 shapes with 8 different boundary conditions. Our display is
ng incremental mesh data via slice-style management for guidance quality. We also design an incremental mesh rendering alg
p velocity optimisation framework. The proposed method employs torque data captured from past steps through a Series Elasti
ch together enhance the input variables for the fuzzification process. The enhanced FS-LSTM approach is evaluated using rea
order to both avoid explicit labeling procedures and have powerful nonlinear fitting capabilities. Experiments involving 10 health
y maximising the benefits of all agents involved) to human-robot cooperative tasks promotes a more seamless and natural inter
ve been identified in the annotation of grasp bounding boxes within the popular Jacquard Grasp. We propose utilizing a Human
ance of the return from the current state. We iteratively improve this estimate during training using a Bellman-like recursion. On
UAV operators, regulators, and companies developing this technologies to produce a more capable workforce capable of consis
m using continuous-state Partially Observable Markov Decision Process (POMDP), and use a physics-based digital human sim
tion pairs). Our approach manages to produce queries which are closely aligned with those labeled by humans, while avoiding
allow for humans to provide correction to robot action in real time, but intervention policies are often based on explicit factors re
Using our proposed simulator, we create a high-quality dataset with vehicle-pedestrian interaction data from safety critical scen
gn PERCEIVE, a human preference-aware rebalancing and charging framework for shared electric micromobility vehicles. Spe
ntly large safe region. In the online part, a remote control-mapped primitive is designed to fast encapsulate pilots' inputs, and th
ough an electromyography interface. Instead, to react to sudden load releases, we set a power limit on the tendon, and continu
n of robots to design an elastic interaction controller. The controller not only actively adjusts the deformation of the gel to provid
hen subjected to workspace restrictions due to unplanned interactions and contact loss. The first approach employs an augme
AR) device to modify a robot’s movement space in 3D. A user can define virtual borders in the real world with the tablet and
Markov Neural Operator designed to capture pedestrian behavior across diverse scenarios. Our paper models trajectory predi
ations to the learning process that encourage these RL policies to maintain additional caution in unfamiliar situations. Specifica
participants to select among various paths with different levels of} time-risk tradeoffs. The data shows that participants exhibit
ven mobility. However, they fail to leverage the diverse contextual cues of a home environment. We present a novel machine le
esize diverse pedestrian motions via local coordination factor (LCF). Our experiments have demonstrated that the manipulation
eural networks, limiting the planner structure to simple ones, e.g., sampling-based planners. Despite the ability of gradient-base
ented. 1) By assembling directed detours formed by nodes and segmented paths formed by edges, a generalized path consisti
ajectory-and-environment continuous-optimization. We show that the evolutionary method is well-suited to explain sampling-bas
m the demonstration. For the real-world experiment, the proposed system leads to satisfactory grasping abilities and learning to
ognition performance, the study proposes an EEG Vision Transformer model coupled with a 3-D spatial representation to captu
from multimodal data (e.g., images, sounds, haptic signals). We demonstrate the effectiveness of our technique by benchmark
tion formulation to introduce a distinctiveness metric that aggregates individuals with similar preference profiles. Analysis of ano
we show that the semantic aspect allows SCONE to learn self-confidence faster than existing approaches, while also achieving
ttributes. They can be used to segment LiDAR point clouds into two parts: traversable and untraversable parts, and then an act
ptible to pixel electrode drops. The sparse electrode-skin disconnections rooted in issues such as low adhesion, sweating, hair
ults indicate that the proposed method outperformed other state-of-the-art methods (PI-IRL, BP-HMT, RNNIK-MKF and CMk=5
ogy trick, which poses a significant threat to trust maintenance. The human's trust in the robot will collapse when they discover
ns. Finally, results of our user study indicate that Haru imitating human player’s game strategies via our method is perceive
udies, we propose a system for guiding visually impaired users running outdoors using a drone-based robotic system to locate a
s of an extremely low-speed U-turn test are used to show that the proposed methods provide stability while maintaining drivabili
paration. Our system incorporates a multimodal active perception module, a human-in-the-loop long-horizon planning through a
the future system states. This method was then tested on 10 real human subjects’ data for the interaction task and the sim
nd automatically creates a new path using this map. We believe that large language models (LLMs) can be used to understand
including priority definition, spatial semantic information about ad-hoc geographical areas, and reference waypoints, which are
hrough holograms from a mixed reality device, gaze can be utilized effectively to estimate the properties of the defect in real-tim
constraints of a large-scale user facility. We demonstrated this framework by integrating a robotic arm to pick and place a sam
test sites as part of the DrapeBot project. The goal of the project is to enable human-robot collaboration in the process of carbo
MG) sensors. The study explores optimal sensor placement for maximizing the correlation between muscle activity and the estim
erefore, in this paper, we introduce the first language-guided active sensing approach that allows users to observe specific parts
e kinematic state-dependent and pure gait phase-dependent control.} These combinations were tested during overground and t
ect from robot rooms, the results of which will inform both the iterative design of the robot room and the thinking of our commun
ain and compare different deep learning models. Results show that automatic trust estimation is feasible, which opens up to us
controller is employed to regulate the robot's motion and interaction force with the target person, enabling the robot to maintain
eralization ability.
bution temporal logic (scLDTL), a logic that is defined over predicates of states as well as belief states of partially observable sy
at depends on the robot's future behavior. First, we propose a novel model-based formulation of conditional behavior prediction
m beliefs. We propose a novel blend of epistemic planning and POMDP techniques to create a risk-bounded AI team assistant

complementary mapping pipeline is developed to utilize the built-in SLAM capabilities of the AR-HMD as shareable data. Our s
. The resulting networks predict the intention of the observed human or robot based on input RGB videos. They are trained in a
ment times. The proposed framework is experimentally validated on 7 manipulation tasks and we show that CoBT achieves app
amework plans a collision-free trajectory for the robot. Subsequently, predictability is taken into account and a set of virtual tube
e authority allocation in this shared control framework considers both the limitation of learning-based methods and the higher d
n study with 12 participants. In this study, participants were trained on the use of one of the previously mentioned audio commu
red autonomy (SA) which enables synergetic collaboration between human operators and robots by dynamically allocating auth
sue, we propose a framework that learns a robust caregiver’s policy by training it for diverse care-receiver responses. In ou
e then propose a priority-changing shared control method to effectively combine the surgeon and robot's strengths by dynamica
f Online Corrections (DROC), an LLM-based system that can respond to arbitrary forms of language feedback, distill generaliza
roposed. This novel guidance scheme enables operators to effortlessly and simultaneously control both orientation and position
s. The framework is implemented on a mobile manipulator with a weighted whole-body Cartesian velocity controller, which mee
trol authority is allocated seamlessly between the human operator and the autonomous agent based on the prediction of opera
this paper, we further extend the FIC framework to bimanual operations, allowing for stable and smooth switching between diff
ge. In this paper, we use LLMs as translators between NL information and a structured AI task planning problem, targeting hum
ot collaboration. We formulate the problem using a generic multi-agent planner, show how to generate visual explanations thro
es path preservation, exploiting redundancy to minimize task execution time, while explicitly adhering to the constraints imposed
ped object. First, human safety is defined as bounded approaching velocities between human and multiple robot links based on
return trajectories that can be travelled at greater speed and in less time, while guaranteeing the safety limits according to the
essing distinct abilities are served as domain experts. Our proposed navigation agent, namely DiscussNav, can actively discus
arious aspects of the behavior but also within the action selection process, which is closely intertwined with personality-depend
on of preferences to account for highlights: state-action pairs of relatively high information (high/low reward) within a preferred tr
dels in the context of real-world co-painting do not perform well because they (1) do not understand the constraints and abilities
materials. Our methods achieve respective 40.2% and 51.1% of mIoU on COCOStuff-10K and DMS datasets, surpassing the p
s aim, we also present a legs tracking algorithm that utilizes 2D LiDAR sensors integrated into the mobile base to monitor the h
ions for libraries composed of Via-Point Movement Primitives (VMPs). By building on Riemannian manifold theory, our approac
f post-fall recoveries, we designed a task-space impedance control with high damping and low stiffness, where the SuperLimbs
. We show how this approach enables greater robustness to variation relative to geometric and reinforcement learning baseline
competence, as well as their trust in the robot moving forward. It is valuable to characterize the effect of these biases within hu
ations of motion in the environment and which can be used for data-efficient, interpretable motion prediction (MoD-LHMP). To a
 arrow context of object classification on static image datasets. In this paper, we combine ideas from continual learning, semant
es. In this way, the robot can select some complementary strokes and draw them on the canvas. To train and evaluate our inpa
ourney is the user training required to achieve an acceptable accuracy of the control. In this paper, we aim to identify the best t
our new method and system significantly outperform baselines on multiple metrics and have potential practical value.
aesthetic understanding. In the robotic drawing stage, an optimization method was presented for drawing smooth and physical
ual life-sized representation of the cumulative 3D map overlaid onto the real environment. As such, the operator can effectively
based on a walking trajectory that follows the law of conservation of energy and formulated by energy-states. We evaluated ou
rusion experiments with a bipedal robot. The results confirm the model accuracy on the given type of granular terrains. The mo
ontroller shows comparable cost of transport and greater adaptability to terrain changes compared with certain existing method
ors aligned with the data distribution of mixed reference motions. Additionally, we employ a specific Integral Probabilistic Metric
atly reducing the computational effort of online planning. With tailored objective function and feet complementary constraints, th
he spikes from the motor neuron in the third layer, a reflex matching the intensity of the stimulation can be evoked. Experimenta
obots, its implementation using both simulation and hardware data for a humanoid robot, and a method for estimating lead time
he global body dimensions. The middle muscle layer consists of flexible patches and layer-jamming bars with tunable stiffness f
erlying robot sensor suit. To validate the approach, we show how the proposed sensor fusion can be integrated into a two-leve
Deployment Gap and Feature Gap between simulated and real data. We investigate the effects of these two types of domain g
and produce improved multi-modal BEV features. Specifically, we devise a Semantic-guided Flow-based Alignment (SFA) mod
D detection framework for VIC3D task, Enhanced Multi-scale Image Feature Fusion (EMIFF). To fully exploit holistic perspectiv
s. We use a vision transformer to extract image features, which are passed into a transformer encoder that reasons about temp
ontrastive learning, ensuring balanced dependence on fog depth hints and clean matching features. This framework helps enha
epresentations across spatial and channel-wise dimensions. Through these attention blocks, more globally generalised represe
r an ideal situation where all vehicles are equipped with identical LiDAR, ignoring the diversity of LiDAR categories and operatin
a new method, dubbed FastOcc, is proposed. By carefully analyzing the network effect and latency from four parts, including th
ding 3D annotations. LPFormer consists of two stages: firstly, it identifies the human bounding box and extracts multi-level featu
ield inferior prediction results, particularly concerning smaller and long-tailed objects. To address these issues, we propose Mo
ficult scenarios like occlusions can harm tracking performance. To address these issues in a unified framework, we propose Sh
on gap on existing multi-agent perception systems. To break the data silos, we introduce the Feature Distribution-aware Aggre
nt perception system. To address this concern, we frame the task as an adversarial attack challenge and introduce ADVGPS, a
int training, promises to solve this issue. So far, however, the evaluation protocols between the two methods were incompatible
ces and complementation for individual unaware instances. Taking camera-based vehicle-infrastructure perception as a typical
LiDAR imaging model with the spatially non-uniform distribution of point clouds caused by LiDAR structure. The fused results ar
while causing unnecessary bandwidth cost. To mitigate these problems, we propose a novel feature aggregation framework ca
proposed methods mainly use cooperative detection results as input to a standard single-sensor Kalman Filter-based tracking a
s driving, where a network is trained to imitate the control commands of a human. In this setting, two situations arise: 1) Closed
RIL) that learns the mapping of cross-domain sensorimotor skills and utilizes it to generalize to unseen tasks given a human vide
ates where the future states are predicted by a transition model. A trajectory contains numerous actions before the task is done
erties of visual foundation models enables unprecedented learning efficiency and generalisation. Videos and code are available
ard function for reinforcement learning. The challenge is determining how to ground visual demonstration inputs into a well-sha
how that it is possible to optimize for learning the reward function up to a behavioral equivalence class, such as inducing the sam
n unfolding policy and learning from demonstration to transfer unfolding actions). A mixture density network is introduced, with
parison, offline metrics such as mean squared error and inference time are used. In addition, the optimized models have been e
 bjects with low-dimensional configuration spaces, such as planar polygons. We instead propose a practical sampling-based app
t, as an example of utilizing the rolling orbits, a manipulation method to rotate the ball around a vertical axis is developed. Finall
nt simultaneously. Based on screw theory, the kinematic model of the spherical mechanism is established, and the maximum av
e large feature reduction may cause difficulty and inefficiency in policy training. To address this issue, we proposed Curriculum-
ost part do not capture the rich, nonlinear behavior our robots need to exhibit, shifting the burden of discovering these behavior
e to maintain desired contact mode regardless of parametric errors. Real world experiments were conducted to illustrate the acc
tate space by following actions from these controllers. We are the first to demonstrate learning hard-to-explore finger-gaiting in
about robot actions. The method, trained in a simulated environment and then deployed to a real robot, is applicable to various
moving during manipulation. We present a control strategy based on coupled dynamical systems, whereby the fingers move in
ps one after another. The synchronization between the parallel mechanism’s motion and the suction cups during locomotion
zation techniques. Subsequently, the trained model predicts grasp quality, guiding the optimal grasp pose to the robot's controll
speeds, and learning from partial trials. The experimental validations on a 6-axis industrial robot with a laser tracker absolute m
c surface treatment. RoboGrind is evaluated both under laboratory and real-world conditions in the context of refabricating fibe
propose the LLM-driven Multi-vehicle Dispatching and navigation (LiMeda) framework. It comprises an LLM-driven scheduling
ex Network (SCRNet), which leverages the Retinex-based structure and is guided by the principle of spatial consistency. Specif
sk execution. As a proof-of-concept demonstration, the proposed framework was evaluated in a bi-manual setup, where the mi
police district in downtown San Francisco that demonstrate RoSSO's capabilities on our new formulations and the prior work.
n can take care, with little effort, of the mobile base navigation, self-collisions, and collisions with the environment, as well as se
mptions about the system dynamics or underlying uncertainty distributions. By parallelizing our approach, we are able to demon
e. We present a unified, interpretable, and efficient autonomy framework that moves away from cascading modules that first per
vity is also enhanced by leveraging the differentiable structure of the parameter space. The proposed safety filter is experiment
p between vehicle dynamics, road slope, fuel consumption, and control commands using the attention mechanism. Based on th
uced, which comes into play when the desired terminal state becomes infeasible considering the disturbances. Both of these co
ry proposals for the autonomous vehicle, which are further refined by a sampling-based technique. This combined approach en
logy streamlines complex specification extraction from map images and enables the efficient generation of long-horizon robot m
MPPI is more efficient and requires substantially (20-times) less number of rollouts and (4-times) smaller prediction horizon wh
anner that leverages these two contributions to produce dynamically-feasible trajectories. The proposed framework not only red
ormance, closely matching or outperforming the former.
del. Our proposed techniques address challenges arising from the sequential nature of trajectories, and potential benefit of prior
which learns to predict the evolution of future riskmaps conditioned on the future path and speed profile of the vehicle. The mo
ethod is evaluated on a set of robotic manipulation benchmarks and is shown to outperform state-of-the-art reinforcement learni
that Motion Memory can contribute to reduce the planning time when facing a new planning problem. We provide extensive exp
oposed to alleviate the issue by incorporating de-confounding supervision. Additionally, to better capture the environmental cue
al paths in the training set, making them incapable of finding good paths in some scenarios. In this paper, we propose a Multim
s of subgoals for the MPC to follow. Our method, called Subgoal Diffuser, generates subgoals in a coarse-to-fine manner, prod
ransitions from sequential runs of kinodynamic RRT and a bandit algorithm decides what region to sample at each timestep. Th
method provides a tight superset of the poses which could explain the measured points. This naturally detects ambiguities relate
the aim to improve the probabilistic prediction of condition metrics, specifically RUL and SOH. Therefore we propose to identify
ompetitions and serves as a portable test and data generation system that captures and grades performances, making compari
conveniently calculable without the need for any additional boundary condition measurements. This eliminates the need to use
ing machine then accurately stitches along the generated path automatically. The vision system utilized in this study achieves a
 of concave areas. Therefore, an automatic segmentation method is developed to find areas with a uniform machining strategy
solved in the Decentralized Partially Observable Markov Decision Process (Dec-POMDP) framework. Results show that the pr
n to the edge, that ensures safe and collision-free navigation even in the presence of failures in the connection between the on-
CBS in the mixed traffic. Additionally, virtual goals are introduced to search CAVs’ priority and outlets along with path findi
lso analyze the effects of contrastive divergence in various autonomous driving scenarios, under multiple network infrastructure
study), rendering most of the database wealth wasted. We propose a novel scenario database design and the first ODD-based
g of a mixture of AVs and human drivers. The analysis establishes high reliability upper bounds on human supervision requirem
s are used to compute forward-looking risk assessments along the ego vehicle's path, guiding its navigation. Our method aims
s, allowing the method to produce a more accurate estimation of destinations; 2) achieves map-compliant predictions by gener
nts involved are addressed appropriately with convex approximation, such that the convexity property of the reformulated optim
CBG) model for this purpose, which allows the users to specify the level of courtesy of the generated trajectory while ensuring re
s their perception of the driving environment, their utility evaluation of predicted future environments, and their driving style in te
ce, to mitigate the occurrence of safety-critical failure cases like false negatives, a safety-aware training strategy is needed to e
anges and reduces the model's learning capability. To alleviate this problem, we propose a new Polar Cylinder Balanced Rando
tates accurately. STT consumes rich appearance, geometry, and motion signals through long term history of detections and is
collaborative perception framework that incorporates communication optimization and a judger mechanism to facilitate CAV dat
up to 53% on realistic benchmark instances.
equency domain, which generates an estimate of lumped uncertainties. To streamline the design process, we utilize a coordina
paper, we present a fully-fledged framework that monitors active vehicles using camera images and state information in order t
ypes and 3D bounding boxes, our method focuses on motion segmentation and clustering, enriched by a motion estimation mo
ecise observations use global information, such as environment outflow, and local information, i.e., vehicle positions and veloc
prompt tokens. Instead of simply converting perception results from the separated train model into pure language input, our no
nning cost, we propose a learnable context-aware cost function with latent interaction features, facilitating differentiable joint lea
ctions. To resolve these issues, we propose the Semantics-aware Interactive Multiagent Motion Forecasting (SIMMF) method
ese labels to perturb the data by deleting non-causal agents from the scene. We evaluate a diverse set of state-of-the-art deep-
ocity increases in more than 60% of the 230 cut-in scenarios from the highD dataset, without sacrificing safety.
for a set of task scenarios, decomposes the control into nominal components that are effectively solved by conventional contro
emand, applying such methods to a large urban environment can be computationally expensive due to the large number of tax
ic scenario libraries collected from other AV models with distinct behaviors, attempting to extract transferable information for cu
, network architecture, and training paradigm. Experiments show that our method outperforms other SOTA methods.
y prediction. Our framework explicitly incorporates map constraints into the network via two carefully designed modules named
e a reference trajectory inside the selected driving corridor, which can be used to guide or warm start low-level trajectory planne
ehicle behavior descriptions. eMem-NDT is a neural-backed part of a pre-trained deep learning model by changing the soft-ma
Additionally, the triggering conditions of warning are mostly rule-based threshold-checking given the current state, which lacks
orates a Bayesian inference approach to actively learn the agents’ parameters, i.e., other drivers’ model parameters. T
t to learn situations that require controlled traffic rule exceptions. Our method is applicable to arbitrary RL models. We successf
ing (RL) framework with policy transfer and enhancement, to improve the convergence speed and ensure a high-quality policy.
ion, proximity, and driving speed. Additionally, we develop a ''dynamic traffic graph'' using Convolutional Neural Networks (CNN
from semantic maps during agent state encoding, our approach takes into account both the social features exhibited by agents
ility of the intention estimation model. In this paper, a subject-independent continuous motion estimation method combining con
her SOTA models that only use the arm's position, making it particularly well-suited to assist in tasks where the operator's inten
 G filtered signals and derivatives; allowing for a simple SVM classifier to achieve unprecedented accuracy and realtime perform
e neutral face and microexpressions. To extract a reliable optical flow field from the expression information, we design an optica
onstrate are important features for predicting driver intention. Using our handcrafted features as inputs for both a transformer an
rios at scale, and augment it by adding complex language queries for 40 object types. We demonstrate that a robot using our m

sources of information and challenges caused by inadequate modeling of ego-motion and scale of pedestrians. To this end, we
ectory predictions to discern potential hand motion endpoints. This is statistically combined with gaze fixation data using a weig
nergy and fast closure of the claws by linking the motion of an ankle linkage and the claws through soft tendons. We study the d
ed self-healing electrostatic actuators. Our artificial muscle design decouples the high-voltage electrostatic actuation from the p
tecture utilizing an Archimedean spiral cam. The design and fabrication process are discussed, considering system design con
plex geometries while maintaining consistent scrubbing force. This study uses a hybrid force-position control scheme and a low
hm are conducted under same actuating forces and external loads. Results show that there are multiple solutions to the kinetos
on path: total bending angle and equivalent tendon length. Then, based on the optimal control theory and minimum principle, th
configurations with different shapes, achievable with the same actuation inputs. A global stability criterion is also established wh

grated approach of experimentation, modeling, and validation, offering preliminary insights into the torque transmission charact
s are subsequently processed through a predefined, differentiable forward kinematics model, resulting in a decoded representa
eformation. Particularly, the dynamic model of the robot for the whole jumping process is also developed relate to multiple facto
to control the traversal of a one-dimensional path in such a way that a reference path length of a multi-dimensional path is appr
ed-form solution to the inverse kinematics problem. These relationships are derived from the unique characteristics of the bisec
ons. An implementation of solution is available at https://github.com/aabecker/dubins3D
novel performance index using an approximation of the capacity margin. The proposed index is continuous and differentiable, c
ted using a modified variant of the Particle Swarm Optimization for the robot design problem. In the proposed approach, we gen
escriptor. Concepts from screw theory were used to make this representation time-invariant and also invariant to the choice of b
onstitutive modeling. We therefore focus on developing and applying primitive deformations that each observe this constraint. B
ach to achieve the desired directional pre-impact flux for the robot through a combination of a dynamical system (DS) for motio
ves of RASCAL and how they address some of the limitations of existing robotic systems in this area, such as scalability and re
ve-joint space is proposed, which not only simplifies the kinematic analysis, but also decreases the kinematics nonlinearity. By
nditions, both well known in optimal control theory. We will also (i) prove that there is a link between some singularities of the CD
motions, allowing to predict the impact of design choices. The results have been validated experimentally on two cable types a
ffect are proposed. First, the required wrench of the MP is mapped into the twist space by the motors' torque-speed relationshi
ectories for the robot’s dynamic behaviors and stabilization of these trajectories using a Time-varying Linear Quadratic Reg
control, the stability and performance of the closed-loop system are guaranteed. We demonstrate the GP-enhanced control exp
Specifically, we quantify the conservativeness by the magnitude of the energy function, and we reduce the conservativeness by
CBF), where we parameterize the class-K function by a neural network and train it together with the reinforcement learning polic
of failures to occur. Then, we propose a modular and hierarchical method for safe robot manipulation by augmenting base skill
alities. Moreover, connecting predictions from such models to the environment at hand to ensure the applicability of these pred
s of the ``approach" phase to bring the wooden dowel closer to the hole and the ``insertion" phase. While the proposed model a
scribes the mutual relationship among joint angle, visual information, and tactile information from the feet. We aim to train this n
ensive relearning. In this work, we propose three strategies that utilize semantic similarity and relatedness between the variable
etwork controllers due to its quadratic complexity. We leverage efficient approximation methods in ES to propose three new CM
 lability of depth images or models of a target to be operated on. Therefore, in this study, we achieve more accurate imitation le
novel Contrastive Learning loss for training Deep Neural Networks (DNN) that, when used together with an Imitation Learning l
pt the learned movements to the task scenes with different obstacles, which limits their adaptability since obstacles are everyw
ve to move the robot to accomplish the task. The GNN-policy has been trained as a node-classification approach by looking at
tained in the task demonstrations. The segmentation method was evaluated using a peg-in-hole task and bottle-lid opening task
ew skill is obtained to be used for other similar tasks. However, it can also be dangerous in terms of task execution safety. This
ount of unsupervised data, easily accessible, while using few-shot LfD to train the second, leveraging the benefits of learning s
probabilistic uncertainties in the form of a deterministic point cloud map. A neural network sampler encodes it into a latent embe
Barrier Transformer (ConBaT), an approach that learns safe behaviors from demonstrations in a self-supervised fashion. ConB
dently and incrementally. However, obtaining hand-engineered rules, such as PDDL descriptions is hard, especially for unseen
ty compensation is relatively slow and mechanically complex. The varying payload causes an unintended joint movement, i.e. t
, our method significantly reduces training time by 50% as compared to traditional reinforcement learning methods. WayEx obta
ge. Instead, we show how robots can learn several variations of a surgical technique from a smaller number of surgical demons
re expressive, it is able to fit a wide range of human demonstrations well, also without a non-linear forcing term. We exploit this
in-the-wild learning framework, we show that with only 3 minutes of the teleoperated demonstrations, augmented by diverse a
uses a simple and more realistic ranking of groups of demonstrations, without requiring pairwise comparisons over all demonstr
ting the user to sketch out demonstration trajectories on 2D images of the scene, these are then synthesised as a generative m
ncorporates additional properties that enable learning multimodal, distributional examples. We show that our method is able to l
o ensure global stability throughout the learning process. We validate our approach through extensive simulations and deploy th
s, which yields an optimization problem that can be tackled by sum-of-squares techniques. Finally, our approach can handle ob
ses the actual shifts in the test-time distribution. We take a different approach, where we make use of a single demonstration (a
ntrol barrier functions, respectively. Our approach outperforms baseline DS learning techniques on the LASA handwriting datas
energy function that is consistent with the provided demonstrations. The model presented in this paper employs a simplistic ne
rgy function is learned to capture the demonstration preferences via constrained optimization algorithms. With the learned neur
maximizing policy divergence between task- and constraint-oriented policies to obtain a transferable constraint. Evaluating our m
ons and use them as a rough estimate for solving complex planning problems instead of as a strict solution. During the process
ral Difference Learning for Model Predictive Control (TD-MPC) with Implicit Q-Learning (IQL); and 2) we propose to use IQL-TD
to produce desired behaviors. To address this limitation, we introduce SLIM, a multi-critic learning approach for skill discovery
model of cutting mechanics, embedded in a simulation environment, can be used to rapidly train the system without needing la
kill chaining with offline reinforcement learning. As a result, SPRINT pre-training equips robots with a richer repertoire of skills t
sentations acquired through contrastive learning by the Teacher Encoder and the Student Encoder to collaboratively estimate t
a quadrupedal robot and the results show that the controller is able to reproduce highly dynamic behaviors including sprinting, ju
based RL using distillation. Specifically, TWIST leverages state observations as readily accessible, privileged information comm
of the evader’s behavior, and the strength of the modeling assumptions in the fully-observable policy. We deploy our polic
tion pairs in the replay buffer equally when trying to learn from the replay buffer. In this paper, we propose a simple yet effective
e robots while harnessing their redundancy. In this work, we consider the snake robot as a modular robot and formulate the con
riments on CartPole, Ant, and Humanoid environments, along with real-world deployment on the Unitree Go1 quadruped robot.
of data augmentation on MARL algorithms. Guided by these insights, we present AdaptAUG, an adaptive framework designed t
while maintaining sample efficiency and provide the user the choice of picking a network architecture that satisfies their inferenc
with more training where prior work saturates in performance. We demonstrate that continued training with APRL results in a p
compliance. Also, we accelerate the torque-based-policy training process by pre-training the policy to remain upright by compen
 propose a Decentralized motor skill (DEMOS) learning algorithm to automatically discover motor groups that can be decoupled
nd users to reconfigure the data to cater to their own platforms. These tools and infrastructure alongside the dataset are useful f
he complex underwater environments can induce significant domain shifts across different scenarios, seriously deteriorating the
ntroduce a co-simulation method based on a multi-body dynamics model and discrete element analysis to obtain high-fidelity v
e, metric-space features directly from single-frame lidar scans, thus reducing the effects of occlusion and odometry error. Our a
strate the potential of our network in localization, we have designed a Geo-localization Node (GLN) that implements geographic
modeling factors that influence robotic traversability. The key contribution of this research is a two-stage traversability estimatio
eometric characteristics of the terrain. Second, a time allocation function based on offline reinforcement learning is proposed. Th
arios and propose a learning-based automatic trajectory generation model, which we call Multi-Agent TRajectory generation wit
sk training stage, and further show effective transferring in terms of both sample efficiency and performance.
-dimensional state space due to their non-compact representations. Moreover, they cannot collect high-quality training data thr
em as a set of keypoint embeddings and their relations. It then learns an action-conditioned forward model using contrastive es
h exploration, and a complex robotic manipulation task challenged by both exploration and long-horizons. Results show IGE-LL
fers from low sample efficiency. Integrating RNN with off-policy RL could enhance sample efficiency, but this often compromise
tivity graph to tackle nonlinear dynamics and complex configuration of the deformable object. Both simulated and real experime
n object segmentation. We zero-shot transfer from simulation to a real robot by using Segment Anything Model for promptable o
y of motor skill learning. We then show that CIPs can be used for plan execution in a zero-shot fashion by sequencing learned
rnatively, Reinforcement Learning (RL) based methods simply treat the target and its motion as part of the observation of the w
gradient algorithm to find a simplified dynamics model that explicitly maximizes task performance. Specifically, we parameteriz
idance. TEP utilizes reusable samples as the source, with the aim of learning reachability information across different task scen
s for this long-horizon task. We first sample demonstration from a set of heuristic policies and trajectories consisting of a set of r
actile representations and evaluate how each affects real-robot manipulation results after transfer. We conduct our experiments
x that provided details about action costs. Taking these representations as observations, we designed an attention-based netw
ntation on simulated scene point clouds. Then, we learn a SIM(3)-equivariant visuomotor policy on top of the pre-trained visual
essive attacks on the state space and the expected state-action values to emulate real-world perturbations such as sensor nois
tic framework for robot action space decomposition to control large 3D deformations of DLOs. In our approach, two deep reinfo
dback rewards. Specifically, BioSLAM provides a novel dual-memory mechanism for maintenance: 1) a dynamic memory to eff
efficient HRL-based framework to achieve this with a novel scoring method, named Predictive Neighbouring Space Scoring(PN
e achieved through an attract-repel reward term motivated by the Van der Waals force. We demonstrate the effectiveness of ou
sion detection in whole-body tactile sensing, the efficiency of the simulator is severely compromised. Thus a distributed training
nition head that generates global descriptors, and a metric localization head that predicts the 3-DoF transformation between the
grids to represent the scene effectively. Especially when dealing with object edges, our method can effectively encode the geom
ount of training samples, and the resulting policy, without fine-tuning (zero-shot), is unable to navigate efficiently in new scenari
onmental complexity and show great interpretability. Furthermore, we propose an environmental-complexity-based navigation m
ncoders, pre-trained on street images, to present novel applications for field-of-view expansion, single-agent topological explora
L research and its real-world deployment. Our evaluation of six tasks led to three key results: (a) the diverse nature of the tasks
diversity within a single policy by maximizing an information-theoretic objective. This diversity enriches each agent's experience
o provide spatial attention to the RL agent to boost the training efficiency of the pipeline. We evaluate the pipeline against varia
projection-based optimization can improve the convergence and learning performance. Unlike many existing works that require
explore regions of the state space equally. Past methods on encouraging symmetry for robotic tasks have studied this topic ma
ne-search-based methods grows strongly for computational complexity. Encouraged by the adaptability of reinforcement learnin
offline policy learning. However, data collected in real-world scenarios rarely contain dangerous cases (e.g., collisions), which m
cement learning, where the progress along the labyrinth's path serves as a reward signal. Additionally, we exploit the system's
ures. However, most existing topology-based exploration tasks utilize classical methods for planning, which are time-consuming
self-supervised manner and takes as an input a LiDAR pointcloud and RGB image. The network predicts a binary segmentation
motor learning, we develop an incremental training procedure that initially freezes some degrees of freedom and gradually relea
les the acquisition of complex motions like pronking and back-flipping, effectively from scratch. Additionally, our method simplifi
mber of evaluations required while tuning the controller parameters. We evaluate our method on a legged locomotion applicatio
ng existing ones, enabling incremental repertoire expansion and preserving original motion quality. We measured the transition
el motion generator that gives trajectories of parameterized human-like motions to the robot from multiple modalities of human i
earning (RL) method to generate physically realistic robot motions under language commands. LAGOON first leverages a pre-tr
implicit status of current joint condition based on past observation history, and 3) progressive curriculum learning to allow a sing
ng a multi-task training architecture capable of actively perceiving and overcoming two types of leg joint faults. The architecture
introduce a goal synthesis network and an observation selector to support combination of heterogeneous skills, each with its un
ed by the control policy. This allows agents a chance to gradually recover from disturbances before attempting to carry out their
ment learning with limited exploration for further finetuning the policy on more challenging terrains. By conducting comprehensive
traditional locomotion controls that separate stance foot control and swing foot trajectory, our innovative approach unifies them
veloping robot parkour on a small low-cost robot with imprecise actuation and a single front-facing depth camera for perception
s consumed by a risk metric to extract risk sensitive value estimates. These are integrated into Proximal Policy Optimization (PP
ally, we employ a distributional value function learned by quantile regression to model the aleatoric uncertainty of environments
sh a constrained optimization framework for the primary objective of enhancing task performance and the secondary objective
ocomotion skills characterized by motion parameters, such as velocity and body height, may not be adequate to navigate robot
ustness. To address this issue, the STL-guided framework generates optimal and safe recovery trajectories that simultaneously
oposed controller is validated in simulation and through an experiment on a bipedal robot platform, Gazelle, which confirms its e
e discrete-time barrier functions and a sampling-based approach to approximate sets that are maximally forward invariant. Las
of the stance foot. We then present the results of this controller on our robot Nadia, both in simulation and on hardware. These
time parametrizations. The proposed planner constructs the complete motion plan as a sequence of motion phases that can be
earned latent variables are disentangled and directly correspond to different gaits or speeds, such as moving forward, backward
ses of biped robots. We also announce a biped robot dataset consisting of LiDAR, inertial measurement unit (IMU), joint encod
y. The approach’s efficacy is demonstrated by a high payload-to-weight ratio verified with multiple tests, with a maximum ra

ithout prescribing a specific gait. The code is open-source, and a video synopsis available at https://youtu.be/1NKQ0rSV3jU.

t, lateral spinal flexion, which is able to modify the position of footholds, holds promise for optimizing balance during trotting. Th
nal motion imitation for legged motor skills often requires expert anima- tors, collaborative demonstrations, and/or expensive mo
esign of the mechanism is detailed, as well as the methodology to exploit the mechanism in the locomotion control framework.
of Ceres' gravity (0.029g), the robot can reliably jump to commanded positions up to 6m away. Our real-world experiments sho
training periods. In our study, we show that drawing inspiration from animal motor control allows us to effectively train a single
texts indeed offers optimal stability. The study also scrutinizes the influence of the robot's structural parameters and gait configu
hrough whole-body behaviors, and can reach far-away targets with gait emergence, enabling loco-pedipulation. By deploying ou
bstantial amounts of real interaction data since it solely focuses on training the world model, concurrently allowing for rapid polic
obot using data. The predicted state errors are then used to adjust the predicted future robot states generated by MPC. By suc
 al quadratic programs that share their optimal solutions according to a one-step communication delay and an agreement proto
mics, RL effectively overcomes the inherent limitations in reference motions imposed by modeling simplifications. We validate t
pical steady wheel gait of the linearized model is numerically generated, and its extension to the nonlinear model is discussed.
ody and limb dynamics, limb workspace, limb motion continuity, body tip-over stability, and power consumption. The performanc
echanism, the proposed controller can learn the control input without resetting the system at the end of each iteration. Simulatio

aces. We present novel algorithms for plane feature matching, tracking and registration for mapping within a factor graph frame
s through underactuated domains and hybrid impact events. To address this, we utilize the fundamental solution matrix that ch
ysical constraints, an advanced offline trajectory optimization method is employed to generate various acceleration trajectories,
uristic offers an effective way of providing candidate footholds for a legged contact planner. Additionally, we introduce a guiding
r demonstrate the effectiveness of the trained neural network to predict the disturbances and actuator failures based on historic
rmulation of the RL task, which we show to be vital for stair climbing. Furthermore, the methodology leverages an asymmetric a
rimless wheel to represent quadrupeds, using passive dynamic walking as a method, which has been well-demonstrated for ob
meras and is spatial-temporal consistent across the whole sequence. We apply our proposed framework to the popular nuScen
ding, we cover multi-floor buildings, gardens, urban and highway scenarios. Combining handheld and car-based data collection
he input and generate input-adaptive dynamic convolutions, thereby significantly enhancing the model's adaptability to complex
atasets of images. This representation can naturally handle image sequences and provides a general and granular feature repr

mera field-of-view overlap as similarity ground truth for learning. By optimizing image descriptors to align directly with graded s
thods. Therefore, there is a need to investigate the impact of DA on the generalization ability of VPR. To achieve this objective
ion based on transformer(SFFM) to enable full interaction between semantic and structural information. Furthermore, we introd
antic misalignment problem between Camera and LiDAR features, we design Cross-modal Interaction Transform(CIT) module
odometry (QVIO) to offer energy-efficient 3D motion tracking. In particular, we first quantize raw visual measurements in an intu
ation, we have specifically tailored a pose estimation and mapping algorithm that incorporates Semantic Label Filter, Dynamic
LP to encode the signed distance function (SDF) of the submap with an option of semantics. Learning the map using this struc
these features from stereo frames. Moreover, to prevent pixels violating photometric consistency assumption from negatively a
crucial for lidar depth loss. We use submapping to scale the system to large-scale environments captured over long trajectories
e localization at centimeter-level. Furthermore, we propose a novel map update approach by implementing high-quality data ass
ynamic objects removal for ground vehicles. According to the observation time difference between the object and the ground wh
We create separate implicit models for each detected object and train them dynamically and in parallel as new observations are
PVM. OctoMap-RT employs massively parallel ray-shooting using GPUs to generate occupied and free voxel grids and to upda
n on the boundary between traversable and non- traversable regions. Moreover, a measurement correction based on the pose e
ance-based segmentation masks during training. We show that this simple, yet effective, technique drastically outperforms rece
sion module (FSM). The first strategy involves building a novel GFN using a newly designed guide filter layer. The GFN interpre
geometric data augmentation, predictive world modeling, and an information-theoretic regularizer enable our model to predict a
our devised cross-metric knowledge distillation to narrow the performance gap between the teacher and student models, maint
bal localization with only a 3D LiDAR scan roughly aligned in the gravity direction and a 3D pre-built map. This method required
ormance on SemanticKITTI and validate it in real-world environment. Experimental evaluations show that our method achieves
eralizable visual-audio navigation. We meticulously design two auxiliary tasks for respectively accelerating learning representati
, 3DIML leverages a two-phase process. The first phase, InstanceMap, takes as input 2D segmentation masks of the image se
f non-standardized urban features. To achieve consistent ground features, the Motion Compensate IPM (MC-IPM) technique is
 in large-scale environments by loading local map information. Secondly, global localization according to Branch-and-Bound Sea
nd segmentation, energy point detection, and energy curve refinement are designed to extract DEC features from a 2D radargr
ents in pursuit of precise scaling and color details within mapping outcomes. However, the outcomes are subject to extrinsic cal
e selection scheme that detects uninformative directions in the point cloud registration and explicitly selects patches with compl
d to propagate the posterior probabilities of particles over the neighbor particle graph. The proposed method is capable of evalu
ly estimates the gravity direction and IMU state and helps perform global localization in 3- or 4-DoF for system initialization with
d opposing) distance matrix sequence matching method. We evaluate SPOT on a publicly available dataset with 6.7-7.6 km rou
mproves the match performance with the destination image. Experimental results on multiple datasets show that the proposed m
sly improving the overall SLAM performance. Through experiments across 3 different public datasets, we show that our approa
methods merely provide cross-view patch-matching results, which leads to unfeasible in integration with the SLAM system. To ad
a feature-based visual odometry learning system with an effective scale recovery strategy in this paper. Additionally, we propos
tary-analogous environment, show that UMF outperforms significantly previous baselines in those challenging aliased environm
Convexity (GNC). Additionally, we propose a robust Incremental Fixed Lag Smoother (rIFL) for efficient GNC-based optimization
related objects, the semantic-focused patch tokenizer exploits attention-based channel selection and spatial partition, which effi
e-to-frame data association. The LiDAR frame-to-frame and the inertial measurement unit (IMU) preintegration measurements
letion model for filling in the missing scene content in a geometrically consistent fashion to support pose registration. We evalua
mans can comprehend and interpret their surroundings for localization purposes. These modules jointly learn how to effectively
feature interaction among the neighboring images in the sequence. A hierarchical feature aggregation (HFA) module is further
extract these features, one for global and one for local, this strategy is suboptimal in terms of computational efficiency. In this p
ry computational costs with redundant data. These challenges are addressed in three folds in this paper: (1) A NeRF is designe
rages large uncurated sets of images through a multi-task learning framework. With JIST we also introduce SeqGeM, an aggre
oaches. Our CCL constructs a contrastive feature pool and utilizes contrastive loss to train more transferable representations o
gh Gated Recurrent Units, which also considers semantic insights and robot height from a Vision Transformer autoencoder app
ptured in nearby locations in Euclidean space and their embeddings in feature space. Experimental results on the large-scale p
address this, we train on real colonoscopies a deep global descriptor achieving high recall with significant changes in the scene
netic field is measured using a three-axis magnetic field sensor embedded in the magnetic device to be localized. The method is
selection process to identify the Top-textit{K} most similar candidates. In the fine stage, we estimate pairwise overlap among th
ine heat point clouds, by which the size of the original LiDAR map is compressed. Then, an end-to-end online pose regression
he network by points sampled along the measured beams, and our loss is designed to learn a valid distance field. Additionally,
scene contexts, and the associated multi-frame scans could also improve the performance of underground place recognition. W
method is designed to select the valid G2S prediction. The selected prediction is then fused with the SLAM measurement by so
ances computational efficiency by diversifying input features, reducing redundancy, and expanding model capacity. EffLoc exce
, and dynamic vehicle removal. Unlike previous semantic-aided approaches, the proposed method can improve localization acc
ation between the query and database features. It does not rely on or affect APR network architecture, which is flexible and co
he images. In this paper, we propose to utilize these minimal available labels (.i.e, poses) to learn the underlying 3D geometry o
iveness and symmetry. Next, the orientation filtering module estimates the unknown 3D orientation of the query image and res
g Compressor} (GPC), in which the compression is reversible, a crucial requirement for the downstream 6DoF pose estimation.
d SNN that trades biological realism for efficiency. VPRTempo employs a temporal code that determines the timing of a single s
mple way to solve normal equation on manifold. Both simulations and real world data experiments demonstrate the proposed on
ts) without any re-training or fine-tuning. We demonstrate that general-purpose feature representations derived from off-the-she
te binary features, and an improved spatial filtering step that better preserves performance when the databases are tuned for li
 d maps (with limited views) and optimally fuses IMU, monocular images, and synthetically rendered images within an efficient f
e Invariant Extended Kalman Filter (InEKF) to fuse IMU, odometer, and camera measurements for consistent state estimation a
orms, including flying UAV and driving vehicle. The experimental results support the observability analysis and show increased
he proposed approach fuses geometric information from miniaturized Time-of-Flight~(ToF) sensors and semantic cues. The se
ansmit antenna are placed on top of a slowly moving service robot, and 88 trajectories are recorded. Each trajectory includes 20
sing from various special motions. Through numerical simulations, we validate our key observability findings and examine their
centimeter-level global positioning results with scale, which is effortlessly integrated and favorable for low-cost robot system de
estimator to provide dead reckoning localization for a small autonomous surface vehicle. The proposed method has only one in
ates at a high rate for online control. The proposed method is lightweight and can run on a computationally constrained micro-a
r self-calibration, robustness in static conditions, global navigation satellite system (GNSS) outliers and shifts, and robustness t
tric noise model crafted specifically for non-Gaussian multipath disturbances. Additionally, we present a novel Levenberg-Marq
ment regularities akin to the constraints imposed by floor plans. This work investigates the possibility of leveraging these radio m
age descriptions and extracting correspondences based on conceptual similarity with image observations using a Vision Langu
ward and effective approach is introduced for updating particles when dealing with sharp-peaked likelihood functions obtained
and glass. In contrast to this, commonly used map representations, such as occupancy grid maps or point clouds, lack these fe
n. Furthermore, to enhance spin inference early in the ball's flight, where limited observations are available, we integrate human
des a more reliable translation of the sensor, compared to the errors linked to IMUs due to noise and biases. Its short-term acc
nments. We compare several odometry estimation methods, from direct integration of Doppler/IMU data and Kalman filter senso
e propose a hierarchical one-shot localization algorithm called Outram that leverages substructures of 3D scene graphs for loca
es Jacobian (FEJ) methodology was used to properly fix state linearization points in the optimization-based VINS, which seems
learning approach for a handover task. The validity of the approach is demonstrated through experiments with participants on
pproach termed NaviFormer. This approach handles navigation tasks based on sequence modeling to obtain the data-driven ab
gthened by updating a solitary inducing point derived from newly samples, in conjunction with the learned hyperparameters. In
y a recurrent state-space model (RSSM) that accurately predicts the transformation of the latent representation over time. Exte
atory and end-to-end training properties of RL. The resulting policy effectively manages both short-term decisions through the M
ings can parameterize discrete and distinct modes representing high-level driving behaviors. We propose to perform fully react
erse range of simulated deformable objects so that the policy can adjust actions based on different object parameters. At the tim
o a full system for box loco-manipulation to achieve the task of moving boxes from one table to another with a variety of sizes,
onian structure in the dynamics model to improve data efficiency. We design an observation-space loss that relates motion pred
ns due to computational constraints. And even with a perfect model and enough compute, MPC can get stuck in bad local optim
ng approach that aims to balance robustness with adaptability. We train a System Identification Transformer (SIT) and an Adap
We posit that a significant challenge to widespread adoption of robotic RL, as well as further development of robotic RL method
s, and use it to reduce the input space when learning the dynamics. We then learn the unknown constraint manifold by approxim
this paper, we develop a system for leveraging large-scale human video datasets in robotic offline RL, based entirely on learnin
g them to further expand their reach and increase mission efficiency. In this work, we used a robophysical rover model and per
on is guaranteed, and thus has minimum limitation on exploration. Similar to Recovery Reinforcement Learning, we utilize a de
orithms, hierarchical best-first search (HBFS) and parallel Monte Carlo tree search for multi-primitive rearrangement (PMMR) to
s both visual inputs from the scene, and goal-specific language prompts to plan intermediate reorientation poses. Specifically, t
w-level execution in these systems. When deployed on a mobile manipulator equipped with an eye-in-hand RGBD camera, For
ting the objects without prior knowledge of their physical properties. To accelerate on-hardware training, a focused training stra
ossible grasps to accomplish the task without considering such topological information is very inefficient, as many grasps will n
everage the property of real-world scanned point cloud that, the point cloud becomes less noisy when the camera is closer to th
le or pick-and-place rearrangement actions in addition to pushes; we show that by running it as a depth-first search improves p

gnificantly boosts (by up to 60%) lane detection and topology prediction on current state-of-the-art online map prediction metho
jections. Secondly, we concatenate the learnable features with a Fourier feature positional encoding. The combined features ar
el map and decoding signed distance fields through a shallow MLP. We advocate for a robot-centric local map to boost model t
n a hash-encoded NeRF which significantly boosts up the pose optimization process. To account for the noisy image gradient c
insightfully integrated for effective supervision of QuaidricsNet. Besides, a novel pattern-comprehensive dataset with quadrics s
equently, we propose the Height Stack Test, Ground Layer Test, and Surrounding Point Test methods to precisely and efficient
y initialization and makes scene geometry easier to learn. Besides, we present a coverage-maximizing keyframe selection strate
. We derive a probabilistic formulation that propagates shape and pose uncertainty through two novel loss functions. Unlike cur
h with segments as nodes, where edges are formed by a) associating segment-level descriptors between pairs of consecutive
ogical graph, and propose a learnable map-reading method, which parses a variety of 2D maps into SAMs. Map-reading extract
division and a novel pyramid-like subspace division, is proposed to propagate particles and update the map efficiently with the
and non-moving ones. Additionally, we propose to fuse these predictions in a probabilistic representation of the dynamic enviro
of such a set of lasers produces an array of points where each point's neighbor is known due to the known firing pattern of the
-based point cloud and a sequence of panorama images with initial coarse camera poses, our objective is to jointly optimize the
obabilistic way. To reduce the redundant computation for simple planar objects, extracted facets from a Gaussian Mixture Mode
ovide theoretical analyses of the accuracy of the depth image projection and time complexity of occupancy updates. Furthermo
nt is employed for effective online DOR. Then, in scan refresh, region-wise ground plane estimation (R-GPE) is proposed for inc
ur work can be used on top of any SLAM/GNSS estimate to improve and refine the initial trajectory. We conducted different exp

etical models are derived to facilitate the control of the robot and the understanding of its nonlinear behavior. The impact of the t
tional diving robot and a suction sampling system. The proposed concept consists of a series of rigid links connected by angled
for underwater manipulation that bridges the gap between simulation and full field trials. The underWater ArmVehicle Emulator
iniaturization of amphibious floats. The dynamic characteristics of the small PVBS are linear fitted, and modeled. The originally
rates propulsion and yawsteering for fast course corrections. Two independent pectoral fins provide pitch and roll control. We e
Lagrangian method, and the hydrodynamic force is calculated by the quasisteady flow model. Third, extensive experiments ar
bility to collect replicas without returning to the boat or to the shore. We also defined a protocol for sediment collection and exten
ose a new expression called Zero Moment Point (ZMP) bias that can indirectly represent the leg-terrain interaction force, remo
tion in 3D cluttered environments consisting of observed static and dynamic obstacles. The problem is modeled using path-opti
encode global context. The network is trained in a supervised fashion on the forward-looking underwater dataset, FLSea. Eval
ction to reliably compute 6D pose estimates in different underwater scenarios. We test our proposal with 4 objects with symmet
water application scenarios. Consequently, multibeam imaging sonars emerge as the preferred choice for perception sensors.
ncoder-based framework (CVAE-SM), which is capable of generating diverse latent codes for improved uncertainty quantificatio
me densities of the underwater environment. We introduce a logistic regression model to separate water from the scene and ap
we demonstrate that robust deep visual models can be developed based on CaveSeg for fast semantic scene parsing of underw
dentify different benthic habitats. Combining these maps allows us to determine the level of acoustic activity within each habita
ch each octree incorporates a transformer-based deep entropy and an arithmetic encoder. Besides, we introduce the untied cro
and which dynamics are compatible with a fair estimation of the cable shape through the proposed models.
model in numerical simulations, comparing the hybrid model's performance to that of a model for fully submerged UVs and to t
nts in regular and irregular waves reveal the effectiveness of the algorithm. The algorithm ensures a healthy initial attitude and in
ey are extremely power efficient but not very maneuverable. Gliders are designed for long-range missions that do not require pr
reference trajectory) even in the presence of model inaccuracies and perturbations. To address this challenge, we present a ro
yzed and optimized propeller-leg by simulation combined with the open water test, the propeller-leg's thrust in the water before
ras and lenses mounted in water-tight dome enclosures. The proposed camera models treat each layer of effects as a ‘sphe
the learning-based methods work well for coarse alignment, and are better at recovering rough transforms consistently at high
Angler: a software framework for developing localization, control, and decision-making algorithms with support for sim-to-real t
es by the number of group options, not agents. We demonstrated the effectiveness of this approach using a hypothetical `explo

ary Algorithm (EA) and a decentralized Monte Carlo Tree Search (MCTS) method. Both methods incorporate the charge availa
g ASVs as well as with static obstacles and current flows. We evaluate the performance of the proposed Distributional RL base
mework suitable for large-scale, feature-sparse environments. Our strategy leverages SLAM graph and occupancy map data ob
d capacities, and superior concealment. Specifically, the submarine mode makes Sea-U-Foil unique among previous studies. In
etworks (GNNs) to both classify acquired sonar views and reinforcement learning to choose the next best view to capture base
nvenient mechanical and electrical connections for end-effectors. To compensate for the disturbances that are present in the co
of scans is not only based on navigation, but based on the graph itself. Scans formed from scan-lines can be optimized by onlin
nce process by seamlessly transitioning the target's state between the different observability stages, intrinsic to any dynamic do
semantic segmentation network is used to detect the water region in an image, and back-projection to determine the water clea
ng water conditions. Both AUV and DS utilize cameras to determine the water background color in order to calculate the desire
gions from LiDAR measurements. A safe collision-free path for the ASV is generated within the navigable regions using model p
and inertial sensing, the model is validated regarding the accurate estimation of the refractive index and its benefits for robust o
esults in significant reductions in the orientation error and also in recovery of the true yaw orientation with respect to the magne
exhibit an accurate initial state for successful state estimation. Additionally, for wider application, we extend the sensor data of
gmentation. Experimental results are provided based on real data collected from a dam shaft scenario.
own the complex planning problem into a series of manageable subproblems and solve them sequentially. Our approach signif
ethered drill configuration for drilling operations. The novel gripper mechanism enables drilling on large planar surfaces and eve
osed strategy’s accuracy, efficiency, and robustness. This study makes a substantial contribution to present this novel con
r-of-mass (CoM) position changes when cargo is loaded, leading to undesired attitudinal motion due to unwanted torque gener
line utilizing tactile feedback for force tracking via a hybrid motion-force controller and a method for wall texture detection durin
alize the jumping take-off of a meter-scale FWAV from flat ground. To lower the mechanical complexity, we eliminate the desig
d. We design a perching/unperching module composed of a single servomotor and a magnet, which is then mounted on the tilt
d load due to the hybrid dynamics that vary with the cables' tension conditions. This paper presents the first aerial transportation
e method is tested in simulation and is experimentally validated with flight tests. Comparisons between simulation and experime
ed by the maximum descending speed, it is necessary to jointly analyze the flow field and the energy consumption during desce
control by the internal force induced in the manipulated object by non-vertical cables. Furthermore, we formally study the sensi
ther. Winding and unwinding the tether allows the pod to move within the canopy, and activating the propellers allows the tethe
plification effect of the floating base is quantitatively analyzed by virtue of the aerial manipulator kinematics. Intuitively, it is found
s to predict the aerodynamical forces experienced by the aircraft due to its own motion in real-time and, therefore, will be useful
textbooks, is a novel figure that generalizes fixed-wing doghouse plots to tail-sitting vehicles. This flight envelope figure sugges
l study of ground proximity effects for EAD propulsors, both individually and as quad-thruster arrays. It focuses on multi-stage d
ed additional thrust to maintain altitude and airspeed while turning, or a UAV with different curvature bounds and costs for the si
 -rural mix logistics case study in Southern England. Results show increasingly risk averse paths with decreasing future UAS op
nerated through multimodal search in a dynamic time-domain. The velocity errors relative to the moving surface are alleviated.
tation about its body Z axis as well. In addition, it has the unique feature of a coaxial motor configuration that allows for a pitchin
ten occur due to perturbations caused by modeling uncertainties, sensor noise, and environmental uncertainties. Even small at
sition approach is designed to distribute the output to the torque and thrust commands for each subplane, with no need for com
nifold. Furthermore, our method facilitates the application of usual techniques designed for particle filters in Euclidean spaces,
e reality gap, we incorporate a learned residual model and an adaptive method that can compensate for modeling errors in thru
y in the rolling mode to take photos of the surroundings at different tilting and panning angles to construct a panoramic image. B
arameters. The method consists of applying multi-objective constraint-based optimisation to a multi-body winged drone with traj
es. Moreover, to strengthen the adaptability under significant disturbances, a novel switching strategy is introduced into the con
h the lesser cost of rotation as predicted by a dynamical model of the UAV derived from first principles. This problem is relevan

s, creating unsafe operating conditions and potentially damaging the platform. In order to ensure safe pushing on these surface
een the UAV and the load. This lightweight construction is reliable and robust while not relying on active forces to maintain the g
mplish various aerial interaction tasks without manual adjustment. The proposed energy-saved trajectory planning approach in
ontrast to standard VRP and capacitated VRP, we must additionally consider waiting times for payloads lifted cooperatively by
quadrotor-based modules. Each IdentiQuad can be operated alone like a standard quadrotor, but at the same time allows task-
ations of the platform in the joint space and task space respectively. The proposed underactuated platform is found to be more e
insertion costs, and an adaptive speed planning framework. With that, the optimal path is searched in a graph using a discrete
magnetic coupling effect provides a spring-like characteristic that can be utilized to implement a series elastic actuation enablin
tem generated 31.3 N/cm2 sheer stress and a total holding torque of 7.1 N·m while consuming only 2.5 mW/cm2 at 500 V. W
ldable wings are measured by the F/T sensor. The experimental results indicated that its aerodynamic performance is basically
ristle-bot units equipped with body-embedded permanent magnets actuated via torque imposed by an external magnetic field. A
Coils (C-MAC) mechanism, is specifically designed to be manufactured through a monolithic 3D print. This approach enables to
up to ∼4◦). This is the first concept of growing robots to enact the robot’s modularity and Transferability for future dep
method to determine backdrive torque is also explained. Finally, experimental results quantifying the dynamic performance of t
evice, the control experiment of the motor is carried out. The influence of gravity compensation device on motor driving torque a
eriments showed satisfactory results for positional accuracy and repeatability with adaptability over a wide range of curvatures.
grasping stability and practicality than existing grippers in order-picking tasks. First, the design of the mechanism and model co
modules during task operation. We demonstrated the effectiveness of the proposed method by achieving assembly experiments
uated by comparing the motion properties of the proposed robot with the robot without elastic rings and synchronism mechanis
bot can carry objects and transport them to a predetermined destination. The design and operational modes of the robot were e
he thigh frame to minimize the distal mass. High payload timing belts and unique linkage mechanisms are utilized in the transm
n appropriate solution for towing heavy floating objects because of its ability to provide greater pulling force. The main part of th
distortion experiment results demonstrated no data distortion, showing an average error of up to 0.40% for detecting obstacles e
torque controller and the high-level compliance controller, confirm the effectiveness of the proposed hardware-software remoti
icularly in regions with multiple colinear crease lines. Furthermore, the design exhibits one degree of freedom and features pivo
e resonance frequencies for the principal axes of the beam, the vibrational orbit of the tip of the beam can be controlled in the s
of two coaxially nested cavities, where the inner and outer cavities extend synchronously by inversion and eversion of the film r
ndulum (MIP), and other modes. In this work, we derive a dynamic model and formulate a nonlinear model predictive controller
¼L), the optimized strip structures can enhance the stability of eSKT and facilitate the penetration of nutrients and oxygen, lead
 ved therapist-patient interaction, safety enforcement, and performance quantification. The whole body control of the two exoske
a tracking method of in-plane ultrasound catheter's tip and detection of vascular boundaries utilizing intensity-level iterative Hou
n, the main challenge is that the motion model between the input of the magnetic field and the output velocity of the target obje
mapping between actuation space and shape configuration and tip force estimation model are respectively established for furthe
ntegration of robotic systems, creating a high cognitive load on the operator and prolonging the surgery time. Moreover, misalig
VIC, and PVI-HVC. The different actuation of each had a direct effect on the global tibia kinematics, but the average deviation
surgical task autonomously, human supervision is critical to monitor the robot’s execution and, if necessary, stop the robo
lop a track-based stereoscopic endoscopic robot (TSER) which is equipped with four tracks positioned at the corners of its bod
port both training and research in the sector. In the past ten years, a great amount of research on the dVRK has been carried o
n of compliant controller parameters during ultrasound procedures. This optimisation is formulated as a quadratic programming
validations demonstrate the system's precision and capability in simulated surgical tasks, highlighting its potential for advanced
esults in improved interpretation and disease diagnosis. However, the interpretability of 3D ultrasound reconstruction of abdom
sitions in a standard 12mm-radius spherical eyeball model, achieved by linking floating instruments with on-the-retinal objects b
es rather than the skin surface. To this end, the self-organization mapping is successively used twice to unify the input point clo
tral valve to be centrally positioned is estimated based on this scan. Second, the probe is rotated once in the yaw direction whil
racy and robustness of small blood vessels. It incorporates two encoders to exploit the maximum potential of the Doppler signa
ectory planning was devised. In the freehand US 3D reconstruction and registration experiments, the registration errors for hum
d slow response speed in the drilling and milling status. Aiming at these problems, an elliptical torus-based 6-A fiber optic F/M s
a-operative motion errorswhen driving both hand-held and robotic-assisted surgical tools may result in significant stress on the
the vertical force. Second, fuzzy control minimizes the total error of both the vertical and feed force by adaptively adjusting the
urgeons perform surgical procedures while looking at the eyepieces of a surgical microscope and the same applies for robotic m
ntial to significantly enhance surgical safety, streamline efficiency, and alleviate the physical demands on the surgeon. The pap
g high-touch areas in the surroundings, enabling the disinfection robot to achieve superior disinfection efficiency. In addition, we
peratively visualizes the hidden lymph nodes by projecting the corresponding 3D pre-operative model, and virtual fixtures are us
n-robot interaction forces during surgical navigation. We propose a hybrid control method that leverages both force and vision i
SYS is carried out, and the shape of one of the driving wires, when bending, is fitted by a two-segment polynomial curve. Then,
amping force, and provide real-time force feedback. A leader-follower bidirectional force feedback control strategy was propose
surgical field. However, conventional 1R1T RCM robots typically employ complex structures with numerous joints consisting of
del of the proposed mechanical amplifier, and the performance was validated through finite element analysis and prototype fabr
reprogrammable magnetic continuum robot (RMCR) enabled by magnetic reprogramming modules (MRM). Actuated by shape
he fluid environment, but also realizes the intestinal anchoring and targeted therapeutic delivery functions against the constant
tly bolsters the flexible robot's ability to adaptively track mobile objects and resist external interference. Experimental results un
we introduce a novel method capable of directly estimating a CDM's shape from FBG sensor wavelengths using a deep neural
gle the inequalities and provide proof for the unique existence. It transforms the interdependent inequalities into independent bo
numerical solution for calculating its electromagnetic force and used it to guide a hand-held control panel for navigating the robo
the opacity of the human body. In this paper, we provide new methods and tools that enable the 3D control of a magnetic swim
hybrid robotic endoscope control system based on fusion model of natural language processing (NLP) and modified YOLO-V8
alongside unsupervised classification using K-means clustering. The SE base-learners comprised Support Vector Machine (SV
catheters, and guidewires within them. This work proposed a solution using an adaptation of a state-of-the-art machine learnin
cy of the robot is related to the position, distance, and insertion angle. The result shows that the performance is better when wo
. To address this challenge, this paper introduces a novel environmental motion analysis system that employs a grid layout with
fety feature, that is, minimizing the contact force between the surgical instrument and sclera surface for avoiding tissue damage
t learning and damping control approach to regulate the motion scaling ratio for different operations during offline training. This
’s distinctive flexibility originates from a conventional rotating joint design and the incision pattern in the flexible material. In
ling were established to investigate the parameter design. An easier insertion procedure was introduced and demonstrated in e
stable phototherapy of Helicobacter pylori. Each integrated unit is equipped with wireless charging capabilities, producing an o
mented. When the catheter force is too large, shrinking the grasping distance of FTG can avoid the surgical risk. This method p
hape and force along highly deformable flexible robots by using sparse strain measurement from a single-core FBG fiber. A thin
al tool, a vacuum grasping tool, two 6-DOF manipulators, and an autonomous control system. The letter introduces an ontology
ed-order trajectory planning module that depends on just two parameters. By performing 25 experiments on custom-designed s
ctor that uses standard microsurgery instruments. An initial teleoperation validation was carried out by eleven participants, who
usting the pose and position of the robot, the accessibility of such systems is significantly optimized. At the same time, the acce
Results from a human-in-the-loop case study with eleven operators indicate a 47% reduction in guidewire navigation times, wh
ator to provide an efficient (
ding of CAST is actuated by pushing/pulling distance, and the maximum diameter is only 1.7mm with high dexterity after moun
ewire manipulation. Experimental results demonstrated that the proposed trial-and-error and centerline-guided methods succes
el. With this new design, the new device can induce bending motions at its tip for active steerability to reach desired aneurysm

mpression. Real-time finite element simulation allows multiple control inputs along the length of the needle with full three-degree
pproach aids in dynamically estimating parameters such as inertia using regression analyses on data from real robots. The exc
es of high accuracy, high speed, and high success rate. It can endow magnetic microrobots with flexible material selection, arbi
erimental results. A launcher mechanism was designed and integrated with the NiTi actuator, and this mechanism was used to
imode locomotion, micromanipulation, medical imaging, and biosensing. The proposed microrobot is essentially Pickering bubb
per actions, returning to a default state (normally open or closed) when the magnetic field diminishes. This work advances magn
hode by redox reactions. The EOP generates a self-propulsive velocity due to counteracting forces of electroosmotic flows prod
aneuverability and opens up new challenges for modeling and control. We detail the design methodology and fabrication, follow
eloped a micromotor that consists of three geometrically heterogeneous helices that can be selectively driven within a specific n
l specimens. Using two ex- ample experimental cases as proof of concept, we show that our automated platform is more precis
control algorithms. In this paper, a micro-assembly task of miniature thin membrane sensors is considered. For human demons
on each arm. With this setup, an optimized motion trajectory planning under standard knot-tying protocol is also presented in o
ow efficiency. This paper aims to address the challenge of acquiring dense depth information during robotic cell micromanipulat
and in situ Gaussian blur compensation algorithms, we achieved rapid automatic focusing and illumination shadow compensati
coordinate system of the robot is calculated. And the cutting trajectory is generated and optimized. Finally, the end-effector is c
ximately, and 20° around the z-axis approximately. The micromanipulator has weights of 0.8 g, dimensions of 16 mm × 16 m
gh SCM and holds a top speed of 27.4 cm/s. In this work, we propose a novel strategy for designing the insect-scale robot with
ure robots in their respective targeted workspace to achieve functionalities. It reports that the existing miniature robots with a m
wever, previous CNN methods have struggled to accurately segment tiny objects, requiring staining or fluorescence techniques
its intrinsic and hand-eye parameters through a two-step algorithm. The calibration data is then utilized to execute a SIFT matc
R) force sensor. The imaging technique was used to detect the tip of the glass capillary and the contact to the cell membrane, w
d by automatically merging 50 AFM elementary topography patches of dimension 0.9 μm × 1.2 μm based on feature match
icromanipulator as in existing robotic systems, this paper presents an automated, compact three-dimensional positioning stage
long and curved shape and its uneven width. This paper presents automated techniques to measure sperm morphology param
 ay for using PAM for robot-assisted microsurgery. This work proposes a fast PAM scheme with robot-controlled microtrajectory
method to rotate and transport mouse oocytes, which utilizes acoustic waves and micropipette to generate acoustic radiation fo
latform, extendable to match the dimensions required for in vivo applications, marking a noteworthy advancement. Our method
oduce Sim-Suction-Pointnet to generate robust 6D suction grasp poses by learning point-wise affordances from the Sim-Suctio
e sensor range are unobserved, but the attributes of objects that fall within sensor view can be resolved via repeated observatio
vironments by adaptively adjusting searching dimension based on environment complexity. Additionally, we propose a real-time
compose exploration, navigation, and manipulation skills. To achieve this, we design an abstract high-level action space around
e friction coefficients, which provides robustness to frictional uncertainty and other force disturbances while also substantially re
e adjustment of several, independent environment parameters is usually not considered in the AMPC formulation. In this work,
rasping pose observation state for reinforcement learning. Moreover, on-the-fly fusing the grasping poses enables a direct asse
model predictive controller (NMPC) to control the legged robot. On the robotics arm side, we adopt an impedance controller to re
pproach, ActPerMoMa, generates robot paths in a receding horizon fashion by sampling paths and computing path-wise utilitie
sample-efficient and adaptive learning process, capable of accurately inferring model dynamics in real-time even in operating
rmer to learn a latent dynamics model to predict material deformations given a grasp action. We design a novel action samplin
ed racing. To this end, a Model Predictive Contouring Control (MPCC) (a specific formulation of the standard Model Predictive C
om motion and can even estimate contact interactions without requiring ground truth force data during training. This enables th
daptive control, we employ an adaptive system identification framework and demonstrate its efficacy in enhancing the performa
do-inertia matrix. Also, the resursive version endows the estimator with the capability to deal with time-variation of parameters b
n dynamics during manipulation. To demonstrate feasibility of output-based control without state estimation, direct input-output
a sequence of the future system states. An LSTM encoder-decoder with an attention mechanism is used to capture the dynamic
ving a convex optimization problem based on the history of past robot motion, the GPR augmented motion model generalizes to
ual metrics. The proposed metric is derived from a thorough analysis of eleven existing human-likeness criteria and is applicabl
ntroller loop to respond to contact pressure. We evaluated the effectiveness of MiBOT in a controlled study and assessed subj
methods as a seamless plug-in, and its consistent performance improvement underscores its efficacy. Furthermore, a new me
ntrolled in-the-field recordings by robotic agents or smarthome monitoring systems. We present two SynthAct datasets: AMARV
ons, grasping effect read from tactile images and grasping strategies for different positions. The characteristics were then utilize
based whole-body controller and a multi-task optimizer for behavioral adaptation. On a screwdriving scenario, we demonstrate h
namics and interaction with various devices. We develop a new algorithm using low-dimensional representation and hierarchica
put is an image, a lightweight backbone named multi-stage and global feature enhanced network is designed for 2D encoding, w
oving horizon estimator (MHE) algorithm was used to track experimental biosignals. Two participants were equipped with EMG
ehensive theoretical analysis for this model, focusing on two fundamental characteristics of LJ-based continuum robots: shape
t growing robot for early breast cancer detection. It was demonstrated that for most regions of operation experimental results co
rajectory. Through comparative experimental evaluation, we found that the new control method outperforms its conventional co
that this constitutes a more general methodology than PCC. Simulations and experiments on the real robot attest to the superi
Convex Iteration for Distance-Geometric Inverse Kinematics for Continuum Robots), an algorithm that solves these nonconvex
omatically tune the hyperparameters and generate the controller. In the process, we extend AutoMPC’s capabilities to han
based labeling algorithm, our approach identifies the true order of the detected key points and then reconstructs the shape usin
ol process. The proposed controller can effectively handle the soft robot's complex nonlinear uncertain dynamics and external
ence. After presenting a 3D polynomial description of soft robots, we study their intrinsic properties, e.g., input-to-state stability,
proposed pressure control algorithm could push the closed-loop pressure control to 19 Hz cutoff frequency. A combination neur
e. This paper studies the state estimation of a pneumatically actuated soft robot using the measured forces and torques at its b
 he goal position. The proposed controller was tested on a modular soft robotic arm, I-Support, by tossing four objects of differe
niques from 3D character animation to skin our simulation with a high-resolution surface mesh, pointing a way toward Mixed Re
. In this paper, we propose a proxy approach, as exists for humanoid robotics. This proxy is a simplified model of the robot that
arget position, followed by a closed-loop proportional controller that minimises the system’s steady-state error. Feedback is
To improve the proprioception pipeline's training efficiency and model prediction consistency, we present POE-M. POE-M pred
n systems with symmetry, we offer a method that can be applied broadly to robots with first-order, low-pass actuator dynamics,
then propose a controller that compensates for the elasticity, and we prove its stability with Lyapunov methods without relying
ces and grip strength, which can be modeled by an exponential fit. This investigation serves as a nexus between biomechanica
ction weights suggest that the increase in walking speed attributes to a reduction of foot sliding penalty weight and weight increa
an optimization-based control strategy to minimize the risk of unexpected detachment. The robot prototype has demonstrated v
ad carrying capacity of up to 2.8 kg for vertical climbing and 0.5 kg for inverted climbing. The robot also demonstrated the ability
y, suboptimal control parameters, or insufficient power. We propose a solution that asks the thrusters to do less high bandwidth
g these two approaches, we take a step towards geometric motion planning for high-dimensional systems. We test our method
resent a hierarchical online hybrid primitive-based planner for autonomous navigation with wheeled-legged robots. The framew
o ambiguity and sparsity in the reward. In this paper, we propose a method to directly learn a neural network controller to satisfy
ork can be used to build an analogy of a shortest path tree, thereby enabling optimal concatenation of sub-problem solutions in
mon representation for path planning and also for finding multiple distinct paths. Yet, challenging environments with multiple na
ution and guide the solution to converge to fit it. In the proposed method, the target mode is roughly estimated using a modified
directions. In order to accomplish that, we use an entropy-satisfying upwind scheme that converges to the true visibility polygo
splines that leverages the symmetry and similarity of clothoids. These techniques are demonstrated with numerical examples. T
ralized planners. Our planners compute trajectories that allow the robots to reach precise SE(2) goal poses. The effectiveness
r paper tackles this challenge, integrating ergodic search methods with energy-aware coverage. We trade off battery usage an
ions, we present Reconstructed Bi-directional Informed RRT* (RBI-RRT*) path planning algorithm. The algorithm acts as RRT-
el elastic obstacles as external potential fields and formulate the robot statics in the potential field as the extremal trajectory of a
d, and the overall distance traveled by the robots.
o combine the strengths from both sides. We define point cloud representations of free states. We perform Neural Focus, which
ical subroutines, such as forward kinematics and collision checking. We demonstrate our approach over a diverse set of challe
the efficacy of our methods at persistently gathering information while avoiding risk, robust to randomly-generated environmen
ry representations and a lack of consideration for time optimization. In response, we introduce a novel trajectory optimization fr
algorithm are studied in simulation.
igh fidelity. Finally, optimization via gradient descent was attempted but found to fail because the search space is far more non
ion plans, potentially sub-optimal, within a user defined time bound. This framework has been demonstrated to be effective in a
ned policy's critic network is then used to evaluate the quality of dynamically feasible velocities generated from a novel context-a
es, they are much harder to adopt without specialized training datasets. We propose EDMP, an Ensemble-of-costs-guided Diffu
hese clusters are then inflated into large, full-dimensional, polytopes. We evaluate our method on a variety of robotic systems, a
m such as connectivity (valid single pursuer motion), coordination (multiple pursuer motion), and tracking information (evaluatin
alled APP, is developed based on the costmap, which is widely used in commercial service robots. First, a bidirectional vertices
ate the configuration space codimension-one manifold, and we apply a local elastic update to fix the triangulation when part is i
2D map of the environment.
ath to visit those targets. To increase safety and efficiency, we propose a hierarchical state validity checking scheme, which lev
ch to deploy on low-weight embedded systems. The proposed pipeline has been evaluated in many simulation scenarios, varyin
efficient trajectory while considering the single-rotor-thrust limits. Our approach excels in computational efficiency which only tak
esign a motion planner. In this work, we develop (i) a procedure to construct parametrized environments, (ii) metrics that chara
ty. Additionally, an adaptive NMPC with a hierarchical disturbance compensation strategy is designed to overcome unknown ex
ization. Facing the challenge of finding feasible hitting actions under extreme circumstances, we propose a two-stage planning
direction, Bézier curves are employed in order to predict the future trajectory of the target, which moves in an arbitrary way. T
cell increases over time and is nullified by being in a sensor range of a vehicle. Thus, the proposed multi-vehicle planning appr
sensor systems, namely Human Handle Device and Cable Attitude Device, have been designed to estimate human command
or distributed 3D printing. Moreover, by harnessing the dexterous flexibility due to the 6 DoF motion of UAV, the framework ena
d consider it in the controller design. To mitigate the adverse effects of backlash coupling disturbance, an admittance control-ba

modified controller structure is designed to achieve the passive FTC ability for two and three rotor failure. To avoid the control
flat, thereby allowing us to calculate a slosh-free reference orientation. Through the utilization of a cascaded proportional-deriv
s this gap by proposing a new scale vehicle platform, called JetRacer-4WS, based on the open-source JetRacer autonomous v
Robust Nonlinear Model Predictive Control (OL-DR-NMPC) approach for limiting the risk of collision through considering the wo
d in this direction, this work first derives a deviation prediction methodology and exemplifies it using three largely employed con
per presents a novel ILC control scheme specifically designed for D-CDPRs, with a focus on reducing tracking errors over repet
er, Admittance Control layer and force decoupling layer. To mitigate the effects of mechanical resonance, a high-pass filter (HP
e robots during locomotion. The middle-level trajectory planner employs single rigid body (SRB) dynamics to generate optimal g
ons deemed unsafe, 3) Safety-critical control incorporating control barrier functions, 4) Gait transitions based on safety levels, a
physical limitations of robot manipulators to achieve repetitive motion planning and RCM control at the acceleration level. Addi
se Bayesian optimization. We demonstrate the method for synthetic benchmark and compare its performance to established B
ep phases to express the timing of segments' spatial movement. According to the length of movement periods, we classify the m
hance feedback portions through error-driven control gains, while the latter exploits finite-time convergence of discontinuous sw
found in wafer scanners, for example, where high-precision motion is required despite the resulting high inertial forces involved
articular use case of the C-NRIC framework, we design contact-responsive motion controllers that enables a robot to react to u
planned trajectory tracking (PT) controller related to the robot. In precision surgical manipulation scenarios, precise tracking of t
geously, the involved force subspace in contact direction is fully dynamically decoupled from dynamics in the motion subspace
used to provide stability guarantees for the interaction controller with uncertain environments, while an online learning policy al
xploration. It also enables independent control in multiple configurations. We achieve force-position hybrid control in task space
s. However, to handle new task commands like desired hand or footstep locations at a desired walking velocity, we must retrain
ould serve as predictive model in the next step. Koopman Bilinear Model Predictive control (K-BMPC) is proposed to solve the t
ect's shape transition model, which in turn is composed of a geometric cutting model and a cutting-surface-deviation model, bas
asks. This framework includes an adaption module, which continuously learns a residual of the hybrid model to minimize the ga
s are validated through an extensive experimental evaluation, cumulating over 7 km and 1.8 h of driving data over three distinct
by determining subspaces of the configurations space independent of the primary task. We call them orthogonal foliations beca
hich rearranges and optimizes consecutive edges to minimize action cost. We solve six rearrangement scenarios with a Fetch
ng for more productive exploration of the space. We evaluate the proposed approach in a set of mobility and manipulator postu
tructure size and shape affect planning time. The planner solves 80% of the randomly created instances with up to 12 cubes in
e sliding control point algorithm and motion smoothing restriction algorithm are designed. The former ensures that the remainin
arla is reported to discuss the smoothness and robustness of the proposed algorithm. Finally, we present a road demonstration
ergy usage. In this paper, we present IKLink, a method for finding joint motions to track reference end-effector trajectories while
quation through neural networks for motion planning and do not require expert demonstrations for learning. Inspired by these ap
polyhedral obstacles and constraints imposed on the maximum thrust of the UAV and its maximum tilt. We demonstrate the ap
ploit occlusions and sensor field-of-view to make planners return a trajectory which is thought to be collision-free, but is actually
onstraints either assume simple robot and battery models or are computationally expensive. To address these problems, we pro
r humans by requiring intricate simultaneous translations and rotations. Remarkably, these complexities also present computat
cking is challenging because the admissible heading direction of the robot is highly constrained by the SEF constraint. As a res
trol signals that steer the system towards sampled configurations by RRT. CBF-INC is learned as Neural Networks and has two
mobile printing or spraying applications. Our method decouples the problem into: (1) planning an optimal base trajectory subje
w how VQ-MPT improves planning times and accuracy compared to traditional planners in simulated and real-world environme
that would be difficult for trajectory optimization or sampling planner alone. We present tasks of traversing challenging terrain th
ntly combines quadtree and hash table data structure to enable UAV to analyze large scale terrain point cloud in real time. Base
ur project page: https://sites.google.com/view/unkidnappable-robot
ce (AUROC increases from 84.5 % to 91.2 %); the distance at which an accurate classification can be achieved improves from
avoiding negative feelings. It is hypothesized that the non-verbal cues will help avoid psychological reactance without diminish
man level of handover fluency, this paper investigates a bimanual handover scenario in a naturalistic, complex setup. Specificall
ed LIM2N, to solve the above new challenge in the navigation field. We achieve this by first introducing a multimodal interaction
The e-skin captures the magnetic field changes caused by subtle deformations through Hall sensors, employing deep learning

truct the left and right hands, which are ineffective in enhancing the interaction between both hands. To address these challeng
gainst missing keypoints than previous methods on action segmentation datasets. Moreover, we improve the performances fur
ed camera for human analysis. Different interaction characteristics, including touch location, human pose, and gaze direction, a
a joint of extensive public data, and show superior generality to diversified and challenging open-world scenarios. In the experi
ework termed Open-woRld Interactive persOnalized Navigation (ORION), which uses Large Language Models (LLMs) to make
obot should then identify the target object by interacting with a human user. Based on the task setup, we propose a new robotic
work applies a linearized one-step inverse solution with Tikhonov regularization to convert raw boundary voltage into a noise-red
ommand by interacting with users via question generation with LLMs. We believe that proper recognition of the given command
a deep reinforcement learning in simulation and deployed on physical Haru in reality, which can show politeness during group d
age Interface, a voice user interface for interaction and configuration of industrial active exoskeletons. The evaluation of the no
DoF position measurement and 2 DoF image measurement derived from the LED. We theoretically analyze the performances o
t analyzes the position, velocity, and attitude constraints within the IAP. Leveraging constraint relationships in the IAP, we propo
hm based on neighbor constraints to improve the robustness of the system. We demonstrate the effectiveness of our method o
ce. We evaluated our ROS 2 implementation on five different datasets, and in a real-world experiment with three robots commu
on for multiple UAVs, which robustly supports scenarios with lost global positioning and weak co-visibility. A method for Visual G
try calculations. The server back-end establishes a global reference by harnessing the shared data, reï¬ ning the joint pose gr
deband tag per robot, eliminating the need for any ground anchors, thus allowing deployment in GNSS-denied environments wi
sociated ambiguity as a point-wise uncertainty. This uncertainty, modeled by combining the state covariance with the acquisitio
ii) is capable of identifying and discarding incorrect loop closures caused by perceptual aliasing, and (iii) effectively associates a
unov functions under mild technical assumptions. Finally, we seamlessly integrate the dynamics models and the observers into
roblem that maximizes a reward matrix learned via imitation learning. We design a novel graph attention transformer network (G
This work proposes an online task and motion coordination algorithm that ensures an explicitly-bounded estimation uncertainty
 additional evidence based on simulations, that at these critical system sizes, the system performance separates into two phase
e to handle any CSLAM problem formulation. We demonstrate that MESA exhibits superior convergence rates and accuracy co
ing to different stages, rather than optimizing multi-objective functions. In the secondary phase, having poor performance divers
es, we incorporate awareness of VO performance into our policy by introducing pose estimation-related punishment. To aid robo
e RoCoBench, a 6-task benchmark covering a wide range of multi-robot collaboration scenarios, accompanied by a text-only d
vironments with obstacles. We provide our agents a curriculum and a replay buffer of the clipped collision episodes to improve
causing performance degradation. In this work, we get inspired by the genetic mechanism of social animals' instincts, and prop
large search space often makes its online use limited. Here, we present the Multi-Level Action Tree Rollout (MLAT-R) algorithm
uxiliary signal to learn advantaged actions. We further build a Multiagent-Race environment, and extensive experiments are con
ed, overall performance is generally degraded from the case with full communication. In this paper, we propose a reinforcemen
ncertainty in the location of neighbors, together with Control Barrier Functions to address the exploration-exploitation dilemma t
and the estimated probability density, such that the robot is responsible for serving the events in its assigned region. As the tea
hronous communication technologies. We show in simulations with up to 1000 robots interacting in arbitrary patterns that our s
oach leverages spatiotemporal GNN, named STGNN that encompasses both spatial and temporal expansions. The spatial exp
Based on the error model, we formulate a joint optimization problem and utilize semi-definite relaxation (SDR) to furnish a lossle
ically whenever communication links are available, propagating information on a peer-to-peer basis. We demonstrate the perfo
age paths without factoring in energy constraints. Subsequently, we cluster segments of these paths and employ graph matchi
can communicate with and hold on to. To better understand the challenges and opportunities of these and other space-filling ma
oposed controller in managing a quadrotor team with multiple leaders.
, which can then, at runtime, be dynamically allocated to the current robot team. The system allows to statically model an indivi
ame and a two-player game, then employ an iterative best response approach to solve the resulting optimization problems in a
a human-in-the-loop (HIL) model. This model enables real-time integration of human directives into the coordination process. A
ces and hyperarcs are transitions between elements. Existing methods use based representations where vertices are full comp
d B) dynamic uncertainty. This objective is accomplished leveraging GTSAM, an existing toolbox for the formulation of factor gr
ation utility, and an algorithm for gathering high-utility observations. To demonstrate this framework, we also supply complex dyn
hieves comparable place recognition results but also maintains a compact representation, suitable for sharing among robots to
ime, with the locally and temporally relevant parameters; (ii) the application can influence, also at runtime, the navigation behav
en considering uncertainties is a complicated and time-consuming process because we need to explore, figure out the size and
een communicating robots is integrated by sharing planned trajectories and projecting them onto the SOGM. The simulation re
via algebraic logic simplification, and are interpretable by robot teammates and human operators. We propose this intent repres
We examine the utility of partial belief space propagation, which allows polynomial runtime to decrease. We validate our findings
nt synchronization of observations, or the presence of a central controller to coordinate joint actions. Instead, we focus on the s
hms. The results of our experiments show the proposed algorithm’s capacity to strike a balance between curbing map unce
llocation and formulate it as a Mixed-Integer Linear Program that includes all the aforementioned features. Then we use an off-
particularly develops a Graph Reinforcement Learning (GRL) framework to learn the heuristics or incentives for a bipartite grap
activities that take place may be categorized in terms of resource development, mine-, rail- and port operations. From mineral
wed by each vehicle and coordinates the AGVs over time. In the paper we describe the coordination strategy we propose, which
trict communication limitations in emergency rescue scenarios, we have verified that our algorithm can efficiently converge to a
ence. At the same time, the agents in the group are also allowed to perform additional tasks of interest while maintaining a coll
nts in bearing formations of quadrotors as a kinematic mechanism with analogous properties to find geometric conditions for th
er-iteration performance improvement guarantee and polynomial computational complexity. We reformulate the multi-robot patr
 cilitate smooth overtakes. The decision-making process of the robots is distributed, relying only on local information. Practical a
aracterize our simulated environment, the robot swarm, and the inspection algorithm. We then employ a noise-resistant heurist
caffolding is. Prior works focus on end-to-end approaches that simultaneously plan for block placement order and inter-robot co
les feasible 6D wrist poses is trained in an end-to-end fashion to reduce the dimension of search space. Furthermore, a fast op
3D printing and carbon fiber reinforced manufacturing of core parts the platform is easy to replicate and highly modular. This p
the robot hand to have a soft, durable, and even functional structure. Specifically, the soft fingertip was crafted from TPU and c
g motion. Specifically, we formulate the discrete trajectory parameters into a temporal-based parameterization, where the prior c
hand and objects during the demonstration process. To overcome these challenges, this paper presents a novel solution in the
ensors on each finger link detect contact with humans and adjust gripping force. We conducted experiments in which a humano
eratively approach a cast skin design which preserves the robot’s dexterity in terms of range of motion and speed. We dem
einforcement learning (RL) algorithms. However, when applied to mixed datasets, the performance of BC deteriorates; the perfo

g suitable allocation with UAV-USV landing convergence and collision avoidance among UAVs into constraints with the assista
sence of communication and sensing constraints. Specifically, we consider the problem of preserving connectivity for a group o
ic mobile robot; 2) In a team of first (second) order non-holonomic mobile robots, the control input of the leader must be limited
h aspect. We use Gaussian Belief Propagation (GBP) as the inference mechanism, which permits parallel, on-demand or async
cally switch between QoS data flows with different priorities. As such, we introduce a robotics oriented analysis on the impact o
udGripper is intended to form a community testbed to study the challenges of large scale machine learning and cloud and edge
motion planning. FogROS2-Config can reduce the cost by up to 20x. By comparing with a Pareto frontier for cost and latency b
the underlying opportunistic routing protocol. Our system is task agnostic and practical to implement and operate on robots equ
passivity controller by augmentation of a tolerable passivity-shortage. Tracking offsets are eliminated with a position drift comp
es problem by selecting correspondences between the query image and the resulting bundle. We facilitate this process through
stage module to enhance the spatial-temporal point feature interaction along the time dimension. The aggregated feature also
scene. As a result, our elaborated sampling strategies effectively retained useful semantic information and narrowed the differen
(DE) scheme leverages relative depth information to realize the effective feature lifting from 2D to 3D spaces. Furthermore, a d
yle gap and a weather gap. In this paper, we present an unsupervised domain adaptation framework for object detection that c
e semantic segmentation, we efficiently compute the 6D poses by least-squares fitting. To address the ambiguity issues for sym
atent variables in a Variational Autoencoder (VAE), features from RGB camera and LiDAR sensor data are fused to improve th
the search for inlier 2D-3D correspondence. Meanwhile, a separate branch is designed to directly recover the metric scale of t
proach that first harnesses geometry supervision to facilitate coarse 3D feature perception and subsequently refines the feature
le to identify pixels of unknown objects as out-of-distribution (OoD) samples. However, when applied to images with many unkn
3D points while being more expressive than the pooling-based PointNet. Experimental results show our PVTransformer achiev
nism that can iteratively decouple feature maps into object-centric features and corresponding pixel-features used for distillation
igns for point-based transformer detectors. First, we propose Aggregated Multi-Scale Attention (MS-A) that builds multi-scale to
e the future pose of the bounding boxes. Our method is flexible in motion model selection. We explore three motion models in o
nt research has aimed to improve the backbone network to address this performance gap, we argue that there exist a room for
he image content through decomposition and maintain model efficiency by obtaining semantics from low-resolution image. To fa
y of the point-based detector by efficiently modeling longer-range inter-dependency while including only a negligible overhead.
a novel edge-aware consistency constraint among edges of each task is presented, which exhaustedly exploit geometric cons
and stabilizes the training. The trained object detection models can deal with difficult detection cases better, even though only
h uses the domain invariance and contextual understanding of a pre-trained joint vision and language model to predict the class
based methods. The proposed module capitalizes on the LSS mechanism to establish dense associations between perspective
(Detection After Prediction), consisting of a two-branch network: (i) a branch responsible for forecasting the current objects' po
rectly used to detect small targets. Secondly, an involution block is adopted between the backbone and neck to increase chann
CLIPUNetr - a new CLIP-driven referring expression segmentation network. CLIPUNetr leverages CLIP's strong vision-language
tegrating multi-scale feature fusion and temporal information, we propose that these techniques may not be sufficiently equippe
to multiple unlabeled LiDAR scans from past traversals at locations of interest (potentially from other high-end vehicles equippe
D data in context with in-distribution data, which we use to fine-tune existing segmentation models with maximum entropy trainin
sing robot interaction and a designed body frame-invariant feature. We demonstrate that the relative linear and rotational veloc
detected obstacles, and by incorporating this propagated uncertainty into optimization constraints, the planner effectively navig
derstanding in 3D point clouds. We further introduce a new text-point correlation method to learn the semantic links between po
resses the inherent ambiguities present in pseudo labels by introducing an Evidential Deep Learning (EDL) based uncertainty e
across frames. Furthermore, we present two alternatives for hand and object pose estimation. The optimization-based pose est
specifics of radar measurements. We extensively evaluate our approach with real data to demonstrate its effectiveness and es
r the wholeness of object data. To this end, differently, we propose a dynamic object classification method of low-resolution dat
umptions on the appearance of objects or environments. This allows deployment in complex scenes such as multi-storied buildin
VIP500. We then use these datasets to quantify the differences between VI-SLAM point clouds and dense RGB-D point clouds
Moreover, a context-based feature fusion module is introduced for robust and preservative sensor fusion. Finally, in order to add
attention-based tracking for sparse radar scans to include appearance features and enhance performance. The final associatio
ovements of objects (e.g., pedestrians and vehicles) inside the BEV.However, most of the current single-sensor or multi-sensor

umber of test domains including the SAX Segmentation benchmark, which includes labelled test data from dense urban, rural a
a unified manner, leading to a compact and efficient model. EfficientDPS can be trained and tested in an end-to-end manner vi
bustness, this paper introduces a novel framework named Robust Collaborative Perception against Temporal Information Distu
d point cloud instance segmentation via label propagation achieves comparable average precision as recent supervised deep l
cts are overlapped by each other. Our algorithm exhibits remarkable performance without using any deep appearance descripto
ramework, WLST, designed for WDA on 3D object detection. By incorporating autolabeler, which can generate 3D pseudo labe
fusion step that addresses data corruptions and makes sensor fusion for 3D object detection more robust. Through extensive e
verse datasets, evaluating its efficacy in detection, segmentation, and path prediction using various metrics. Notably, it outperfo
guidance of radar Bird’s-Eye-View (BEV) features and combines the radar and camera BEV features. For instance-level fu
kle this challenge, we propose a novel One-vs-All (OVA) active learning framework, known as OVAAL. This paper explains how

each iteration, we query the region a single frame covers on global coordinates, labeling all frames simultaneously. We then tac
ce. This situation is referred to as the cold start problem. To address these problems we propose BaSAL, a size-balanced warm
on anchors. In this method, a quantization standard for trajectory disturbance is designed for adaptive control. Subsequently, we
between the two aforementioned modes autonomously for better trajectory tracking performance with special heading requirem
g because the resulting solutions may violate the real constraints. We address this by leveraging convexity within the proposed
allel-friendly symmetric stair preconditioner. We prove that our preconditioner has advantageous theoretical properties when us
solver at its core. We show that MPCGPU increases the scalability and real-time performance of NMPC, solving larger problem
des with waypoint constraints and adopting separate sampling intervals for trajectories between waypoints, which significantly a
rs from noisy measurements using optimal control. Using experimental human demonstrations of 3D contour following and peg
vel algorithm based on factor graph variable elimination that can solve for the global optimum of the retiming problem with quad
 hat address this issue clearly present a trade-off between optimality assurance during transition and computational costs. Here
od is asymptotically optimal and derive finite-sample error bounds. Simulations demonstrate the high speed and reliability of the
h which it is possible to precisely control the energy-transfer timing. A prevalent challenge in the realm of CERs is the automatic
d. We propose an efficient DDP-based algorithm for trajectory optimization of articulated soft robots that can optimize the state
experimental validation on a torque-controlled manipulator in challenging tasks for which accurate force tracking is necessary. W
rimitives leading to a low symbolic complexity of the resulting expressions and a geometric intuitiveness. We demonstrate the u
ession, which improves the accuracy of the forward invariant tube as data is collected. Given a safety objective, we also provide
these challenges, we propose policy neural CBFs (PNCBFs), a method of constructing CBFs by learning the value function of
ins stable, safe, and passive. The effectiveness of the developed approach is showcased using a team of mobile robots in a hu
nal analysis on the closed-loop system driven by these controllers. We then extend the formulation to handle hybrid systems w
bot without any simplifying assumptions by employing a Differential-Dynamic-Programming algorithm. We achieved 6.8kHz, 1.9
ned by reduced data generated via a physics-based simulator and Principal Component Analysis (PCA). We use Differential Dy
ng and strongly dependent on the exploration strategy. Thus, our algorithm learns a “good” control policy via TO-guided
the Indy Autonomous Challenge (IAC) with closed loop experiments at the Putnam Park Road Course in Indiana, USA. The res
usly perform complex tasks despite real-world unpredictable disturbances such as wind gusts. To develop the algorithm, we ca
ssible for balancing the biped robot be a pregiven desired bound gamma(> 0). To achieve this objective, a robust controller is de
a challenge in learning the optimal control policy precisely due to imperfect data. In this paper, a novel data-driven robust polic
nal trajectory while computing optimal hyperparameters for a stabilizing time-varying linear quadratic regulator (TVLQR). RTC-D
nsus among the agents. To do so, we use an ergodic control formulation called heat equation-driven area coverage (HEDAC) a
m linear dynamical systems with control limits, balancing an Atlas humanoid robot on a single foot, and performing a whole-body
lable on most industrial and service robots. For each robot, CODAK uses a trained recurrent neural network to anticipate collisi
n leverage Koopman operator theory to acquire a learning-based linear system model that approximates the follower's feedback
usly deployed to assist the convoy by repairing edges, reducing the cost for the convoy to traverse those edges. The convoy is
n. We relate the spectral coefficients of the search map to each other in three different ways. The first method maps agents to p
el benchmark for cooperative multi-robot social navigation. SAMARL employs a unique spatial-temporal transformer combined
complete algorithm, Order-based Search with Kinematics Arrival Time Scheduling (OBS-KATS), which significantly outperforms
mation to estimate the unknown environment via convex optimization with linear constraints and an encoder that utilizes the dec
el to predict the deviation of a drone's trajectory when the drone enters the wind field of another drone and then use the reserva
onstraints, such as a distance budget and limits on the robot's velocity and acceleration. Our approach can be used for IPP wit
ropose a method to develop and benchmark different multi-robot coordination algorithms in realistic, unstructured and human-s
explore conflict-free paths. Our experiments demonstrate the effectiveness of our method in reducing computational demands
atic and dynamic collision constraints are treated in a unified manner. Dynamic collisions are handled based on shared position
ecution. Compared to other variations of multi-robot MPC like joint, prioritized, and distributed, we demonstrate that CB-MPC im
inter-robot collisions and resolves them by imposing constraints on the first level. The third and final level uses the resulting so
bstacles/agents beyond their FOV. To address this challenge, we propose ALPHA, a new framework combining the use of grou
ng paths for the other robots. We formally prove that the algorithm guarantees the complete coverage of the unknown workspa
ain to leverage the improved efficiency provided by topological guidance. We demonstrate our method’s ability to efficiently
fective heuristics that reduce the number of variables in the MIP model, thus improving its efficiency for large-scale MCPP prob
a control frequency of 100Hz. The proposed framework is validated through statistical analysis of the collision classifier network
rithms. Using a point cloud representation for the incoming event streams, the proposed method achieves the fast detection of
We demonstrate multiple contiguous hours of fully autonomous uninterrupted drone operations composed of several cycles of
privileged information about the surroundings and unconstrained computational budget is developed to provide demonstrations
complete map creation. In each iteration, the collector strategy stores and validates frontiers detected during exploration and se
rated using Unreal Engine, detailing the methodology of replicating the real-world scene, for producing diverse slant-angle dron
nal resource problem and improve the real-time performance, the velocity information of the UAV is further fused with the detect
pressing redundant ones at high speeds by harnessing a GPU device. PyCA uses our new Multi-Location Per Thread culling st
ork only utilizes the onboard inertial measurement unit (IMU) to detect and estimate collisions. We further design a collision reco
function, Gamma, which quantifies the total observed risk for a given pointing direction. Finally, by finding the global optimizer f
observations, making them unsuitable for incremental mapping, or are not real-time viable, especially for applications where hig
een environment while the DRL-based planner determines the best navigation goals that can be safely reached to provide the m
e the skeleton to efficiently generate a minimal set ofspecialized and informative viewpoints for complete cover age. Based on
ocations of potential cues such as doors and rooms, (2) a circular decomposition for free spaces used for target assignment. U
nt occlusions due to high obstacle density, as is the case of forests. Moreover, the problem of long-distance wireless communic
ocation to train a deep reinforcement learning navigation policy that offers low-latency computation and robust sim2real perform
orld experiments. To this end, we propose a learning framework for agile gap traversal flight, which successfully trains the vehic
it satisfies either of two conditions: (i) it is occupied by an obstacle, or (ii) it remains unmapped. At each time step, an occupanc
€œflow shadowing” because we take advantage of the fact that a downstream whisker shadowed (or occluded) by an upstre
ion and tracking (DODT) method based on an RGB-D camera, which is designed for low-power robots with limited computing p
safety, agility and the quadrotor dynamics. The perception objective is achieved by maximizing the number of covisible features
es to control every plant's growth and hygiene, from seed to harvest. Such a system surpasses current farming robots in scalab
nterest can be directly fed into our temporal convolutional network, which estimates the position, strength, and spread of the en
dule named RepNeck is designed to comprehensively fuse features extracted by the FEM, further enhancing the speed and ac
natural patterns. We deploy our algorithm in a production-scale robotic farm, where groups of growing plants must be matched
ovel two-stage pipeline for dynamic scene reconstruction. In the first stage, our approach takes scene scanning as input to reg
e introduce thermoforming as a technique to efficiently fabricate tactile sensitive e-skins that conform to curvilinear surfaces. Th
cipitate detachment of the object from the end-effector, potentially incurring damage to either the object or the arm. In this pape
he two modalities. In particular, we propose a grasp prototype adaptation module, which aims to facilitate fine-grained local fea
parent objects and lighting conditions. In this work, we propose Residual-NeRF, a method to improve depth perception and trai
avoid a typically expensive line matching process. Specifically, we first construct sparse 3D points by tracking feature points, wh
ate the required drawing speed for the desired stroke size. This open-loop system works effectively even without sensor feedb
ures and 2D object features to predict the hand and object scene geometry. With experiments on real-world datasets, we show
ng algorithm with online learning capabilities. After selecting the most promising grasp pose, the robot executes a pinch grasp w

stly, a simultaneous grasp detection and low-light feature enhancement framework is designed, which integrates residual blocks
ng collected LR tactile sequences. In addition, we present an active exploration strategy to enhance the reconstruction efficienc
ges vision-based affordance predictions from a neural network with interactive kinematic sensing in an analytical model. Our wo
mal and tangential direction. However, current approaches only react to slip and therefore react too late to perturbations. The o
module to compute their relative cross-pose. Taking point clouds of two objects as input, SCN predicts a set label for each point
ose refinement method based on joint loss is proposed to enhance the pose estimation accuracy. This refinement method empl
a graph-based model of the DLO. This model remains accurate as long as the DLO is under tension and is dynamically update
rocess to help improve grasping stability. Utilization of grasp correspondence across observations enables dynamic grasp trac
semantic keypoints on the pedicel. These keypoints assist in the robot's path planning, minimize target contact, and facilitate t
 the object of interest. In shared control, the user then guides the robot end-effector across a virtual hemisphere centered aroun
s. In this work, we present a novel approach for liquid class identification and state estimation leveraging electrical impedance s
s, a capability we refer to as Interactive Visual Failure Prediction (IVFP). This enables autonomous verification of grasps during
actuated hand. By solely sensing actuator positions and torques over a period of time during motion, we show that a classifier c
es of Freedom (DoF) poses of multiple objects within a couple of minutes of operation. Small objects such as bolts and nuts are
o their reliance on a fixed viewpoint for perception, often require redundant manipulations to randomly adjust an object's pose f
prior, leading to a novel solution possible only with the event-camera. To our knowledge, this is the first method for 6-DoF obje
sh of points for representing the object to smooth. In a second step, we transfer the resulting behavior in the real world with imi
nd truth linear velocity based on polynomial interpolation. Using Chebyshev interpolation and a pseudospectral parameterization
os. We also provide a unified decision-making process model that can operate effectively across different scenarios, serving as
posed work gives an overview of the benchmarks, datasets, and competitions that have been proposed and adopted in the las
evaluated for object perception, grasping planning, and motion planning. We believe that our benchmark will be a valuable tool
ies, however, they suffer from biases, such as length of trajectories. In this paper, we propose a new benChmarking paRadIgm
nce. To evaluate the proposed metrics, we also construct a new test set for the Waymo Open Dataset, tailored to camera-only
a comprehensive set of metrics for social navigation benchmarking.
ork provides ready-to-use tools and is easily adaptable for the assessment of custom ROS 2 computational graphs. Drawing fro
results validate the framework's ability to characterise cloth properties and compare cloth sets, and reveal the influence that dif
i, might vary over time and can significantly affect task execution. In this letter, we tackle this problem by simultaneously adapti
human skill coaching in real-time in Martial Arts which can be extended to various sports. Our system comprises of the Sawyer
e Model (TP-GMM) to learn the motion of coach demonstration, which can be generalized for the trainees. The 3D path extract
n consecutive or far apart keyframes (in time or space) of a sequential task. We perform numerical studies, simulations and rob
edom haptic exoskeleton. The presented controller decomposes the entire system into subsystems and designs controller at the
ARGCN) for modeling the dynamics of the opening in the presence of the body. We train the model entirely in simulation and v
with respect to the knee shakiness feature. This letter aims to estimate the muscle loading and unloading effects of exergaming
on's external dynamics affect musculoskeletal adaptation outcomes. Accordingly, the current study aims at showcasing the ext
notable risk of receiving careless responses, which compromise data reliability. Recognizing these challenges, this paper propo
ons that were not provided in the training set. We also compared our approach to a purely Transformer-based network, showin
his, we propose a hand-object synthesis method that is designed to generate handover-friendly motions similar to humans. This
emulates human behavior in communicating a motion destination to a human collaborator and in responding to their actions. A
deling to build next-generation robots that better assist humans. In this survey, we focus on the connections between physical h
ction, which allows us to improve the capabilities of our predictor if it is used to estimate the trajectory that the human-robot pair
ering a cobot’s base rigidity level (also referred to as base compliance level or BCL) during pHRC experiments. We explore
a in a wearable hand device, along with a simulation model for predicting interaction behavior by incorporating viscoelastic prop
y weight support. A comfortable and minimally restrictive coupling interface embedded with a force/torque sensor allows to det
naively applying these sequence-based matching methods to multi-agent exploration may require one agent to repeat a large a
they hold with a user-defined probability, so that the MPC is provably safe. We illustrate the results in the self-driving car simul
mials. Theoretically, we rigorously prove that the trajectory optimization problem incorporating DRSCCs is equivalent to a compu
accessible to robots without modern GPUs. To alleviate these issues, we propose a novel (Shield-MPPI) algorithm that provide
samples of obstacle trajectory predictions. The proposed approach reformulates the collision avoidance problem by computing
CP) that utilizes the Monte Carlo Tree Search (MCTS) algorithm to solve a POMDP while maintaining a hybrid belief. We illustra
an exponential growth in the number of hypotheses, resulting in significant computational overhead. To overcome this challen
 bility to reach a set of goal states is one and the probability to reach a set of unsafe states is zero). We compute shields that re
sability scenarios is challenging. We propose the use of a heuristic that bounds the ratio between the expected path cost in our
ovel implicit neural model of stochastic signed distance functions with a hierarchical optimization-based motion planner to plan l
rch-based CPOMDP planning to large robotic problems. We show that if primitive option controllers are defined to satisfy assig
ne. By considering the accumulated transmittance along each camera ray, our Uncertainty Field infers 2D pixel-wise uncertainty
to correct the planning strategy at each new replanning. Our approach adapts the cost function and the sampling distribution o
ends AEP to explicitly plan with respect to dynamic obstacles. Furthermore, addressing prior errors within AEP in DAEP has re
ts of our approach over state-of-the-art trajectory generation approaches and demonstrate its performance in challenging and l
ristic. IBBT interleaves batch state sampling, nominal trajectory graph construction, heuristic computing, and searching over the
pre-specified, interpretable risk metric, and (ii) offers robustness against discrepancies between actual human trajectories and t
wards the goal. The proposed approach is trained in simulation only and can generalize to real world environments on different
orld optimistically ignores obstacles, or causes unnecessarily long routes if the determinized world pessimistically imagines mor
incorporating an arbitrary number of perceptual effects in planning and fusion. Our method models the correlation with the env
the TV-POMDP: 1) Memory Prioritized State Estimation (MPSE), which leverages weighted memory to provide more accurate
onally cheap SLAM front ends in parallel and choosing the most promising feature set online. This problem is similar to the Algo
Particle Filter Tree (APFT) approach with MI as the reward, which features belief state tree nodes for informative trajectory plan
tive and time-critical settings, where the set of initial pose distributions (identified by a perception system) is finite. The finite set
mantic segmentation networks trained with massively augmented datasets to ensure reliable probabilistic occupancy informatio
ach robot from its initial to goal state while accounting for uncertainty evolution with chance-constrained safety guarantees. Ben
r traversability prediction model, which are dependent on terrain characteristics and robotic state/control. Further, uncertainty pr
o enhance trajectory time efficiency and safety under uncertainties. In this study, we begin by formulating a non-convex optimiz
s to learn such graph traversal and teammate supporting behaviors in a data-driven manner. Specifically, through a new formula
n, we propose a policy network based on graph neural networks. Experimental results demonstrate its superiority over traditiona
non-homotopic property among paths, which significantly increases the volume of the robot configuration space. Meanwhile, di
CS has many potential high-utility applications, e.g., for determining parking garage layout and capacity, as prioritized planning c
payload defines the required coalition size. To balance the total travel time of the agents and their waiting time (before task initia
el optimization formulation to find the optimal solution for a deterministic multi-robot task allocation problem with known task out
. In traffic flow prediction, we propose a spatio-temporal graph neural network that integrates local information to predict the ev
emonstrate that this approach leads to improved performance in terms of the makespan of the mission. We demonstrate the us
roposed in order to generate the joint torque needed for walking with a lower-powered actuator for lighter and more compact de
ing by an A-mode ultrasound (US). We first obtained a set of ultrasound dataset from the cadaver experiment, where the groun
dology involves collecting data from three channels of EMG and three degrees-of-freedom (DoF) joint angles and enables us to
iation (std) of 1.9%, 96.6% on average with an std of 1.6%, and 95.9% on average with an std of 2.7%, respectively.
ance parameter trajectories throughout the gait, unifying the stance and swing periods under a single, continuous model. Furth
ve compliant contact between prosthesis and ground. In this paper, mechanics models of the ATP are developed to reveal the v
residual friction, over/under-tuned assistance, or structural elasticity. Fifty-five participants performed reaching tasks while conn
scle contraction. The interface firstly introduces a new concept of motor unit (MU) spike trains, which combines decoupling of th
assistive torque at both hip joints. The D-SEA achieves precise force control with a custom controller based on accurate spring
moment arm and displacement, and implemented them as shape memory alloy (SMA) spring-based actuators. The resulting pr
xion of the ankle. The kinematic compatibility of the proposed exoskeleton frame is evaluated in straight- and curve-walking sce
front-enveloping designs and propose solutions. Furthermore, we validate the usability of our designed exoskeleton through tes
 nts between locomotion modes. This threshold-based method improves real-time embedded processing time performance by u
amework is based on the concept of the divergent component of motion, also known as the Extrapolated Center of Mass, which
average of the three torque profiles while we measured their energy expenditure. Relative to walking with the device turned off
nt a new way of coordinating the user and their device. This paper explores the initial implementation of task-space control for a
ion algorithms, can greatly assist exosuits in understanding the user's environment and adapting their behaviour accordingly. T
eometry-informed distance candidates also improve a pre-trained model's accuracy if the extrinsics or the number of cameras
h another model. So, a pre-existing data set can be used without having to recalibrate the camera. We provide the methodolog
wer and energy; thus, existing demonstrations focus on offline radar processing. RadCloud overcomes these challenges by usin
d cross-correlation that captures geometric and semantic relationships between entities. The respective matches allow us to rec
cloud distribution and density is designed. The codes of the proposed method is open-source and available at https://github.com
ocal spatially continuous storing strategy that allows for fast access to points while minimizing memory usage, and local on-tree
orrection of lidar measurements is still an open challenge mainly due to the lack of clean 3D data that could be used as ground
nd segmentation and terrain estimation performance of GroundGrid and compare it to other state-of-the-art methods using the S
strong geometric and perceptual cues in 3D shape reconstruction. During the inference, the trained model is able to produce hi
ent obstacle avoidance with a very large perceptive radius (30m). We perform extensive statistical evaluations on indoor and o
thm based on Theta*, and a hybrid approach that combines the two. The obtained results highlight the efficacy of tactile sensin
lting in non-smooth paths. To prevent this problem, we first introduce spline interpolation of control input sequence in the MPPI
mprove the same using an additional quality distribution method. This enables seamless integration of the robot’s embodim
re kinodynamic search and spatiotemporal trajectory optimization, which can generate collision-free, smooth, and efficient traje
al-time planning and execution of medical instruments by providing 3D anatomical image overlay, human-robot collision detectio
s sparse, object-centric, and embodiment-independent task representations from a small number of human demonstration vide
n of complex tasks in soccer specific problems. We extensively evaluated our planner in several complex simulation environme
d to run in real-time and is intensively evaluated in simulation and real-world experiments. The results compellingly demonstrat
teforce approach, by integrating EC techniques such as Genetic Algorithm and Big Bang-Big Crunch Algorithm. Comparative a
ce with stroke survivors and validate its potential for future community use. We then present a systematic method for quantifyin
a shoulder actuation testbed for exoskeletons, incorporating a fusion hybrid linear actuator distinguished by high back-drivabilit
pliant trajectories for passive rehabilitation exercises, considering that previous positions of the trajectory are attainable for the p
bling. The used approach focuses on making the ankle joint able to smoothly adapt in real-time to different walking styles defin
riging adjustment (CKA), which combines a biomechanical simulator and real data to build an accurate model of the real system
adic collaboration between the patient, the robot and FES. A patient-driven controller is designed where the voluntary movemen
mobile manipulator (Hello Robot Stretch RE1) that visually and verbally guides users through PT exercises. The robot's soft end
a new method called Recovery-based Supervised Learning (RbSL) to accomplish safety-critical tasks with various goals. To ev
and increasing training stability. This novel approach excels in its performance on challenging Safety Gym tasks, achieving sign
ages the Gaussian mixture model to adeptly depict the distribution of returns. This approach ensures precise, authentic samplin
F design. We present ASRL, a novel adaptive safe RL framework, to fully automate the optimization of policy and CBF coefficie
. We validate our approach on two complex robotic tasks without relying on any prior information about the environment: (i) loc
mpared to the traditional Safe RL algorithms is that it omits the need to constrain the policy search space. This allows SORL to f
s. In this paper, we begin to address this issue through differentially encoded observation spaces. By reinterpreting stored imag
hared and variable task information. We then show that our model outperforms the state of the art on the MT10 and MT50 ben
robust lane detection. Additionally, Bi²Lane employs temporal reverse refinement together with temporal forward refinement t
o mismatches between external physical objects and augmented reality overlay images, adversely affecting the AR display’
 tion selection in this model-free RL setting by introducing additional adaptive control steps based on Extremum-Seeking Contro
by proposing an attack on the training process of the value-based algorithm and the gradient-based algorithm, utilizing gradient
arios, leading to high-fidelity environment mapping and a sampling speed approximately 40 times faster than existing platforms.
address this, we propose a novel approach to better combine the offline dataset and the inaccurate simulation data. Specifically
base policy trained completely in simulation using model-free RL and an adaptation module trained via supervised learning with
ng paradigms: HDC-based supervised learning by demonstration (HDC-IL) and HD-Reinforcement Learning (HDC-RL). It rende
O insertion with reinforcement learning. Firstly, we design a flexibility estimation scheme that characterizes the properties of diff
those issues, we propose an RL framework, Learning Action and Query over Belief LTL (LAQBL), to learn an agent that can c
a prominent role, we acknowledge that many fundamental object properties extend beyond the visual domain to encompass att
n confined spaces to retrieve the desired object using a given robot grasp. Our method actively senses the environment with the
th small relative translations. Hence, extending these MPI-based methods to more complex scenes like large-range or even 36
ces for any point in the task space to the robot motion. Further, we present an algorithm combining the speed of the deep learn
imises for the relative 6-DoF transformation between two arbitrary neural fields, even if those two fields have different scale fac
lf-supervised point cloud representation learning framework, named 3D Occlusion Auto-Encoder (3D-OAE), to facilitate the det
other pre-trained models that build vector descriptions suitable for robotic control tasks, thus capturing "what" the entity is. Thus
ry (SKT), an object-agnostic representation that is highly versatile and applicable to various everyday objects. We also propose
, based on the pairwise inner product of the input measurements. We describe and mathematically deduce the concept, establ
-centric representations. This enables the discovery of slots that consistently reflect the composition of multi-object scenes. The
Nav (zero-shot transfer to a held-out scene in the real world), and 3) the benefits from variations in PVRs, primarily data-augmen
ntion to image space that is most consistent with the external signal. We use a pairwise consistency metric as a representation
s that explicitly encode point clouds are significantly more robust than their RGB-D counterparts. Further, we find our proposed
ously observed parts of the environment unless an extensive sensor history is stored for training. Voxel-based representations
nly the final layer activations, SOFT individuates and locates object-like entities from PVT attentions, and describes them with P

. Our method begins with a pointwise classification preprocessing module, which categorizes all sampled points according to th
2) to inpaint the missing areas of scene color images rendered from different views. Next, we lift these inpainted images to 3D
ation for learning point-pair matches. To further tackle the correlation noises brought by feature distribution divergence, we desi
aggregation in stereo matching. To reduce the triangulation error in depth recovery from disparity, especially in distant regions
multi-view data. Our approach surpasses previous self-supervised registration methods on ScanNet and 3DMatch, even outperf
uts, including stereo vision, various Lidars, Radars, and audio arrays. It offers a unique aerial perspective vital for addressing re
nd present a method that uses annotated objects to learn the objectness of pixels and generalize to unseen object categories i
orithm, allowing for the joint training and inference of both the feature extractor and the clustering. Its central component is a hy
s into the 3D volume along the line of sight with a Gaussian-decay function, centered around the depth prior. Volumetric seman
the model to gradually grasp beneficial depth cues against the weather effect, yielding smoother and better domain adaption. M
a classification problem. We first represent sequences of raw observations as spatiotemporal graphs, which significantly reduce
plications, e.g., where a depth sensor onboard a robot perceives a dynamic scene and has the goal of estimating its own motio
ruct the DLO in 3D space to predict the occluded part of the DLO, and physically smooth the reconstructed DLO. By testing on
using only proprietary distance estimates, and consistent recovery of the albedo of the planar surface, which is not possible wi
pose the Vision-Language Models Goes 4D (VG4D) framework to transfer VLM knowledge from visual-text pre-trained models t
pose ConceptGraphs, an open-vocabulary graph-structured representation for 3D scenes. ConceptGraphs is built by leveragin
the topological geodesic distance to track self-occluding DLOs, and the introduction of a non-Gaussian kernel that only penaliz
utilize network classifier weights in 3D models to balance the training on long-tail class distributions. While previous work in the
of the T-C joint is demonstrated in a flat-foldable and self-assembling robotic arm, providing its applicability in environments with
panels are commonly installed. This is for two primary reasons: (1) most commercially available solar trackers are too large to b
onparametric macrostructures such as appendages, organs and cavities. Although results are provided for open loop control on
n, a simplified model of the gripper was conducted, and the geometric parameters of the gripper are listed in a table for referenc
thodox architectures are occasionally designed. Such arms do not have closed-form inverse kinematics and little insight into the
speed recorded was 0.20 BL/s (body lengths per second) with a COT of 15.82. These results were compared against other robo
and is capable of performing a variety of maneuvers including a full circle with a radius of 0.23 m (1.4 BL) and a figure eight, w
n be further increased by introducing a redundant DOF with a constant angle. As a result, the proposed hybrid parallel robot ach
he proposed concept, we conducted comparative experiments between a common tensegrity prototype and a pulley-guided pr
ots, tasks, and environments? In this paper, we provide datasets in standardized data formats and models to make it possible t
by speeding up: (a) the class of recently introduced RT-2 models, the first VLA robotic policies pre-trained on internet-scale da
ates a lifting success rate exceeding 90.2% when grasping various objects, including items such as 1mm basil seeds, thin pape
y insight is that contact microphones capture inherently audio-based information, allowing us to leverage large-scale audio-visu
a human `plan predictor’ that, given a current image of a scene and a goal image, predicts the future hand and object config
ntation where the set of valid configurations has positive measure. We describe how to use this parametrization with existing m
ce sensor faults and attacks. In this paper, we study safety-critical control synthesis for robotic systems under sensor faults and
. BCBFs are general and can be applied to a variety of robotic systems that use extended Kalman filters as state estimator. We
g to leave the set. We call this set 'maneuver set' or set $mathcal{M}$. We substantiate the algorithm's efficacy through mathe
eoretical analyses of the continuity and continuous differentiability of the proposed CBF. We empirically evaluate the proposed
curated. In this regard, we create a framework for differentiable modeling of cloth dynamics leveraging an Extended Position-b
which is incorporated into an online reachability calculation. Reachable sets are calculated using a computational graph analys
er standardized testing disciplines, to serve as an adaptive, repeatable, and fair disturbance rejection testing equipment for leg
mework to stress-test the vision-based controller offline and mine its system-level failures. This data is then used to train a class
bilistic safety methods such as Discrete-Time Control Barrier Functions (DTCBFs) to ensure the safety of the system up to som
the proposed approach, we successfully constructed an IPC for the vision pipeline. We then designed a control algorithm that u
We employ a scalable approach to generate symbolic controllers for the multi-robot system, utilizing distance functions. We als
approximate control-invariant sets, but classic theoretic control methods break down if the set is not exact. This paper presents
approach for numerically approximating the viability kernel of robot manipulators. Our approach solves optimal control problems
function (CBF) to regulate safe regions in the scene. A model predictive controller (MPC) leverages the CBF-based safety con
tribute to reducing the impact forces with humans, and thus increase the safety of collaborative robots. Dynamic models, valida
s belief state within this safe set. As a result, we ensure that the risk of the unknown robot's state violating a safety specification
ld model whatsoever; the observations need not be I.I.D. and can even be adversarial. The theory extends results from conform
pends on the design and conservativeness of the chosen backup controller, especially in our setting of human-driven vehicles in
pondences. Based on graph representations for both point clouds, we extract equivariant node features aggregated in their loc
bution and density estimates are combined in a Dirichlet-based evidential uncertainty framework that efficiently computes epist
umerical stability of the proposed SR-VINS is inherited from the adoption of square-root covariance while the never-before-seen
cutive point clouds, predicting relative motion, or utilizing sequential point clouds to address feature degradation, our SeqTrack
e to loop closure make our system a truly low-latency system that achieves global consistency. The neural implicit representatio
ong-horizon manipulation using pre-trained visual representations for robotic control. At a high level, UVD discovers subgoals b
em strength and gripper pose (yaw, roll and offset w.r.t. the apple). Our results show that grasping performance improves when
 ing-based methods, with the former having larger errors and the latter requiring a large amount of data. To address these issue
lysis framework. The device is equipped with an elaborate feeding module and a capturing module for automatically pre-proces
aided by fiducial ground control points (GCPs). Unfortunately, GCPs severely disrupt crop management activities. To address th
C) vines such as Concord. The simplicity and affordability of this technique lends its self to deployment by farmers today or on fu
eep learning to build a convolutional neural network that learns to segment individual leaves. As done in current breeding practi
orward model for predicting resulting deformations, and a contact policy for inferring actions to manipulate tree crops. We cond
ults revealed that the path of agent1, which was assigned the highest priority in both the indoor and outdoor environments, was
he labelers.This paper addresses the problem of reducing the labeling effort required to perform leaf instance segmentation on
e overall SWP analysis process. To achieve so, this work considers three core components: 1) informed planning, to determine
pper. The proposed dual-camera image-based visual servoing (IBVS) controller is designed to deal with the image feedback fro
ntation framework for infield fruits using transfer learning. Concretely, our work is aimed at addressing agricultural domains that
oud data by adopting Hausdorff distance metric and modified breadth first search algorithm. The proposed skeletonization meth
m on a real field, where we reconstruct accurate point clouds and meshes of sugar beet, soybean, and maize.
aracteristics of the object-extracting aerial manipulator, we derive the dynamic equations for two flight modes, i.e., free-flight an
he trajectory optimization underpinning VACNA. In particular, our CEM projects sampled vehicle trajectories onto the constraint
a backstepping approach for the position controller with the CBF filter acting on the position kinematics to produce a filtered virtu
at concepts and structure often observed in urban indoor environments cannot be easily transferred to these novel scenes. Mot
hich allows the efficient recording of robotic sensor data streams. We show that our approach reduces the overall time needed
g-cutting module, and a fruit recovery device. Equipped with a tactile sensor array, it can accurately sense the sweet pepper pe
the roll direction depending on the environment of grooves. An accurate calculation of lateral and angular position of the transp
obile robot with the NVIDIA Jetson Nano module which is running a pH-estimator trained with a Convolutional Neural Network (C
previous work on perception and manipulation. The integrated system demonstrates the ability to autonomously detect and cut p
arameter estimation algorithm. The proposed system features a multi-modal and entropy-based landmark correspondences app
map representation of the environment that is able to both speed up expensive ray casting operations and maintain the high occ
node to maximise its chances of being cut. Previous methods of NBV planning mostly focused on global view planning and us
nd harvesting, and can then generalize to different instances of the same task. For transplantation, the motion constraint arises
datasets during their convolutional encoding and predicts the dataset affiliation of the resultant feature map as a pretext task. O
labeling. In an unsupervised manner, our approach can generate images and corresponding labels to train machine vision syst
ion of logs around the machine, then grasp planning for where to grasp logs, through motion planning and control of the log loa
gative and positive virtual boundaries. While the robot is cruising, the likelihood detection method is adopted for obstacle searc
extract forestry inventory information. Our approach provides per-tree geometric traits such as diameter at breast height togethe
ased objective function. Through this combination it can explore the space of hypothetical inter-quantity relationships and evalu
oward a designated goal by simply maintaining their initial formation and traverse to it. Unlike existing works on herding using fl
on and task-space motion control built on the concept of dowel connectors. Simulation results demonstrate that a collective mod
PLASTR in simulation and on a real-world test setup and compare it to two handcrafted heuristic baseline algorithms. PLASTR
uctures using a set of construction voxels, which do not contain electrical features. This paper describes the design, developme
cts. Object correspondences between missions are determined by clustering the object's learned descriptors. We demonstrate
erwater vehicle (AUV) for manipulation tasks. The proposed task planning algorithm is computationally simple, and we impleme
was originally trained with imitation learning on blasted rock for use with an unknown gravel material, requiring 20 bucket fillings
he difficulty of the task does not change significantly for different domain parameters. In this study, we empirically show that uni
E measurement with pose information to reveal and map subsurface defects. The system consists of three modules: 1) an Omn
 rresponding to the action language. Fine-resolution inference is constructed as a Markov decision process, which takes action
previous grasp datasets is to harness the universal knowledge embedded in these foundation models. We present Grasp-Anyth
perform high-level task anticipation, using the anticipated tasks as goals in a classical planning system to compute a sequence
we introduce a method we call Plan Conditioned Behavioral Cloning (PCBC), that allows finetuning the behavior of high-level p
vable tasks using LLMs. In the proposed method, an LLM is used to collect missing information from the environment using a r
ncy, we construct a hierarchical planning domain that captures the attributes and connectivity of the scene graph and the task a
e maintaining semantic correctness and minimizing re-prompting. In VirtualHome, CAPE improves a human-annotated plan corr
antic knowledge from an LLM to achieve zero-shot generalization to novel concepts. We conduct experiments on Language Au
uild and schedule pipeline components; (ii) A bridge tool named DOS Connect that make cloud-edge bidirectional communicatio
ence, robustness against bounded external disturbances and model uncertainties, non-overestimation of the control gains, and
l control inputs to be defined at once; and (iii) to lump all unknown dynamics from both translational and rotational levels into a
stability conditions of the closed-loop system are formulated.
oiding the delay of the next action. To this end, we compute deviations between real and nominal system states, predicting forth

15-20% of the actual values). However, these approaches fall short in accounting for significant changes in tire models that can
on resulting from the MPC controller, the filter provides increased robustness to the controller in phases of uncertainty as comp
ccuracy of integral and the damping of derivative, respectively, the persisting on sliding surface is eliminated, such that the cha
control formulation based on high-level integral feedback terms in the task-space including the constraints. The proposed metho
To validate the effectiveness of our proposed safe-by-design digital twin framework, we conducted experiments using a huma
ect of orientation in various applications. To alleviate these shortcomings, we present an innovative approach geared toward en
hod that can compute distances 1. at many robot configurations to examine along a trajectory 2. in real-time for online robot con
arios when hands and arms are occluded. Toward this goal, a three-branch multimodal Deep Neural Network (DNN) regression
lanning framework for synthesizing legible and proactive behaviors and demonstrate that our robot planner naturally leads to pr
erved motion of human driver. With these inferred internal states, we predict the future trajectories of human-driven vehicle by f
(CCP) and Average Collision Probability (ACP). To validate the utility of these metrics in facilitating algorithm development and
axed hierarchical control barrier function quadratic program subject to multiple concurrent, possibly contradicting, constraints. J
ize cost by selecting the most `important' of it for adversarial training. In particular, we propose a task-based robust fast (RAST
stablishing a system for safe robotic-diver approach to establish face-to-face communication that considers non-standard huma
om build sensors for multi-day line monitoring, and assessing aging conductors surface properties, to refine their thermal mode
safe fire extinguishing using a 4-m-long flying fire hose consisting of two nozzle units and flexible hoses. An H-infinity-based dis
pplied by batteries, which directly impacts their task execution and mobility. This paper tackles this problem, and proposes a se
anning. Additionally, the global graph is build in an efficient way, using node-edge information exchange only on overlapping re
experiment results show that the proposed algorithms outperform state-of-the-art approaches in multi-object search.

dundant while keeping channel utilization bounded. In particular, a hexagonal tesselation coverage pattern with 3-connectivity is
obot to explore its surroundings at lower update frequencies, enabling the use of lower-performing and lower-cost hardware wh
P that demonstrates the learningless and safe recognition and 6D-Pose estimation of objects from poor sensor data (e.g., occlu
e model to provide scene information for mapping. By projecting egocentric scene understanding into the global frame, the pro
objects in a 3D scene. The LLM then evaluates the spatial and commonsense relations among the proposed objects to make a
m past driving trajectories by labeling regions traversed by the vehicle as highly traversable. Using the self-supervised labels, w
-side points in a smaller voxel and more far-side points in a bigger voxel. However, this causes a problem of the receptive field
y 0 labels (SPINO). In detail, our method combines a DINOv2 backbone with lightweight network heads for semantic segmentat
vised illumination enhancement loss and supervised intersection-over-union loss, thus enabling the joint minimization of both los
r and Dynamic Representation Augmentation Modules. The Deformable Representation Fusion Transformer introduces the def
D heatmap of potential object locations. To manage the computational load, we use Fog-ROS2, a cloud robotics platform, to of
SD-based environment models into Knowledge Graph (KG) representations that facilitate semantic querying and integration wit
mentation quality, compared to a baseline that processes images independently. For 3D map processing, we propose a novel g
g the additional utilization of structural semantic information, such as properties of the intersections leading to frontiers. The algo
we extract region-based embeddings and their associated confidence maps. These are then integrated with the 3D knowledge
-learned association strategies such as probabilistic clustering or voting-based center predictions. Instead, Mask4Former introd
ely on post-processing steps and are often not end-to-end trainable. In this paper, we propose a novel approach that can be tra
s. To alleviate these issues, we propose a novel VSN policy, HSPNav, based on Hierarchical Scene Priors (HSP) and Deep Re
pectation, by learning histograms from available training data. A novel Graph Convolutional Neural Network (GCN) model is dev
his modification and addition to the GDRF provides a key shift from unsupervised machine learning to semi-supervised machin
s short of the necessary precision. To overcome these limitations, this paper introduces a novel end-to-end collision-free space
and then suppress noisy samples based on the graph structure. A subgraph voting scheme is proposed to conduct transductiv
current radar scan with temporal information without passing aggregated scans through a neural network.We propose a full-re
sion strategy is among the worst calibrated, and propose a learned pipeline that combines fusion and calibration, GLFS, which
the satellite map tiles for each sample in nuScenes and release a complementary dataset for further research. To enable bette
nd masked input modalities. The proposed masking strategy prevents over-reliance on a single modality. It also improves the ac
maps using an effective graph-based collaborative SLAM approach that detects inter-agent loop closures. To semantically deco
istribution to camera depth distribution. The method boosts the learning of depth estimation of the camera branch and induces
htweight semantic scene completion network (SCONet) with self-attention to enable accurate obstacle predictions by capturing
hicles. Bringing the visual-inertial (VI) supervision is challenging due to the spatial agnostic of dynamic objects. Moreover, spuri
era and LiDAR evaluation based on the proposed sensor perception model. We also design an accelerated high-quality framew
nt all-day detection by fusing both RGB and event modalities. Our EOLO framework is built based on a lightweight spiking neur
ork. Our method has two contributions. First, our bidirectional fusion scheme explicitly and implicitly enhances the 3D feature v
rning-based spectral knowledge distillation technique known as SK-Net to improve the performance of deep learning models fo
n network used for inference. We accomplish this task through a GAN based cross-modality module which uses an image as in
which is required by downstream path planning and control tasks to improve the smoothness and safety of autonomous driving.
ine fusion-based ultra-wideband-inertial state estimation), in real-world scenarios using public data set and experiments. The pr
hms and accelerate the development efficiency of employing GNSS in multi-sensor fusion algorithms, we open source the GNS
along the timestamp. Motivated by this, we propose a novel joint spatio-temporal reasoning method for event-based moving ob
Fusion Alignment Transformer (GCFAT), which improves the extraction of camera features and provides a deeper understandi
n with four stages, including monocular depth prediction, global scale alignment of monocular depth with sparse Radar points, q
k of datasets specifically designed for indoor environments, with a focus on dynamic object localization. This paper introduces H
s-attention layers at different network levels to fuse these modalities and construct informative features that leverage the streng
sed on the intensity variance. In addition, a suitable geometry is proposed for the calibration and verified using simulations. Our
nore the contribution of the most confident regions of each modality. This leads to structural ambiguity in the modality fusion pr
d and experimental datasets. The IMM is shown to have good consistency properties and better overall performance when com
k, the robustness of the SLAM framework is oftentimes given little consideration. Consequently, even if the robot is physically c
mentarity. Moreover, PVTrack proposes the Siamese Point-Video Transformer for feature aggregation to overcome dynamic env
 the evaluation three open source data sets are used. Since these do not represent the applied bi-modality, a new data set with
as yet to be demonstrated. In contrast, our proposed methods solve this problem. The preemptive impact reduction feature can
s shown to be significantly more efficient than current methods.
g more robust to estimation errors in the surface normal. By applying Gershgorin’s theorem to the closed-form matrices, we
n fits the relationship between image quality metric and exposure via image synthesis technique, and the optimal exposures for
mpared to traditional vision-based tactile sensors, it offers the advantages of i) high spatial resolution, ii) significant affordability
situations, allowing it to robustly estimate the time difference of arrival (TDoA). Our robust-TDoA model incorporates a Mel sca
a Word Error Rate of 40.38% and a Character Error Rate of 28.85%. The direction of arrival of the speech signal is computed
iverse calligraphy images without requiring labeled demonstrations. Our model achieves fine-grained control of various writing u
o faster but often less accurate estimates. Our approach combines the best of both worlds. It exploits the benefit of having a tem
an localization performance on a real dataset of 264 Mars rover panoramas. The proposed solution has also been successfully
capabilities of a highly robust low-level policy to follow 6D commands and a high-level policy to enable three-dimensional spatia
an generation, and plan execution failure. We then propose the self-recovery prompting pipeline, which explores the necessary
eep learning model is employed to forecast quilt state transitions before and after quilt spreading actions. To improve the sensit
eypoints and correspondence into a graph and use graph neural network as architecture of controller. This design utilizes both a
edding from past observations and servos to the desired pose under the supervision of a PBVS that can access the ground trut
ducation activities such as ward rounds. The prototype called ARIS 1.0 was developed, including a three-wheeled omnidirection
echanisms’ characteristics, the proposed cable-driven spherical parallel continuum manipulator offers many inherent advan
of 0.1% position error and 0.5 deg orientation error. This technology will enhance the practicality and applicability of fiducial ma
ose RoboKudo, an open-source knowledge-enabled perception framework that leverages the strengths of the Unstructured Info
n environment of interest. The customized morphology is determined through an optimization technique by considering an envir
ntic errors in the generated plans. We demonstrate the efficacy of the proposed feedback architecture, particularly its impact on
provide validation results and performance metrics on challenging simulations relevant to robotics applications. Our method is m
L, which leverages Generative Adversarial Imitation Learning (GAIL) to emulate real pedestrian navigation in crowded environm
rs of agents. In this paper, we introduce MuRoSim, a multi-robot simulation for lidar-based navigation specifically designed for D
ortant systems such as cloth or shells, have stability issues in complex friction-dominated setups or cannot robustly prevent pe
s in spatiotemporal formations, and the other by travelling-wavy undulations inferred from the same real fish kinematics. Our re
ng with synthetic data pipeline for machine-learning applications. It comes with ROS1 and ROS2 bindings to control not only th
ibly control the sampling process to generate scenes that exhibit desired characteristics. Our experiments show that SceneCon
cktracking mechanism. We modify the popular Dantzig algorithm to get valid solutions under multiple frictional contacts. We co
mic logistics robot. The study breaks down the modeling of the whole system down to creating appropriate colliders for the rolle
vised learning of descriptors. Extensive experiments on both indoor and outdoor datasets show that our TCKDD achieves textit
antic Bird's Eye View (BEV) maps are growing in popularity, but datasets often lack accurate ground truth and are not flexible e
optimization technique, a data association strategy and an acoustic intensity outlier rejection algorithm for reliability and accura
CURL's spherical harmonics implicit encoding. Different from the popular Iterative Closest Point (ICP) based LiDAR odometry te
nditioned environments. The proposed approach combines modalities in a factor graph-based windowed smoother with sensor
the rig. This limits the linear rise in the number of features with the number of cameras and keeps the computational load in che
s: stationary, visual, and dynamic, tailored to handle diverse cases. Furthermore, we develop mechanisms to detect sensor an
resolution implicit field and enhances robustness in challenging environments with sudden viewpoint changes or sparse data c
h detects and eliminates dynamic objects based on the joint errors, while also harnessing the illumination invariance offered by
se-pose constraints while preserving observability and minimizing information loss in the blanket. Moreover, since landmarks a
 pute the maximum elevation map on a bird's eye view projection of point clouds to compute feature descriptors. In contrast, we
e Ring structural information, estimating the normal requires only the range information of the points when a new scan arrives. T
er and the measured Doppler velocities to estimate the vehicle’s ego-motion. We enrich this with a filtering step to improve
mental Probabilistic Consensus, a method that approximates the solution to the combinatorial problem of finding the maximally
efined threshold employed in the hierarchical refinement process of DFM by leveraging a more flexible nearest neighbor search
etry and Mapping framework VOOM using high-level objects and low-level points as the hierarchical landmarks in a coarse-to-fin
ng solution within the textit{Single-Stream} architecture, known as Lite-SVO, which is a lightweight yet efficient multi-feature sha
oach to real data from the feature tracking stage of a real-time visual SLAM system and showing that it yields far more stable p
e local dynamics of events. The mapping performance in terms of completeness and local smoothness is also improved by com
ms layer encompassing sets of wall planes, and (4) a floors layer gathering the rooms within a given floor level. The above grap
approaches, and the major challenges for VPR and how they may be addressed. 2) As a contribution for researchers acquaint
and field of view(FOV) can lead to inaccurate measurements. While prior research has explored radar-inertial odometry based o
ur technique surpasses the accuracy of established methods. Importantly, given its detector-agnostic nature, our algorithm also
confidence maps of both visual factors and it can be further extended to include the IMU factors as well. Extensive experiment
We show that the proposed approach can be used in a sliding window continuous trajectory optimization combined with IMU m
n continuous-time estimation is several times higher than that in discrete-time ones, leading to insufficient constraints on the LiD
be learned to varying degrees throughout online training. The proposed method is evaluated qualitatively and quantitatively on
gistration and the classical extended Kalman filter scheme. We propose an adaptive data association that considers the relativ
mputational efficiency. Furthermore, we introduce a point-wise uncertainty estimation method to augment the accuracy of scan-
rdinate frame. While expressing object points in their local reference frame may seem intuitive, it does not necessarily lead to th
sed on union-find which can save resources and further improve the accuracy of plane fitting. This module can distinguish the k
To represent and solve the resultant optimization problem more efficiently, this paper then adopts and formalises the concept o
clean static point cloud map generation, which reaches the state of the art performance among existing offline methods. DORF
e fly. Finally, the real-time odometry, map, and mesh are published via our broadcaster. The primary contribution of this work is
of robotic perception in the context of simultaneous localization and mapping (SLAM). We show how to formalize the sensor arr
, we design a pre-processing module, which includes extraction, inter-class processing, and post-processing of hybrid point fea
pose into translation and rotation. This approach derives the optimization of a linear form in terms of a translation with reduced
nimization of the largest eigenvalue of adjacency matrix for measurement graph. This problem formulation gives virtues twofold.
blem over factor graphs. We evaluate our method against baselines across a range of simulated and real-world tasks and demo
nd the tags are attached to the robots. Although this method performs well under low visibility, modifying the existing infrastruct
try system outperforms alternative methods using the physical-enhanced components. Our findings also reveal that using the p
ltaneously estimating 6-DoF camera poses. By leveraging the structural regularities of indoor MW environments, we overcome
c verification. In addition, we propose a new supplementary metric that addresses the leaving-out issue of the conventional me
metry) and embeds Contrastive Language-Image Pretraining (CLIP) features in the graph nodes. We use this representation fo
ingle image. Specifically, we consider the bipartite graph encompassing cameras, object poses evolving dynamically, and came
spective gradients. These residuals can be added to a factor graph optimisation problem, either as frame-to-map factors for the
ecifically, we focus on the problem of selecting the optimal viewpoint at a given location. We compare existing approaches in the
duce a method to achieve implicit surface uncertainty using color uncertainty, which reduces the time needed for view selection
a loop closure there, and the robot’s travel cost to navigate to that pose. By choosing poses that provide the largest reward

sacrifices the inherent temporal resolution of event-based cameras. This paper instead presents a complete stereo VO pipeline
filter builds upon a symmetry that properly includes all the states in the group structure. Thus, we achieve improved behavior, p
lane lines to optimize the positions of the lane line sample points, thus building a reliable lane line map. Finally, we use the bui
e robot. In this paper, we propose a deep reinforcement learning (DRL) approach to train a policy simulating the behavior of non
n. For instance, after detecting that there is an occluded car, the robot may need to actively perceive the car to know whether it

tics. We present the results of a user study demonstrating that users are more satisfied with multimodal than unimodal explana
based on attention mechanisms is proposed, which leverages the spatial-temporal graph to enhance robot navigation. We emp
ively, despite processing time limitations.
the human condition where participants displayed longer utterances and shorter response latency. These observations sugges
requirements. Based on these requirements, we built the robot-mediation system prototype. We conducted a pilot study using
e. Therefore, in this paper, we proposed a ChatGPT-powered adaptation system (ChatAdp) for human-robot interaction which r
econfigurable responses and adaptive adjustments in various radiation directions, gains, and modes. The interdisciplinary appr
thod tailored for continuum robots. Our approach embeds safety constraints derived from the robots' posture states, facilitating
ensitivity of the structures. We present the design and fabrication of the proposed gripper in detail. A theoretical model is introdu
he initial position by the threads, the actuator also returns to the initial state. To avoid the jamming effect during the deformatio
ctuators may not appear suitable for the harsh conditions that prevail in the deepsea, but when combined with a rigid, bistable m
astening, and can easily release the vine robot. This approach enables carrying of loads of at least 28 kg (limited by the tensile
try and thus must search for solutions in a bigger space. Instead, we propose to leverage the symmetry for sample efficiency b
odel parameters and deviations from model identification of the soft continuum robot. It also provides a qualitative measure for
r for the harvest of blackberries. We adapted an existing gripper concept, initially testing it on a sensorised raspberry physical t
two walls, turn in two directions, and adhere to various surfaces. Different motion and control strategies are proposed based on
hape change, and rolling, crawling and bounding gaits. Furthermore, Tetraflex is able to roll onto and engulf objects then subseq
nificantly enhance structural stiffness as well as improve the model accuracy drastically; 2) paired with a novel flow-controllable
tive Eutectogel packed in an Ecoflex tube (WEE), which is sensitive, stable, and easily manipulated. The wire and its array facil
ability within confined space. In response to the proposed solution, the designed ECR with two segments helical spring structur
ve acquired the constitutive characteristics of this actuator. Subsequently, an innovative gastric endoscopy robot is developed b
synergy of the body actuator and wheel-legs enables the robot to achieve multiple crawling gaits, including gecko-like crawling
ic gripper created for this purpose. The gripper has a compact form factor and retractable fingers with specialized TPU finger p
g structure on both sides by the central muscle part to simulate the process of the manta ray’s lateral fins fanning to propel
rasping and locomotion tasks.
inchworm actuator array that can rotate an antenna 88 degrees in incre- mental steps of less than 3.4 degrees with zero-power
design RObotic Modular Endoskeleton Optical (ROMEO) fingers, which are underactuated two-segment soft fingers that are a

erities via a pressure normal to the surface. Cooling down below 60◦C makes the sponge “jam” to retain its deformed
gns. We then used a dual-material 3D printer to fabricate those wrapping-based origami structures with the required mechanica
nt 4 degrees of freedom (4-DOF) control of the liquid robot in planar space. Based on the planar 4-DOF control, the liquid robot
ure of the problem. In this study, a simulation-based design optimization approach is proposed to address this challenge. The pr
rns and input loading on the soft actuators’ deformation. Theoretical guided design of lateral-climbing soft robots and explo
etons, which exhibits a high extension ratio and low input pressure. With the antagonistic actuation of tendon-pulling and air-pu
n. The design can be directly 3D printed using a multi-material additive manufacturing process without any assembly process a
S score patients. Modelling and testing revealed extension torques of up to 2.7 Nm at the MCP joint and 0.67 Nm at the PIP an
size of a human finger, it can twist 5400 at an air pressure of 300 kPa. In general, SFTTA based actuators can twist 9.6 degree
g fluids. This study introduces a soft retraction mechanism that depends exclusively on the existing pneumatic force, eliminating
e environment. Here, we propose a plant-inspired method of steering by asymmetrically lengthening one side of the vine robot t
ed. In addition, a multifunctional manipulator that is built by integrating three modules in a series is demonstrated. A specific arc
ure, the manipulator can regulate it’s posture under greater loads of up to 1kg at the end-effector, with reference to the flex
nt localization, we propose a Recurrent Multi output Network (RMN) for predicting 3D contact points under continuous contact c
ignificant control over sensor and lattice design, while the pressure-based internal sensing provides measurements with minima
an interestingly be used in conjunction with FEM. FEM is being used to train initially the CNN model and for the control of accur
botic gloves, and has high flexibility, high output force and good fit. In addition, soft robotic gloves generally lack the ability to se
ssurized to deliver impact forces. We explore this inherently low-cost and simple design and its potential to replace current syst
ations in stiffness resulting from the different sliding states. The stiffness characteristics of the structure were validated through
rther characterize the OSM response within a fully-assembled soft robotic retractor to both folding and the application of forces
error of 3.08 mm, that is 7.7 % of the robot length. The robot’s functionality was shown in validation experiments, including
The optical waveguide sensor was embedded with carbon fiber to limit the stretching of the waveguide, which led to the optical w
on of local deformations in the soft body. These measurements inform a finite-element- method (FEM) based model to then es
sed to replicate motion from tactile demonstration only. To evaluate this approach, we explore the performance on increasingly
iscoelasticity into consideration. Further, a state observer is introduced to predict the unobservable viscoelasticity. Then, an En
ut compromising the softness of the device. This work presents a visual deformation-based force sensing framework for soft co
actuator can be increased to higher than three times its original stiffness by jamming. In addition to jamming, the proposed actu
œgrowing” and steering toward either a light or heat stimulus. We present basic modeling of the concept, design details, an
robot's geometry within any predefined regular shape, we employ B-splines to allow generation of freeform boundary surfaces,
a thermally-activated reactor incorporating a biocompatible hydrogel valve that enables control of the biochemical reaction of s
nges driven by a pulsating conductive fluid. The sensor demonstrates good repeatability and distinction of single and multiple p
re side, this is matched with exchangeable control software modules based on the FMI standard. Those can be interconnected
rove resiliency and elasticity. In order to facilitate involved DevOps processes, we present and release a tooling suite for autom
ments on real and simulated robotic platforms demonstrate these advantages.
machine shops. All benchmarks are accessible through cobra.cps.cit.tum.de, a platform to conveniently share, reference, and
he performance of existing GSL algorithms. GSL-Bench features high-fidelity graphics and gas simulation, featuring NVIDIA's I
LM), we present Cook2LTL, a system that translates instruction steps from an arbitrary cooking recipe found on the internet to
novel system that uses ChatGPT to automate the process of programming robots for adaptive assembly by decomposing comp
reduce the total execution time w.r.t. current approaches leveraging a single arm or multiple arms with synchronous motion pla
altogether. To overcome these limitations, domain specific extensions must be developed and integrated into the software pipe
ainty Based Image Super-resolution - can enhance lunar surface imagery by improving their resolution by a factor of 2, outperfo
complex control dynamics of floating platforms. Unlike traditional control methods, this PPO-MPC approach learns from MPC p
rol and demonstrates the effectiveness of the approach when applied to two different tether-net systems, with 4 and 8 maneuv
lter tracks the target's relative orbital and attitude motion, and its accuracy is ensured by robust on-ground training of the NN us
e-grasp success. For effective learning of the pre-grasping approach task, we developed a reward function that offers the agen
en explored in the past and shown to reduce the domain gap between simulations and real-world scenarios. Event sensors hav
y surpasses traditional window-based hazard analysis by resolving sub-cell size obstacles and terrain gradients at the individua
quares state estimator to determine the full state of the spacecraft. We validate VINSat in simulation using real nadir-pointing im
novel estimator is used for the unmeasurable modal states. The significance of accurate SYSID in controller design and perform
tweight design. Additionally, a compliant docking method is provided to generate lower contact forces along the Z-direction. Fur

dded softness of its shell adds new unmodeled dynamics into the system that impair commonly used control schemes. This pap
RL 4. Docking under loads of up to 550 N has been successfully tested with the new design. The experiments presented includ
y bounce away from the surface. To solve this issue, we propose a simple passive mechanism and a control method of a robot
he variable wheel diameter ellipsoidal robot (VWDER), which combines the characteristics of two-wheel differential driven robot
cheduled controller. However, its design for spatial SMS is not trivial and has not been studied yet. Therefore, the aim of this w

n model of the manipulated object, which estimates the robots’ movements by monitoring the displacement of surface point
Experiments performed on the da Vinci Research Kit (dVRK), an open research platform based on the clinically deployed robot
sessment Working Set (JIGSAWS) dataset. Our model outperforms the state-of-the-art (SOTA) with 89.5% accuracy for gestu
cessfully learnt a policy for the needle pick-and-place task, whereas the state-of-the-art TD3+HER method is unable to achieve
main randomization that allows for robust sim2real transfer, and a recurrent memory layer that enables reasoning about random
ith 3D computer graphics software, has been shown as an alternative to minimize annotation costs of 6D pose datasets. Howev
rgical automation, but their effectiveness relies on extensive data acquisition due to the absence of prior knowledge in success
PT-based architecture to automate surgical robotic tasks. The key to our approach is developing a goal-conditioned decision tr
in two main aspects through theoretical analysis and experimentation. Firstly, the extensible segment enables the bronchosco
shaft detection algorithm and probabilistic models. We apply our proposed approach in both a structured environment as well a
t based on optical flow from consecutive frames. To accommodate appearance change and alleviate drift, we then propose an
n real surgical tasks, the real-to-sim gap needs to be properly compensated. In this work, we propose an online, adaptive param
natomy via a graph search algorithm, A* search, with a cost function that enables the trajectories generated to balance betwee

m a total of 28 subjects to train activity recognition models at the gesture and motion primitive (MP) levels and for separate rob
models in simulation, demonstrating that the trained policies can complete the CCC in 85% of cases. We also show that applyin
earning and imitation learning algorithms to facilitate study of robot learning to augment human surgical skills. ORBIT-Surgical
manifold. Then, the recently developed Reproducing Kernel Hilbert Space (RKHS) algorithm is introduced to overcome the ``bi
p reinforcement learning (RL) has been investigated as a solution. However, hardware validation has remained a challenge due
To accomplish this, we employ a modified Kirchhoff rod model for each tube (with offset stiffness center) and embed the constra
uiring prior knowledge of tissue properties or needle deflection shape and amplitude. Our optimization process includes a repro
c location sensing and magnetic actuation. The proposed catheter uses a magnetic tip to enable magnetic steering. In addition
inverse mechanics together with endoscopic visual servoing lead to autonomous valve positioning. Benchtop experiments and
s-dimensional registration between 2D US and 3D CT, a decoupled two-stage registration approach combining initial vessel stru
usly learns and updates a very abstract model of the actual system, in the form of a dictionary of three independent key-value p
uce also a minimalistic fully onboard obstacle avoidance method, and assess the flexibility of our method by introducing an obst
is paper, we introduce M-ARK, a system to support studies on multi-swarm interaction. M-ARK is based on the synchronisation
sical interactions among robots. Each Morphobot has a changeable shell. It is functionalized for programming local interactions
ests demonstrating phototaxis and neighbor-detecting behavior. This paper also describes the method for manufacturing the co
ocal interactions among agents that affect each other's motions over long and short distances, we use non-reciprocal coupling
f using a small group of robots—the shepherds—to coordinate a relatively larger group—the sheep. In our study, the group
ual drone on a commercial, desktop machine along with a sensors and communication plugin on Gazebo Sim. The plugin trans
ation of motion planning algorithms to reach point goals within explored locations and object navigation policies for moving towa
algorithm that combines stream-based motion planning with an efficient, constrained task planning strategy. We validate COAS
al formulation that is based on dual quaternions, but adds more tools for geometric modeling of the dual-arm tasks. We also pr
hmarks, we demonstrate the efficacy of our approach, showcasing superior performance in comparison to state-of-the-art tech
onal obstacles containing geometry information for efficient motion planning. For exploration and simultaneous trolley collection
e a preemptible computational sub-problem of a priori unknown compute effort. Infinite branchings represent an infinite choice o
g the scarcity of available gas measurements. In this work, our goal is to push the boundaries in terms of continuity in sampling
hability graph can incorporate environmental information (obstacles) to provide the planner with sufficient geometric constraints
guide the search, we introduce heuristics to generate and prioritize subgoals. We evaluate our approach on various manually d
is shown with a realistic application.
Information(HAPFI), effectively leveraging the historical data from diverse modalities that agents collect while interacting with th
ulator arm, and we demonstrate its ability to learn about the place-dependent accumulation of garbage items. Case study also
olution space obtained by introducing minimally necessary coordination-induced constraints and not relying on prevalent simplif
ws and precedent constraints, its extension to multi-robot scenarios, and a method to quantify plan robustness. We demonstrate

an and it is enhanced with a probabilistic model that estimates the probability of sampling a new robot configuration towards th
ased on Monte Carlo Tree Search (MCTS) where each robot iteratively selects the next sampling location through communicat
uccess rate and energy usage. We have implemented the algorithm and tested it in both simulation and physical experiments. T
able. In order to broaden community engagement, RAMP incorporates fixtures such as April Tags which enable researchers to
s such as screwdrivers near human operators, the knowledge of injury severity needs to be extended so that robot integrators
e and formulate the notion of a strategy's affordance, which indicates its predicted benefit to the solution, and use it to guide the
ly transforms FOON into PDDL and leverages off-the-shelf planners, action contexts, and robot skills in a hierarchical planning
M task tree outputs into a graph and perform a task tree retrieval to avoid questionable nodes and high-cost nodes to improve p
agents. Subsequently, tasks are reallocated among all agents based on this solution and the solutions are iteratively optimized
e (RRT) and use greedy algorithms for task allocation. However, the classical RRT algorithm is a fixed growth step, which leads
ck on the human side. In this paper, we propose extending the "Marionette" interface by adding a wearable haptic interface to c
nfidence-weighted versions, WHSMM and WTP-HSMM. We evaluated various weight metrics that serve as teleoperation trans
to reproduce in a natural and intuitive way the manipulation task. A model of the smart servos is derived, exploiting the feedba
mmunication perspective is not investigated in the current literature, although communication imperfections and resource limita
teleoperation. In this study, we developed an immersive, bilateral teleoperation system, integrating Digital Twin-driven Mixed R
oposed using a JND and a rate-of-change threshold. An offline estimation method is also proposed when the geometry and imp
e effectiveness is demonstrated by conducting user studies of drawing reference shapes using the implemented robot avatar co
efficiently, still limits performance due to narrow workspace and insufficient kinesthetic feedback output compared to requirem
Primitives, which additionally allows adaptation from user commands to improve the shared control skills, during both design an
pose, such as a virtual reality controller or a finger-mapping hardware controller. To evaluate the efficacy and user-friendliness o
er limbs motions experiments.
ace context information (e.g., user preferences, object to place, and current scene information), our proposed method outputs a
into the environment, although these techniques may sometimes compromise the perceived realism. The experimental evaluat
Co users complete drone piloting tasks 32% faster than with a traditional remote control and report less frustration in a subjectiv
erator's view and using the same interface for direct, shared, and fully autonomous control. In a human subjects study (N=19),
rall accuracy. Furthermore, we propose a multi-window method that assigns proper hierarchical prediction windows of input dat
eting strategy for manipulation tasks. Additionally, the paper discusses the role of whole-body haptic feedback for wheeled hum
 tes it autonomously in 3D. The AI component of 3D Autocomplete was presented in our previous work, where we introduced a
d find the critical role of aggregation in fine-tuning performances. Inspired by such discoveries, we propose a novel dynamic agg
tion to improve the performance of rare objects. Specifically, we develop Valid Ground-based Insertion (VGI) to rectify the imba
n paired data, we propose a novel approach for learning the mapping functions between state and action spaces across domai
ally, a lightweight PointAdapter module is arranged near target tasks to enhance prompt tuning for 3D point cloud understandin
main shift accumulation on multiple geometric spaces (i.e., 2D, 3D Voxel, BEV) makes BEV DA even challenging. In this paper
4-DoF top-down grasps, our evaluation model and optimization procedure demonstrate generalization to 6-DoF grasps and nov
cy on a virtual agent over time can be decomposed into different frequency bands, which can be progressively learned via a se
policies trained in simulators to physical systems. In this paper, we treat the reality gap between a physical environment (target
e of AI robots. The core concept of the system involves the utilization of various wearable sensors, such as attachable IMUs an
pe adaptation to uneven ground. To verify the effectiveness of the designed foot, GRF measurement experiments were conduc
s been developed, which could provide operators with additional limbs to perform tasks beyond their capabilities. Compared to
d a flexible suction chamber. By analyzing the leaked airflow of the climbing robot's suction units, we propose a pose control m
hree and adopt the A-star algorithm. With an assumption that the relative posture between the end effector of the manipulator a
' low gear ratios. The system features a compact but high-payload tool-changing system, allowing the robot to efficiently use cu
e Line-of-sight between devices. All these limitations are addressed by our new solution, Mono Wi-Fi Scene, which requires onl
erface for Telepresence (Co-RIFT) and begin working autonomously, yet an operator can still take control when needed. Throu
embly, and part fixation poses common challenges. This project proposes a vision system which helps users to manipulate obje
ypoint estimation, 3D human mesh represents closed 3D surface shapes, aligning with pixels in 2D images while having measu
ed with object classification, voice command recognition, and hand tracking systems. The system was tested on a collaborative
ed on a single depth image by shape completion and grasp prediction (in
hanges. In a departure from existing work, we propose a novel architecture for ad hoc teamwork that performs non-monotonic lo
ed by the sensor property. For example, the lighting condition is often poor in the underground car park, and the LiDAR is not po
to move freely underwater. We built an underwater robot simulator using Gazebo. By utilizing the simulator, we can efficiently i
ct detection algorithm, determines the target's position in sonar images. Combining the particle filter algorithm with the detectio
e the mechanism to form a torus gripper, and conducted experiments to show its effectiveness.
s. This system, named the Music intervention Using Socially Engaging robotics (MUSE), has the social robot, Pepper, guiding p
nd experienced operators, because of the complexity of the procedure and the steep learning curve. Moreover, both the patient
ottle and delivers it to the user. The software was based on the Robot Operating System and utilized some of its solutions for th
nstrations to auto-tune the parameters in the DMP framework. This method is also valuable for using in conjunction with Reinfo
one foot while using the other as a pivot. Due to the robot's low weight and small foot size, limitations exist in ground-foot static
s of their respective parent companies. They must be capable of determining the value of services, engaging in negotiation, an
s detected by an inexpensive optical flow sensor, allowing quick grip force adjustments to stabilize the grip. A full shared contro
nment works as a server and the robot cell as a client. Pose of the virtual cube is updated as the physical robot checks the wor
ssive actuators and structures that mimic animal limbs. EEWOC overcomes these limitations with an unconventional mechanica
ementioned goals. While the system is designed to run autonomously, it can also be teleoperated through a variety of multi-mo
ded soft robotic gripper. In the first part of the video, we not only show that the gripper can perform nondestructive sampling tas
0 body lengths per minute on flat terrain, effortlessly maneuvers through diverse terrains, including ice, sand, grass, and rocky
t half, is the world's first device to achieve tip growing motion while performing active directional steering, presented at ICRA20
gle regulator weighs only 35g. We demonstrated that the buoyancy regulator system can switch from ascent to descent in 20s.
 ies-Parallel-Reconfigurable configuration more practical. The transmission adopts the tendon-driven ways with low friction, goo
ural movement trajectory and muscle activation patterns can be disrupted, leading to discomfort and reduced performance. In t
ask model and interaction method based on FSM. The approach shown in this video attempts to apply mixed reality to wearabl
e then utilized to select the most promising plans. We validate the performance of our Ensemble LSR (ENS-LSR) on simulated
and fisheye cameras. Our implementation succeeds in tracking objects when the camera speed is reaching 2 m/s.
l planes and shape priors. However, these sources are not always available. We thus propose a widely available new source o
bot under larger environments where the cameras may be moved around without extrinsic calibration, or maybe when in motion
d achieves state-of-the-art accuracy and time efficiency among query-based methods on multiple multi-object tracking benchma
is end, on the trail of SORT, we propose AgriSORT, a simple, online, real-time tracking-by-detection pipeline for precision agric
2) these approaches do not allow for dynamic changes to the number of UWB anchors in a tightly coupled estimator. Our appro
pe bias and use it to formulate a maximum a posteriori problem for further refinement. After this refinement, we proceed to upd
is gap, by GPU-accelerating robotic volumetric mapping. Nvblox delivers a significant performance improvement over the state
n i.e. when the goal becomes visible. Crucially, an exploitation module teacher-forces the exploration module and continues driv
s, and wheel-terrain interaction. We theoretically demonstrate that the proposed T&R system is globally stable and robust while
hat our proposed method can achieve an improvement of 1.38-13.38% in terms of text-to-image recall on different benchmark
s. A machine learning model is designed and trained to jointly achieve multi-object semantic keypoint detection and classificatio
redictive models and estimate the mutual information between model weights and outputs, facilitating dynamic horizon planning
ps. To tackle novel visual observations in unseen environments, CMN learns a deep perception model that maps the visual inpu
ker markers using blinking light-emitting diodes (LEDs) achieve long-distance recognition, they are constrained by the number o
efficient NeXt1D convolution blocks with a fully recurrent structure based on feature pyramids to decrease computational costs
e names as reference landmarks, ensuring accurate turns and stops, especially in unfamiliar places. To address this insight gap
of the topological navigation pipeline. This makes subgoal selection more efficient and enables leveraging large-scale datasets
Knowledge Graph with Visual Perception (AKGVP) method for object-goal navigation. Technically, our approach introduces con
on tasks, and traditional map-based methods that neglect visibility constraints. Our experiments, involving the first phase of the
ation task. It is straightforward for recent foundation models to indicate an object by pixels, and with pixels as the goal specifica
nner. GeoAdapt uses geometric consistency as a prior to improve the robustness of our generated pseudo-labels against doma
lanning approach that generates local plans based on geometric and semantic information. The system is trained using the Imp
(IR) and (2) training the navigation policy based on human-labeled IR. By training the navigation policy on IR instead of raw da
eferences for visually novel terrains, which the robot should adhere to, can often be extrapolated from established terrain prefe
te local detail optimization, we propose a multi-patch perception pyramid block that models the correlation between different siz
arance similarity, to yield instance-unaware candidate keypoints. Then, the entire graph with all candidate keypoints as vertices
n from multi-view images, and employ volume rendering techniques to establish 2D renderings, thus enabling direct 3D superv
cost, and two gated recurrent units iteratively improve the results by tracing historical information. In this way, our network is a g
g. Incorporating semantic priors with self-supervised flow training, Doduo produces accurate dense correspondence robust to th
pecific encoders. Specifically, the large-scale pretrained knowledge in MLLMs allows for easier fine-tuning to downstream robot
fective comprehension of point cloud video. We also propose a multi-level contrastive approach for both modalities. Through ex
al-flow-based mask designed for dynamic object removal, (3) a self-distillation training strategy to augment the training process
is. In this paper, we present a novel method for video-to-events stream conversion from multiple perspectives, considering the
h tasks. To this end, we propose attention-based temporal fusion for multi-object 6D pose estimation that accumulates informat
detail, we design a novel two-stage training strategy, incorporating Low-rank Adapters and Convolutional Adapters, which not o
oint strategy to enhance the distinction between known and unknown classes in the feature space. Additionally, we address the
 s hindering both the speed and real-world applicability of the 3D recognition process. Our proposed method leverages a multi-s
on network that fuses the textural and structural information from image frames with the high-temporal-resolution motion inform
ons (Neural ODEs) to model the complex dynamical system depicted by videos as a nonlinear ordinary differential equation. Th
verages generative adversarial networks (GANs) to fuse salient image components from the two modalities. Our network comp
hat automatically adjusts minimum parameters of the 360â—¦ Lidar 3D bounding box to perfectly fit the image appearance of pa
y running evaluations on six benchmark datasets, both static and dynamic, indoor and outdoor, with synthetic and real images.
g of these models demands substantial data and computational resources. Hence, exploring methods to fine-tune these advan
ning NeRF significantly improves the novel view synthesis performance of NeRF and reduces computational overheads. Throu
ased sampling strategy for robust and efficient learning, which not only accelerates training procedures but also conduces to ex
their widespread adoption across diverse applications. Although single-beam LiDAR is more affordable, its limited depth sensin
hanced Outpainting (NEO) that uses extended-FOV images generated through NeRF to train a scene-specific image outpaintin
sues. Specifically, we present a novel representation for $SE(3)$ which follows the property of smoothness of the pose. We the
longside an auxiliary objective based on self-supervised learning. After experimenting with various objectives, we introduce Cro
Master-Slave Relationships (HMSR) among objects and hands, bimanual coordination strategies, and sub-symbolic task repre
e rendering quality while simultaneously reducing training time.Additionally, we leverage sparse point cloud to aid the model in b
wly identified targets while the latter is computationally expensive. In this work, we present Bag-of-Views (BoV), a fully appearan
st occlusion-resistant pattern. The CopperTag detection process relies on three fundamental steps: firstly, extracting all lines fro
inherent geometric patterns within the perceptual data of various agents. Following this spirit, we propose a robust collaborativ
ong scene representation capabilities of NeRF, our method can render high-quality novel-view images with waterdrops removed

we directly solve the pose in one step by building 2D-3D correspondences between target and initial view, thus allowing for rea
autoencoder trained on a nonoccluded human skeleton and exploits the symmetry and angular information of the skeleton. The
supervised depth estimation methods. Specifically, we associate the spatial locations of moving objects with the temporal order
ntroduce an innovative approach that harnesses the strengths of both modalities to dehaze and enhance visibility in a single rea
by exploiting unlabeled images during the training. A consistency loss that makes full use of unlabeled data is then proposed t
ations for vision and point cloud data. Our system integrates the chain-of-thought capabilities of Large Language Models (LLMs
veloped on SAM remain constrained by the inherent limitations of RGB sensors, particularly in scenarios characterized by comp
econstruction compared to state-of-the-art feature extractors on conventional and burst-merged images. This work opens aven
s and data distribution make it challenging to directly apply transfer learning from open-source RGB to target spike data. To this
ctive pillars used for point generation and predicts the center of points and the number of points to be generated for each active
on horizon. Especially in a reconfigurable manufacturing work cell, scaling and calibration of a photogrammetry system require
mproves Cobb angle measurement accuracy by integrating multiple spine morphology as attention information. In the MMA-Net,
as potential domain gaps and a less pronounced intra-class variance. Our approach to address these challenges, using path tr
-stage method for pose estimation of reflective texture-less metal parts, which accurately estimates the target pose using mono
tion and orientation. The PZT is mounted on the end of a soft robotic arm that has three individual degrees of freedom and can
ch save time significantly for collecting and labeling. We implemented real grasping experiments with a robot arm UR10, 2-finge
, which is trained using only simulated data and seamlessly applied to real-world scenarios without additional adaptation. To en
n easily be deployed on embedded AMR hardware. To evaluate our approach, we create new large-scale synthetic datasets sh
per we propose RIDE, a learning-based method for rotation-equivariant detection and invariant description. Following recent ad
up. However, prior studies overlooked two vital problems of the surgical domain: i) large domain shifts from diverse surgical ope
ception method named 'Multimodal Surgical Process Recognition based on Domain Knowledge and Segmentation (MSPR-DKS
 depth data. We conduct several experiments on two different physical robots. The Panda 7DOF arm grasps a static target in 3
ary to grasp foods without damaging them by using external information such as images. In our method, to tackle these problem
demonstration graph in the inference case, we utilize a similarity function that helps select the best demonstration for a specific
mainly focus on image-based visual servoing, which deals with feature errors in the image domain. In this work, we propose an
stness in handling complex scenarios. This approach holds promise for various applications, such as surveillance and autonom
igid-As-Possible (ARAP) deformation model. It requires no mechanical parameter of the object to be known and can drive the o
o prompts using a darkness clue prompter (DCP). Specifically, DCP iteratively learns emphasizing and undermining projections
nd occlusions. Complementing this, we introduce a model-free controller tailored for resilient quadrotor visual tracking. Our syst
unreliability of the predicted positions and feature embeddings of the objects. This paper proposes DroneMOT, which firstly pr
ts plausible object tracks that are resilient to missing, noisy, and kinematically or dynamically infeasible measurements. Throug
, segmenter, and dense optical flow estimator, into a model that tracks and segments any object in 2D videos. Given a monocu
y, in each stage, self-attention is first applied to each branch separately to capture the non-local context information. Then, cros
state-of-the-art supervised models with self-supervised training. We find that when the initialization is supervised weights, mos
ating all possible track hypotheses across the temporal sliding window. A novel graph optimization approach is formulated to so
ding MOT. While there are existing works in probabilistic object detection that predict the localization uncertainty around the bo
ts while walking. It is assessed with experiments consisting in rolling and assembling in an unwinder a heavy and wide bobbin
The main innovations of LiteTrack encompass: 1) asynchronous feature extraction and interaction between the template and s
l constraints according to the parallelism between the quantitative windows. We apply an anomaly correction method to mainta
a singularity-free and differentiable variant of the single-track model to enable seamless integration as dynamics factor into VIO
d when constructing photometric errors, along with the addition of inertial information in the process. This approach enables the
rrier phase measurements, along with the visual-inertial measurements. Inherited from the SRI-SWF, the proposed SRI-GVINS
front-end local tracking) as input and predicts multiple depth bases along with corresponding confidence as output. The final d
loosely-coupled iterated Extended Kalman Filter (iEKF) which fuses together the visual odometry running fast at 300 FPS with
e the point-based and line-based visual residuals from the event camera, the point-based visual residual from the standard cam
ucted on 12 sequences of varying lengths from the EuRoC and TUM-VI datasets, demonstrated a speedup of up to 4.23x in the
ont-end tracking and back-end optimization module designed for monocular or stereo setup towards multi-camera systems, and
itionally, due to the limitations of existing datasets for supporting multi-task detection in athletics tracks, we diligently collected a
Flexible Mechanism (HSFM). The HSFM can change its grounding point while being capable of lifting body weight with one mot
amics, the requirements of the tail can be greatly reduced. Further, it is shown that robustness of stance is significantly enhanc
s in active mode. The absorbed energy can be stored and later reused without compromising exoskeleton transparency in the

ovide customized, safe, and individualized assistance using both human demonstration and interactive learning. Specifically, th
e of structured fabrics that can adjust stiffness under various air pressures. Additionally, it includes IMU sensors to detect lifting
ole. A vector scaling process transforms the force values from the FSRs into the three-dimensional force vector, enabling conti
ve wearers during the exploration process. In this work, we propose a novel preference-based human-exoskeleton locomotion i
hieving this is a complex task due to the shifting center of rotation of the knee in both the Sagittal and Coronal planes. Previous
meter identification. To demonstrate its effectiveness, the control strategy is experimentally validated with five healthy subjects a
me lag seen when using L1-based approaches due to the presence of a low-pass filter commonly used in this type of approach

oposed strategy is experimentally evaluated. The results show that the robot can leap onto a step higher than itself and then sm
d to modify particle weights within the particle filter framework. Subsequent resampling ensures that the contact status converge
 lation was used to simulate an arbitrarily curved rock rolling on undulating terrain and to generate control signals to stabilize it o
ct switches and impacts, traction limits, and actuation bounds. By introducing an arc-length-related parametrization, the trajecto
ntial steering within the oscillator network. Three oscillator models are used as the basis of the CPGs, and their performance is
of Navier’s table problem, only nine-dimensional WFs projected onto the constrained space are identifiable among the twe
ot with an addition structural deformation is designed to balance inputs and degrees of freedom (DOFs), so that the over const
passes through zero. In general, our redesigned control law has no singularities, which can satisfy the requirement that the veh
orizontal velocity, the method uses the Variable Height Springy Inverted Pendulum model for continuously recomputing the fee
computes the manipulation forces and manipulated object task-based reference trajectory. Then, pose optimization aligns the
self-righting whenever possible. In this work, we present a series of challenges that outline why recovery in the wild is difficult.
We have demonstrated the usefulness of the proposed method in environments with low friction coefficients with the environme
nd stabilizes the robot in an upward position. Optimization of centroidal angular momentum is fully integrated into the proposed
sentation is differentiable and smooth in both task and joint spaces, enabling its direct integration to optimization problems. Add

body pulling movements that have not been previously demonstrated using a learned policy. It also facilitates a characterization
n this work, we explore how whole-body bilateral feedback with haptics affects the performance of different shared-control meth
ater Text (UWT) for noisy underwater scenes to establish an important case study. Moreover, we also design an efficient super-
oint cloud representation across domains. Moreover, to learn domain-specific representation and transfer knowledge from sour
robot followed a person to test our framework. During the test, the targets were asked to change their appearance and disapp
compatibility graph of two point clouds represents a robust fitness score measuring their spatial consistency. This score takes
beddings, and only image-paired data is sufficient to bind all action modalities together. Our Action-SGFA can leverage existing
haredConv) that significantly compresses the model and enhances its ability to capture informative features. In PS-VLADs, we
m negative samples. In addition, we propose a fast coarse-to-fine scale search strategy, which alleviates the challenges in estim
e introduce ZS6D, for zero-shot novel object 6D pose estimation. Visual descriptors, extracted using pre-trained Vision Transfor
vercomes the limitations of most action recognition models with the transformer structure using the duplex attention function, fl
tive. The proposed prototype can also improve transparency in human-robot interaction. Also, an accurate torque controller with
re (RMS) and peak load torques for a specific pick-and-place task can be reduced respectively by up to 43% and 44% for joint
paper introduces a mechanically compact design for a novel lever-cam-based variable stiffness joint.
y. Moreover, the continuous linear stiffness profile of the actuator can be customized by solving the transcendental equation of t
n the form of a 1-DoF physical prototype, which is implemented using a modular testbed. We present a hybrid system model, a
acteristics, and crucially, suggest that non-pinwheel designs of cycloidal gearing can be a lower complexity and cost alternative
ective optimization for parameters using Genetic Algorithm is then conducted to obtain the trade-off design. The verifications for
een the two motion modes and possible adaptations of the original design are discussed.
nergy to perform a cyclic task. Variable stiffness springs are a special type of spring with customizable force-deflection behavior
is able to obtain collision-free paths for the UAV and the UGV, taking into account the 3D environment and the tether. Then, th
the mission UAV and the GS. Specifically, to alleviate data storage pressure, a terrain lightweight modeling method is employe
easurements, both as functions of the horizontal velocity of the UAV. The learned models enable the online planner to estimate
rs introduces a new way to recreate orbital motion within an indoor facility, considering the case study of trajectories derived fro
y explore performance-cost trade-offs of the fine-tuning phase along three axes: textit{i}) dataset size (more data increases the
g snap minimization in the differentially flat output space, a computationally efficient algorithm, suitable for online motion plannin
et, data noise handling, and lifting functions selection, are elaborated in pursuit of the theoretical optimality in real applications.

on a novel millimeter form factor 64 pixels multizone Time-of-Flight (ToF) sensor and a generalized model-free control policy. In
uipped with a RealSense D415 camera, closely emulating the MyPaint virtual environment. Our system can perceive its environ
hm with our handmade robot platform in both simulation and real-world environments.
o onset response of feasible plucking motions while minimizing invalid attempts. The resulting models can be used to actuate a
such courses and curricula is effective for sparking student motivation and conveying the unique challenges of programming em
ned with more than 15 minutes of data. In dense real-world flights with two vehicles, deploying our model online improves 3D tra
we call ``latent''. Imprecise modeling of these effects can result in degraded control performance, particularly when latent distu
e overall efficiency and effectiveness of wide-area visual detection systems. Specifically, in this study, we introduce a wide-ang
ero-shot manner, which overcomes the over-segmentation limitation of the current SAM model. Second, we proposed the distil
enerated PDs via error message feedback from the symbolic planner. Our aim is to answer the question: How accurately can V
alman filter, can be efficiently reused to compute the MHE Hessian. This offers linear computational complexity with respect to
mposition and reconstruction. Based on the strategy, MCDEN is more interpretable than most neural network methods since in
c structure of the object determines how we can manipulate it, we propose a kinematic-aware prompting framework that promp
mework operates through three sequential steps: preprocessing, planning, and iterative self-refinement. During preprocessing, a
work for odor detection and identification. The developed algorithm is rigorously tested on a dataset for sensor drift compensatio
ted with interpretable edge features that characterize the interactions. Moreover, we aim to enhance model interpretability and
e data imbalance between static and dynamic points. Evaluations on the Argoverse 2 scene flow task reveal that DeFlow achie
environments, we highlight the merits of our approach. Compared to current leading methods, CoordiGraph notably enhances
hically compliant policy. Next, we extend an existing ethical framework for prima facie duties to belief space and offer an ethica
arned skills to novel tasks through the comprehension of task-specific prompts. This empowers the robot to perform the specifie
subsequently analyze the learned functions. Through nonlinear function approximation, we uncover meaningful insights into the
nd it empirically achieves up to 11x faster detection on realistic robotics settings compared to prior work while maintaining a low
t constraint, before expanding a complete set of sampled future observations episodes and without any loss in accuracy. Moreo
ghly evaluate their advantages and limitations on two use cases: simulated autonomous driving and active perception with a ph
curate representation of the desired posterior distribution in high-dimensional contexts. Empirically, the effectiveness of our app
at incorporates human-computer interaction methods in deep learning security testing. By considering the attributes of model ap
esults show an accuracy of more than 95% in the expected robot’s functioning. The participants judged the system efficien
and by the cutting board. However, training the robot in the real world can be inefficient, and dangerous, and result in a lot of f
extracted from a cluttered scene during the observation. These objects are first coarsely organized and depicted within a scen
meworks that are also able to reason about the complexities of multi-robot coordination. In this work, we compare the task succ
age instruction, thus aiding the policy network in mastering actions for versatile manipulation. Additionally, we present a feature
arch-friendly system, we augmented it with a 2 degrees of freedom kinematic chain to control the goalkeeper rod as an initial se
the current user's instruction, and 2) a slow-thinking system that is comprised of a fine-tuned vision-language model aligned w
ot clear how to document robotics tasks and naively providing full robot APIs presents a challenge to context-length limits in LLM
aining, but also for a refinement of the subtasks and their interfaces in response to challenges arising from discrepancies betwe
an extensive simulation experiment to compare several common replanning strategies and confirm that effective strategies are h
ific types of objects (e.g., door numbers), referred to as semantic anchors, that help to distinguish similar scenes better. In setti
gh success rate on the real robot task of chair preparation. Also, it outperforms several baseline methods on the task of upright
generate 3D assets for simulation by lifting open-world 2D object-centric images to 3D using image diffusion models and query
osition and memory management into LLM planning inference, which could effectively improve the faithfulness of LLMs and ma
the incorporation of latent space simulation for improved training and policy debugging. We use pixel/patch-aligned feature des
nd-motion planning (TAMP) algorithms, making factorization into subgoals untenable. Rather than using LLMs to directly plan ta
ments, and deploy it zero-shot (without any real-world experience or fine-tuning) on the Boston Dynamics Spot robot in eight no
ch combines fast filtering during data transfer to the robot’s LTM with slower yet more precise forgetting mechanisms that a
vironment representations. The latter, leveraging the proposed spatial representation, further enhances the collaboration comb
ms used for enforcement to be customized. Our evaluation shows THEMIS-compiled autopilots can adhere to regulatory constra
sized wrench commands. Additionally, by exploring the dynamics of over-actuated UAVs, the complex planning process is deco
seless environment is reasonable and that the optimized trajectories offers visible reductions in positioning uncertainty in the p
mulations using the PaparazziUAV software.
multicopter only needs to keep the rotor thrust to actively straighten the tethered cable and a constant attitude similar to that of
asks such as segmentation, depth estimation and depth completion, among other tasks required for autonomous drone navigati
lization method. We provide a thorough evaluation of the presented methods and demonstrate that the regressing approach to
opts an incremental sampling strategy based on the probabilistic roadmap constructed by heuristic sampling toward the unexpl

ery of Regions of Attraction in a learned Latent Space (MORALS). The approach combines auto-encoding neural networks with
se, as well as through hardware experiments on a 27 gram quadrotor, demonstrating high-speed trajectory tracking and dynam
showed that oocytes were successfully released one by one with no deformation in shape using the movable microfluidic chip.
e intervals to generate a generalized prediction error for parameter update, such that excitation information of regressor channe
use. The experiments demonstrate the potential of Sim-on-Wheels to facilitate the process of testing autonomous driving in cha
atasets; It penalizes the generation of unlikely scenarios. Our approach overcomes the dimensionality challenge and explores a
ng different views of the target vehicle. During the pursuit, the target vehicle is firstly localized using a Convolutional Neural Net
measured by most-likely FDE and KDE, as well as improved performance on high-speed vehicles, and vehicles within high-risk
e issues, we propose using human feedback for alignment and employ RLHF due to its sample efficiency. We also introduce th
orldwide standards such as ISO 61508, which defines standards for safely deploying robots in industrial factory environments w
n. By fusing both the road topology and trajectory histories of neighboring CCVs, our model learns an interaction score for each
POP achieves comparable results to top-performing methods in open-loop experiments and outperforms the baseline method in
that implicitly extracts the potential future information in an intermediate feature-level, and identifies the interacting entity pairs
entation of roads. This design enables scenario retargeting to various road structures, traffic conditions, and ego vehicle behavio
where agents broadcast queries for unknown conditions and respond with appropriate knowledge using a condition-action-contr
hen necessary, and applying the Hungarian method for task allocation. A local planning algorithm is then applied to identify the
ed for a single prediction horizon. In comparison to a model predictive controller, the proposed approach exhibits competitive pe
baseline of the OWL-ViT CLIP on Wheels (OWL CoW). Furthermore, we study the usage of LLMs for robot navigation and pres
oration goal through the distance in the embedding to the previous samples. In this work, a variational autoencoder, a Wasserst
dynamically to control robots to perform adaptive tasks. Different from other LLM based methods for complex robotic tasks, ou
y DyHGDAT to a CVAE-based prediction system for predicting plausible trajectories. To validate the effectiveness of prediction
duce an evaluation metric for Driving QA and demonstrate our LLM-driver's proficiency in interpreting driving scenarios, answe
tion model that leverages the nature of hypergraph to encode the groupwise interactions among traffic participants. Moreover, w
ious state. Moreover, we suggest employing a downstream, resampling-based algorithm for additional outlier detection. Based
Voyager", which leverages infrared thermal vision to achieve true passive perception in autonomous entities. The system utilize
earning systems. Finally, integrating our findings, we propose a strong baseline model—PlanTF. Our results demonstrate that
The topological structure is inferred from traffic flows correctly and automatically. We highlight that this framework is highly auto
ting. It uses a transformer to aggregate various input modalities and facilitate selective queries on occlusions that might interse
tructed offline by detecting semantic objects, such as ground markers, lane lines, and poles, using cameras or LiDAR sensors.
e of the predictions is similar to the ground-truth, but allow additional unlikely predictions to occur, over-representing unlikely pr
o a tree-structured representation with a dynamic branchpoint based on the scene-level divergence. Moreover, to generate dive
ucted. Then, the scenario matching model based on spatial-temporal graph convolutional network (ST-GCN) is proposed to ma
g terrain to facilitate data-driven mobility; we also present algorithms and their experimental results to show that conventional w
ilding a completely new learning based social navigation planner. In this work, we find that this reaction is unnecessarily extrem
MPC) is proposed to achieve continuous and robust tracking of single, two and three targets in complex environments. Simulat
understanding tasks that rely on both the ability to interpret freeform natural language queries, and in grounding these queries
rically linked positional embedding. Specifically, the DualAT excels at processing and fusing information from multiple camera v
reinforcement learning method based on implicit Q-learning (IQL). When facing OOD observation, our novel localization recov
ctories, leading to realistic scenarios with low collision rates. Our experimental results show that UniGen outperforms prior state
point function (IPF). The basic idea of IPF is that when we are given low-resolution point cloud and a query ray, we generate th
desirably avoids annotating a large amount of data. Extensive experiments demonstrate that CAD outperforms baselines in ter
ediction and achieves state-of-the-art performance on common prediction benchmarks and a human tracking dataset captured
 rein each driver possesses its own policy distribution. Then, the post-trained drivers will be kept in the training loop of the RL al
d skills as actions and can be trained by any off-the-shelf offline RL algorithms. This facilitates a shift in focus from per-step act
ploys Uni-directional to Multi-directional Traversability Distillation (UMTraDistill) to distill a multi-directional traversability network
extensive areas. Hence, this study proposes a cross-modal localization approach using adaptive modeling, employing LiDAR fo
lization with robot navigation. This unified structure, underpinned by an end-to-end learning approach, resolves the inherent con
entation and graph policy, leaving the low-level motion planning problems to classical controllers for better stability and generali
rate. OpenBot-Fleet represents a significant step forward in cloud robotics, making it possible to deploy large continually learnin
development of end-to-end motion forecasting models, we augment the Waymo Open Motion Dataset (WOMD) with large-scale
The effectiveness of the model can be demonstrated using a large data stream with about 3,000 real data points. As a result, it
ure that learns to fuse pre-trained multi-layer representations into a separate network to learn a robust policy. Through extensiv
asets without overfitting. In addition, reliable task conditioning and an expressive policy architecture enables our agent to exhib
asks show that this framework is robust to distractors, controllable by language, capable of understanding complex multi-object
ts of prior experiences and self-improves online to learn a target task with minimal human intervention. Our insights are to utiliz
which retain high performance with a fraction of the computing cost. For this study we focus on the task of motion forecasting u
this work, we study how developers introduce physical unit mismatches by manually inspecting 180 errors detected by the softw
°-90°). The results of flight experiments confirmed the theoretical analysis that the flapping-wing robot can achieve self-takeof
ming results show that compared with ordinary flippers, the locomotion efficiency is greatly improved with the help of origami flip
deflection causes a yaw moment, the coupling between yaw and roll dynamics generates also a roll rate. To validate this simplif
aforementioned swimming motions with different amplitudes and frequencies. To verify the model, we built a multi-joint robotic f
igh water content, the reduced mud shear strength led to a large slippage of robot appendages and a significantly reduced step
ocomotory responses and hydrodynamic information from real schooling fish. In this paper, we present the RoboTwin platform
mation (ASE) algorithm to calculate the real-time swimming direction using filtered yaw angles at a high updating rate of 100Hz
nificantly improve the response of robot at different frequencies. Then we develop the TSCP by embedding shape memory alloy
mics-based self-adaptive mechanism is proposed to improve the self-adaptive performance of the swarm robots in dynamic env
on robustness? 2) What are the effects of using a memory-enabled vs. a memory-free policy network with respect to robustnes

evaluated the effectiveness of the shell in two experimental scenarios: passive righting – balancing the robot on each face of
optimize for directly on hardware with Bayesian Optimization. The force profiles are tracked at the joint level, and added to Cart
but still improves the performance of the platform, allowing it to enter turns at a higher speed without toppling. While the yaw ra
gait of approximately 0.6 m/s and a bounding gait of 0.7 m/s, with desired velocities of 0.8 m/s and 1.0 m/s, respectively. The
and high-precision control. A novel telescopic vehicle mode is also proposed for complex wheel-leg hybrid motion, offering optim
8.7 cm. Continuous jumping of steps and crossing of complex rugged surfaces is realized. Besides, integrated with the power s
erform many other standard burrowing techniques and was able to automatically reach targeted burrowing depths. We hope th
while offering wireless communication capability and sensors for attitude estimation. Our MAV has a total mass of 2.1 g, which
A series of experimental tests show that the robot can jump to a height three times its body length and a distance seven times i
aunching the microrobot into the air at speeds up to 3.171 m/s. At the same time, the microrobot is capable of crawling, when a
o characterize the drag forces acting on the vehicle and estimate the robot’s velocity. This 112 mg robot can reach maximu
ulsor. To improve maneuverability, we evolved the VLEIBot design into the 90-mg/47-mm3 VLEIBot+, which is driven by two pr
neural network, we demonstrate successful learning in both simulated and real forest environments, as well as successful hom
he flapping kinematics of dragonflies, the installation angle and wing pitch angle of the proposed DFWR are optimized consider
e terephthalate (PET) film was selected to leverage the combined merits of metal conductivity and polymer flexibility, light weig
 ed in the mBNNs by combining global chemical modulation (for suppressing spontaneous signal transmission) and local chemi
m and a minimum bending radius of 125 mm. In addition, we developed a Gaussian process regression model to predict an exte
ture to control these two morphing mechanisms. After that, we establish the theoretical kinematic models for both the foldable a
cally intelligent limbless robot for sidewinding. This robot features a decentralized bilateral cable actuation system that resemble
oM and efficient MoI-inducing, we propose combining a 1-DoF tail and a reaction wheel together for spatial reorientation (regula
g themselves into a 3D configuration, akin to origami. To further leverage the capability of generating a wide variety of different
he nematode-inspired cable routing method and compare it with other cable routing methods. Additionally, we derived and simu
ed-order state observer to estimate the temperature of SCPs given the kinematic measurements. Combining the designed obs
pace exploration, which enables accurate end-to-end optimization of the calibration process and eliminates the need for the lab
h mask. The point cloud attributes include intensity, normal vector and segmentation class. Experiments on different real-world
of stereo video sequences. Additionally, we emphasize the impracticality of using three Euler angles and three components in t
unction, and work with a small memory footprint and computational overhead. Both include a preselection of informative data th
ed overlap. Our method leverages long line features from surroundings, and filters out outliers with a novel convergence voting
ard which allows the lidar beam to pass through the white grids and reflect back from the black grids. This innovative technique
he plane motion of the autonomous vehicle are corrected, and an accurate extrinsic parameter is finally obtained. Specifically, T
ate the objects in the two data domains and extract pairs of 2D-3D points. Using the PnP solver, we get robust initial extrinsic pa
for both frame- and event-based cameras. However, using 2D patterns has several limitations for multi-camera extrinsic calibra
tems. Leveraging the 4D correlation volume and correlation lookup techniques inherited from RAFT, we apply them to correlate
neural network (RNN) with customized Runge-Kutta cells, obtaining labeled training data, predicting system responses, and es
optimization. To speed up the optimization process further, and reduce the number of simulation runs needed to find good para
by calculating a globally-aware map-to-map cost volume and hierachical point-to-map cost volumes. The former is used to regre
approach to two different robot-assisted procedures: Transcranial Magnetic Stimulation (TMS) and femoroplasty. We also discu
s, we derive the closed-form relation between joint torques and the object’s inertial properties. Based on the derivation, the
ccurate calibration parameter regression, etc. In this paper, to address these issues, we propose CalibFormer, an end-to-end n
ut the need for a calibration board or other equipment. Our approach takes advantage of the fact that when an event camera is
allenging to rapidly and accurately locate the target and extract the feature. In this paper, LB-R2R-Calib (Long Baseline Radar
ate that with appropriate domain randomization, this approach allows for successful sim-to-real transfer with no explicit pose es
n and react to perception failures end-to-end. To this end, we incorporate both proprioception and exteroception into our policy
nstance of a given target object category. We evaluate VLFM in photo-realistic environments from the Gibson, Habitat-Matterpo
his unified policy results in better overall performance when navigating to visually indicated goals in novel environments, as com
Sharing in multi-agent reinforcement learning. Last but not the least, we implement the proposed framework in online and offlin
GPU acceleration to process collision checks for thousands of configurations much faster than traditional collision detection libr
as motion predictions. By sampling real data and by building on scenario optimization and predictive control, the planner provid
he proposed integration of multimodal motion prediction and trajectory tracking outperforms other non-deep-learning methods in
omputational time. High-fidelity simulations are conducted to validate the proposed method's effectiveness and practicality.
gation in Dynamic environments that uses sequential Normalized LiDAR (LNDNL) scans. We employ long-short-term memory (
omial (QP) trajectory planning. The proposed method determines the terminal state and confidence of the trajectory using DRL

rmine a subgoal, we develop an evaluation method whose input elements can be efficiently obtained on the layered structure. W
e paths of various length in maps of variable size, and obstacle structure.
paper, we explore a simple formulation for a two-dimensional arm, which results in smaller SOS programs than previous sugges
m. However, two roadblocks arise for training a DRL policy that outputs paths:(1) The action space for potential paths often invo
rical tractability. The result is, when existing predictors are coupled with planning and control, one may often find unsafe motion
d is capable of discriminating the safety or lack thereof of two motion plans which differ by only millimeters.
e quality of reconstructed images. The results show that CVCDIC can achieve an MS-SSIM of 0.978 at 0.15 bpp, surpassing DS
urrent robot state to the selected subgoal can satisfy the given time-dependent constraints while being goal-oriented.
problem dimensionality by two orders of magnitude. The reference data for network training are generated based on algorithm
present proof for the forward invariance and the stability of the approach. We compare the efficacy of our method with the base
narrow passages, and navigate through non-solid obstacles in dense environments, we propose an approach to adaptively inf
iciently generates real-time navigation paths, avoiding untraversable terrain in pursuit of the goal. This combination of SGP-bas
ment it in two widely used estimators, the Kalman filter and the particle filter. Thirdly, we propose an active collision path plannin
nimum distances through iterative graph rewiring and updating relevant embeddings. The probabilistic rewiring enables fast and
ock, it bypasses grid-by-grid traversal within the block and directly jumps to the opposing surface. This method significantly redu
ncertainty intervals. We use conformal prediction to learn uncertainty-informed predictive safety filters, which make no assumpti
er of features in the image (corners) and point cloud data (edges and planes) separately to quantify their reliability to augment th
points and applying cubic polynomial interpolation to achieve path smoothing. Simulations demonstrate that the proposed E-RR
quirement to store previous data. Then, a CBF is synthesized using the proposed safety function to rectify the safe control inpu
annotations of common household objects. We demonstrate that fine-tuning a VLM on PhysObjects improves its understanding
n a deep neural network, named PRVNet, to predict the required number of views, allowing us to tailor the data acquisition base
ace areas instead of observing them with extra views. Therefore, we design a deep neural network, named OSVP, to directly p
osed initialization method is combined with a "modality-balanced" transformer decoder where the queries can access all sensor
geometries and textures from the captured 2D images, SM3 achieves integrated optimization of movable parts and joint parame
ng semantic features from range images, which are used as movable object guidance for the motion branch. Our straightforwa
l reconstruction performance without requiring extra priors. Additionally, we introduce an essential mask scheme to adaptively i
certainty in individual estimation, and together with the volumetric aggregation, we can reliably extract useful geometric informa
be applied to daily cloths that have similar topologies as our template mesh, but with different shapes, sizes, patterns, and phys
manner. Furthermore, we propose the novel AmodalFlowNet as an initial step toward addressing this task. AmodalFlowNet co

an upsampling strategy based on both the semantic and geometric feature similarities between sparse and dense samples. Exp
from circum-views to fully exploit the spatiotemporal correlations between frames and capture more comprehensive clues. Fur
orming event-based optical flow network directly estimating flow from event traces without using correlation volumes. We furthe
es themselves, enhancing both communication and computation efficiency. Specifically, we present ActFormer, a Transformer
have shown impressive results among generative model frameworks in recent years. To effectively train DDPMs in the LiDAR
ention generator. This generator is designed to automatically discern interrelations across the three tasks and arrange the task-
e maps. Additionally, we propose a transformer decoder for the segmentation task to dynamically adjust the learned features by
pes are developed using simple materials and manufacturing methods, on which various validation experiments are conducted.
eatures in the stick region. The proposed method was applied to the Tac3D vision-based tactile sensors using continuous marke
e the local normals in a geometrical approach to estimate curvatures between adjacent contact points. These parameters then
(rest position, stiffness, contact location), allowing the coupled robot/environment dynamics to be linearized or efficiently used i
ts are outlined by white markers, while the contact depth characteristics can be indirectly obtained from the distribution of black
een the objects. By jointly learning geometry and physics, our model reconstructs consistent and detailed deformations. We've
problem using nonlinear model predictive control (NMPC). According to the experiments, our NMPC-based planner outperform
ajectories. The theoretical results are demonstrated in 3-D simulations and 2-D experiments using feedback from a high-speed
bution rather than a single constant, and show how this distribution can be narrowed when given a prior on Fn or v. For a given
ng and Abstraction with Guidance). LEAGUE leverages the symbolic interface of a task planner to guide RL- based skill learning
k helps the TTA model to harvest the knowledge of multiple representative domains (i.e., compound domain) and conduct the T
hich adapts to each new example while also generalizing well on past data to prevent catastrophic forgetting. Our approach also
the experience. Specifically, the experience of old tasks are distilled to obtain Markov Decision Process (MDP) data with high c
apt faster and generalize better compared to state-of-the-art RL and hybrid approaches, significantly improving robustness whe
rve as context, 2) a self-guided exploration policy that proposes new tasks in simulation, 3) a skill abstractor that distills recent e
s.google.com/view/halp-submission}.
he required model parameters for VCL. We establish a robotic tactile CL benchmark using publicly available datasets to evaluat
roller that flexibly composes various skills to tackle vision-based manipulation tasks in the lifelong learning process. Our compr
reduced by 58% through mid-air separation. In detail, to compare the attitude response of the two configurations, a dynamic m
ns only. Intriguingly, we find that the discovered policy can not only generalize to unseen object shapes in the simulator but also
pose combining a genetic algorithm with a lexicographic evaluation of solution candidates to overcome this problem and naviga
a prismatic joint and two revolute joints at each end. Each revolute joint has a suction unit that enables locomotion and adhesio
objects with varying numbers of assistant agents, demonstrating its versatility across different object types and multi-agent syste
s with the aim of enabling communication between robots under bandwidth constraints. Our compression algorithms are naviga
mation with UGVs via limited relative pose estimation. The UGVs plan their trajectories in the received map and predict possible
d shape control is accomplished for pattern formation in light of a mean shift technique. The proposed algorithm is evaluated vi
such force feedback data, this paper focuses on the need to provide a specialized dataset, integrating optical coherence tomo
proposed method introduces a calibration model and a data association algorithm to fuse data from panoramic images and 2D
handover failure detection: (i) a video classification method using 3D CNNs and (ii) a temporal action segmentation approach w
oduce new metrics to do so. Finally, our dataset contains a large variety of industrial objects, from small and simple objects suc
enchmark progress, we propose a LiDAR-based model with a Spatial-Aware Transformer for global and local feature extraction
n industrial robotic grasping. By reducing the sim2real gap and enhancing deep learning training, we facilitate the application o
f our knowledge, the first dataset* annotating for OOIs in dynamic scenes. To model OOIs, we establish a classification taxonom
using a system of three Robotic Total Stations (RTSs) tracking moving robotic platforms. Furthermore, we compare the perform
stream RGB-D perception tasks. Data and code are available at https://github.com/shanice-l/RaSim.
g boxes of static and dynamic objects, camera poses, and IMU. The dataset is still continuously expanding. Then, the performa
human-human to human-robot data. We propose a novel architecture, InteRACT, that pre-trains a conditional intent prediction
We exhaustively mine and taxonomize common & rare driving scenarios which are used during evaluation to get fine-grained in
ur system contains three components: a combination of top-down and ego-centric views, natural human behavior in the presenc
vantage of the cheaper human collection along with robot collection. We release a large and highly diverse (29,520 unique inst
ure and properties. Our model then learns to perturb those graphs to generate corner cases using attention and triple embeddin
his paper aims to unlock the potential for an agent to generalize to hundreds of real-world skills with multi-modal perception. To
the human have acted in the same way? By minimizing this counterfactual perturbation, we can induce robots to behave in way
s created using a scenario-based approach, which is widely accepted in the automotive industry. We design a scenario taxonom
lable and seeks to make a contribution towards increasing the safety on inland waters.
annotations for multiple important road objects covering 10 categories and 19 explanation label categories. The dataset also in
s and sensors, making it difficult to analyze and evaluate the method accurately. To tackle this problem, this paper presents LiD
oc and further rendered with the underwater neural rendering (UWNR) strategy to enhance their realism. The testing subsets co
nt cloud data. We open source the dataset, along with an ontology for unstructured terrain, as well as dataset standards and gu
abels to the image space. We trained a suction prediction FCN (Fully Convolution Network) with our dataset generated from the
image. Our method specifically takes pairs of videos as input and yields matched pairs of images along with their estimated re
the first release of this dataset, we provide ground-truth data with over 1,800 manually annotated semantic labels for tree stems
eature similarity estimation network exploits the deep features to compare the symmetry and geometric similarity between real
consistency, task confidence, and geometric saliency to provide more informative guidance. Specifically, the transformation con
hat our two-stage optimization significantly improves the registration accuracy of the feature association and selection module. O
ents a double voting strategy to provide the framework with coarse-to-fine parameter estimation. Experimental results on sever
that our method performs better in adverse weather detection and has higher robustness to unseen weather effects than previo
ce while utilizing only 1% of the parameters and achieving up to more than 10 times faster inference speed. By applying descrip
l Tree-based Attention (TrA) is proposed to guide the progressive convergence of the attended regions toward meaningful poin
the baseline algorithm by introducing a large pre-trained model and human prior knowledge. Remarkably, our novel algorithm c
ons stemming from the heavy computational costs associated with entire model updates. In this paper, we propose a distributio
D shape data. A binocular visuo-tactile system (BVTS) is designed and fabricated for dataset construction. The entire 3D shape
asily and improve accuracy. In particular, it does not require a separate perception system (e.g., object classification, object pos
To this end, alongside a curricular data sampling strategy, we propose a new encoder-based neural density field architecture fo
nsists of an open-vocabulary affordance detection branch and a language-guided diffusion model that generates 6-DoF poses
m RANSAC. Multi-Object RANSAC demonstrates superior plane instance segmentation performances over other recent RANS
the keypoint locations as the center of those markers. We, then, use an inpainting method to reconstruct the parts of the robot
ally, we demonstrate a considerable performance boost is achieved when a state-of- the-art detection model is optimized prima
he need for finetuning in the deployment environment. We demonstrate the performance of our approach on the task of agile, v
r movement by predicting the corresponding schlieren-induced optical flow from a single greyscale image captured by a moving
-resolution camera and an ultra-low-power System-on-Chip (SoC) hosted onboard. We present a vertically integrated system b
. Specifically, for point-level pseudo labels, PCTeacher employs a multi-view fusion strategy to achieve higher precision and re
e use of noisy robot poses rather than precise poses that have to be pre-calculated. Compared to a baseline approach the pea
depth estimation improves depth discrepancies at object boundaries, leveraging a co-optimization network and proxy semantic
his separation promises enhanced performance without sacrificing the efficiency innate to keypoint-based techniques. A vital co
e ego vehicle, capturing complex lane patterns in both urban and highway environments. We leverage the geometric traits of la
the deployment for more tasks in real-vehicle applications. Our approach is completely self-supervised, which means it does no
N2 M2 ), which extends this decomposition to complex obstacle environments, extends the agent's control to the torso joint and
n, potentially capturing the complexities and uncertainties of real-world multi-agent interactions. We also design a navigation sy
s in images, enabling keypoint tracking on symmetrical objects and occluded and out-of-view objects. We employ ourmethod to
iction by proposing the Path-based prediction (PBP) approach. PBP predicts a discrete probability distribution over reference p
memory scheme to achieve dynamics identification, accurately approximating the static hysteresis. We also introduce modificati
es when merging policies, such as utilizing common pre-trained initializations, increasing model capacity, and utilizing Fisher inf
sformer model with prior knowledge and explainability. Our framework features the Transformer model that works with a goal e
action value. In addition, we design a new conflict resolution without additional guidance to emphasize the cooperative nature of
Statler framework significantly outperforms strong competing methods (e.g., Code-as-Policies) on several robot planning tasks.
ce, ranking 1st on its leaderboard (https://waymo.com/open/challenges/2023/motion-prediction/) .
aluating our method on a previously published Active Clothing Perception Dataset and on the real system, we establish that the
impact of using chemical synapses compared to electrical synapses, in both sparse and all-to-all connected networks. We con
ndant modules from those efficient models through a comparison of their contributions. Next, we leverage the efficient model as
ally reconstruct 3D models in a hybrid representation, where the density of the radiation field is formulated by a voxel grid, and t
actile images obtained from a set of ten 3D printed YCB objects. The results reveal a total accuracy of 81.9%, a significant impr
on the surrounding agents, which facilitates the training while allowing for single-motion forecasting when only one human is in
mework (SPCGC) for machine perception. This framework tackles the fidelity issues associated with point cloud compression an
generative AI which are made exact using the guarantees of traditional planning and optimization. We evaluate our system aga
amics models with few-shot real-world data. A safety filter based on the Control Barrier Function (CBF) on top of the RL policy i
for a computation of the learned Cosserat rod model multiple orders of magnitude faster than a publicly available reference mo
e concept of constraint manifold, including dynamics, and neural planning methods. Our approach generates plans satisfying a
k reachability for execution to achieve the target arrangement. Our experiments on a diverse range of simulated and real-world
are made available online at github.com/omron-sinicx/action-constrained-RL-benchmark for further research and development
ropose various uncertainty estimation schemes based on emph{Bayesian uncertainty quantification} and emph{distributional en
ata by proposing a contextual curriculum-based method. We ablate the proposed approach on real-world evaluation data and d
diverse receptive fields to enhance keypoint refinement. By incorporating the perspective-n-point (PnP) algorithm and leveragin
but only a small set of negative examples, thus increasing the applicability of autonomous robot systems. We detail the design
esulting in a single system that predicts multiple grasp candidates for a specific object and task from multi-object scenes. Embe
contacts. To train PAST, we introduce D4PAS: a large-scale Dataset for Part Assembly Sequences consisting of physically va
our designs in the real world with a Franka Panda robot. Please see the supplementary video and https://object-adaptation.gith
nalysis shows that existing pretrained models fail to capture the essential visual features necessary for this task. However, a vis
Prediction Network (APN) and a Configuration Prediction Network (CPN) to model the complex pattern and stochasticity of soft-
ns, it is trained using data augmentation and extreme viewpoint changes. As a result, the student policy learns robust visual fea
rt several interesting empirical findings related to the utility of learning force feedback control policies in both general manipulati
s perform well in terms of different numbers of objects and different numbers of distractor actions. Furthermore, we show our a
d infers the corresponding grasp. Then, we propose two effective separation strategies: Dropping the entangled objects into a b
demonstrations, we show how a pick-and-place policy can be learned and deployed on a real robot, without any robot data coll
on deep reinforcement learning. This control system can perform grasp tasks while dynamically adjusting to avoid end-effector
grasping synergy, an active double-step learning mechanism is introduced to optimize the training reward function for the pushin
e our method’s effectiveness on various robotic tasks, both in simulation and on a physical robot. Notably, in the surgical re
ralization to different (sets of) factors. We design a simulated benchmark of 19 tasks with 11 factors of variation to facilitate mor
ronments, novel arrangements and novel objects, including challenging transparent objects, both in simulation and the real wor
hat can solve complex object-arrangement tasks such as shape-matching,stacking, and even full path-following tasks such as a
vironment. Designing such a complex behavior analytically would be infeasible under uncontrolled conditions, as an analytic ap
ng PPO (Proximal Policy Optimization) by running 768 parallel actors in simulation, for a total of 1,2M environment interactions
lizable Articulation Modeling and Manipulating for Articulated Objects (GAMMA), which learns both articulation modeling and gr
asp distribution model based on Power-Spherical distributions that enables a training based on all possible ground-truth sampl
rom the replay buffer based on rich information due to contact, leveraging the touch sensors in the gripper of the robot and obje
meras, our system distinguishes itself by its ability to directly utilize raw IR and RGB images for transparent object geometry rec
stages can be performed individually, which enables a straightforward application of general offline learning algorithms. We con
n image-only robotic manipulation framework that leverages an eye-on-hand monocular camera installed on the robot's parallel
condensed and hierarchical 3D representation. In addition, a single versatile RL policy can be trained on multiple articulated ob
k completion. We demonstrate our approach on simulated 2D reaching and 3D multi-fingered manipulation tasks.
oposed real robot manipulation benchmark. Our approach learns end-to-end, directly from pixels, and does not rely on additiona
ral action primitives, such as grasping, stacking objects, or opening drawers, strongly improving data efficiency in simulation, an
adapt to different gripper designs. Our approach involves a three-step process that first obtains a rough grasp pose prediction
gate local points into graspable regions and provide global semantic information. Further, a novel non-uniform anchor sampling
antage of the fully differentiable nature of HPN-NC, we provide a three-stage training procedure to servo real world objects. Wit
out any fine-tuning. The efficacy of our approach is demonstrated in simulation and experimentally validated in the real world o
or extracting an enhanced spatio-temporal context to recover high quality fidel predictions of future point clouds. We conduct ex
btly incorporates the multi-level consistency regularization on the predictions and the heterogeneous feature spaces via pseudo
hat utilizes reflectance information based on Retinex Theory to help the model learn a unified illumination-invariant representat
enecks in quantizing the activations of Q, K, V and self-attention in the attention module, as they do not adhere to simple proba
approach efficiently handles radar ghost points and enhances the sparse mmWave radar point clouds to dense LiDAR-like poin
recognition techniques based on transfer learning and metric learning. We show real-world experiments that demonstrate the p
s. The key idea is to autonomously detect valid pixel correspondences by attempting to use a prediction over a new image to p
ly adopted NeRF-based SLAM methods in terms of 3D reconstruction accuracy. Moreover, our approach shows 40.8%(except
s from real images and composes them into new scenes, generating photorealistic renderings and many types of 2D and 3D su
ransport based matching on one human demonstration. Finally, we use online reinforcement learning on our robot to optimize t
e linear objects. We seamlessly integrate this method into the robot control loop and achieve an adaptive shape control framew
uilding blocks, and common household objects. Further, real-world tests are conducted using actual robots to confirm the robus
vered by a custom tactile sensor array. The real experiment demonstrates the superior robustness of the proposed approach a
g space for collision-free grasping actions. Simultaneously, we integrate throwing behavior, showcasing how this action significa
everages the ability of underactuated hands to maintain a passively stable grasp during self-identification movements. This me
tle caps. Furthermore, we introduce a novel visual-tactile fusion network and learning strategy, with one key module to tokenize
y of the object during a contact formation. The contact patch could be used to estimate the stability of the object upon release o
screwdriver and arm/hand. To maintain the fastening effort, estimations are carried out sequentially for the screwdriver's pose
oposed method.
onstrate a significant improvement in performance, characterized by reduced execution time, heightened robustness, and super
g, etc. Using the Mujoco physics simulator, we demonstrate that our framework can efficiently and effectively learn five differen
w/everydayfinger.
lem. By extending its description to include a virtual spring, we have also achieved subtle force application by two cooperated m
apping the robot manipulation onto changes of the modal features. In the Jacobian matrix, unknown geometric and physical mo
ples. To further alleviate (ii), our second contribution is a convex, differentiable and quasi-dynamic formulation of contact dynam
plex space created by contacts. It solves a diverse set of manipulation tasks that require dexterity, both intrinsic (using the finge
hm for in-hand orientation tasks. Our system demonstrates 88% of successful grasp and successfully classified small objects in
gorithm introduces the leader agent assignment mechanism and a collaborative experience reward. The former deals with the a
ion's latency requirement, thereby empowering robots with restricted on-board computing capacity to safely and efficiently navi
to nearby agents. We present and evaluate an algorithm on a large set of transitions between formations with 5 agents and an
e all computations are parallelizable among the agents and only local communication is necessary. Simulation results that scale
distributed and complete, meaning that only local communication is required and a K-serial Stable solution is ensured. Furtherm
y non-convex and unstructured, this method efficiently generates collision-free Voronoi regions using only local sensing informa
we also develop the Incremental DFGO algorithm (iDFGO) for real-time problems by combining the ideas from LC-ADMM and
allocations. For the pre-allocation scenario, CBBA outperforms the competitors in terms of reliability, communication overhead
 roach, thus constructing a 13-dimensional system state-space equation, which is discretized by using the explicit fourth-order R
d integrator is adopted by exploiting the differential variables; then, the constraints of the morphing quadrotor are incorporated i
(NN-guided) sampling strategy tailored for LTL planning. Specifically, we employ a multi-modal neural network capable of extra
s value within a bounded time shift. We use reinforcement learning to solve both control synthesis problems for unknown syste
oposed for comparing this method with the conventional dynamic identification and previous DeLaN implementations. Moreove
ve inverse dynamics algorithm and its linearized tangent version. The approach is illustrated with numerical examples from the
atrix (OSIM), permitting efficient OSIM computation, which we further accelerate using matrix inversion lemma. We generalize t
k, we introduce the formulation of RIMs for flexible multibody systems and based on that propose a general co-simulation sche
ment learning paradigm, RIDER learns to infer its own dynamics by predicting how future robot observations and states evolve
ponsible for planning both the robot motion and desired contact forces, while the Gain Tuner dynamically adjusts the compliance
uded in the no-escapable area. The no-escapable area can be obtained based on the potential analysis considering the contac
duce a novel binocular edge-map-based nonlinear optimization technique. Our primary contributions entail an improved initial p
data with simulation labels. Finally, we show the superior performance of ASAP at generating physically realistic assembly sequ
lysing the system throughput, the optimal fleet size for the system is proposed. The approach is validated through various exam
xhibited ubiquity and high distinctiveness at different location, which makes it suitable for estimating the initial transformation ma
k box component. We apply a policy gradient framework for optimizing action probabilities under such a reward model showing
hese two complementary approaches. Our technique splits the vertices of the graph into a partition that includes a shared portio
bal estimation of the material flow in the system. We utilize a model predictive control algorithm that optimizes the rate of materi
ate through the abandonment of the yaw channel. Meanwhile, the FTC scheme is examined under constant wind disturbances
compare a desired idealized rigid response, which is straightforward to obtain with existing tools, with a measured impact resp
olves movement, actions of the end effector and interactions with the objects of the environment. Since several of the contained
r from failed tasks in the short term, and (iii) learn to avoid failures in the long term. For this purpose, robot executions should b
achieve a unified policy that is capable of switching from four-legged to three-legged walking mode when the complete motor fa
ed during training.
oth simulated and real quadrotor flight data. It is able to detect LOC induced by actuator faults without explicit knowledge of the
hs and composing them through temporal logic. We implemented the framework to monitor for specification violations of 3 top A

ale subterranean environments. Our topological map contains the contributions of keyframes that generate each SER, enabling
d unperching procedures by estimating the position and pose of the UAV relative to the powerline. These measurements are se
models via multiple disparate pipelines that reuse the same underlying sensor data, which leads to increased computation, red
horizon planning fashion. Mode switching is facilitated by associating optimized motion durations with a predefined traversal se
multi-category LiDAR localizability analysis. In the second part, this localizability analysis is then integrated into the scan-to-map
nts the field demonstration of teleoperated pick-and-place tasks using the ULR SILVER2 for which a novel stance control, Grap
wledge represented in the field belief's concentration and flow field, i.e. predicting sources upstream. As a result, our method y
solute Error (MAE) of the turning radius produced by the trained prediction model is 0.58m. Furthermore, it has been observed
posing the inspection problem into different levels. The high-level decision maker first determines the task for the robot and gen
sed-loop control for the intelligent actuation of a set of soft telescopic structures that raise and tilt the solar panels in the directio
aks, by means of image color filtering; and geometric shape anomalies, by measuring joint angles of the robotic arms. We prop
e of mechatronic devices and apply it to lightweight industrial robots. The proposed method allows the identification of the elec
m emphasizes two novel strategies. Firstly, external visual tracking of the UAV informs the landing controls for both the drone a
he low resolution prior acts as a coverage survey allowing the algorithm to perform a relatively exploitative search of high conce
oyment. It is vital to not only be intentional about where to sample in the environment on a given day, but also determine the wo
high-precise pose determination during calibration. The root-mean-square error of the distances to a georeferenced ground tru
ments of particle-laden flows while sampling particles to determine their properties as they are transported away from their sou
sability prediction, eliminating the need for heuristics. Our experiments demonstrate that our method excels at avoiding obstacle
Tip sensor with opaque skin. Furthermore, we develop a Generative Adversarial Network (GAN)-based approach for modality sw
actile sensor is designed for distributed force detection with electrical resistance tomography technology. The sensing performa
eal-time. Within this paper we present a system for the real-time classification of texture using a neuromorphic tactile sensor, a
o estimate tilt torques from gel-based visuotactile sensors. This method does not rely on deep learning, sensor-specific mechan
demonstrate that it is possible to train on a single object and a limited sample of goals to produce precise and reliable policies
een in training. In experiments with linear motors and Phantom Omni devices, using sensorless force estimation, we show that
is tracked by detecting the displacement of an ARTag and is transformed to three-axis contact force via calibration. We further
ator and develop a planar reflection simulator to simplify the tactile sensing hardware. As a result, GelLink only requires one mo
solution depth reconstructions of an object deforming the sensor’s soft elastomer surface. Additionally, we show the advan
dding space capable of predicting a variety of task-centric qualities over multiple manipulation tasks. We collect tactile data from
stimate a full contact state, i.e., contact position, forces and torsion. With that, an experimental benchmark between various con
with enhanced accuracy and efficiency. Remarkably, we leverage the optical and deformable properties of the translucent gel s
To stitch together the tactile images obtained at different camera poses, we use an As-Projective-As-Possible image stitching a
end, two cameras were integrated in a single sensor, whose interface is made of a transparent elastomer coated with a thin laye

on a computational model of human tactile perception that relies on CASR and tactile flow concepts. To this aim, we used a so
ulation, we contribute several efforts, such as introducing appropriate reward functions and proposing a novel goal-update mec
sentation for higher-level downstream tasks. Our method achieves an average of 95.97% IoU coverage on unseen YCB objects
terrain attribute extraction algorithm is proposed based on contrastive learning. It leverages the haptic data generated from the
ated into a geometric-kinematic optimization scheme to enhance existing visual pose trackers. To evaluate our method and to fa
al sensing can bring cloth unfolding to a new level of performance.
nd shear forces significantly more accurately than the state-of-the-art, with comparable accuracy for detecting the contact patch
ensing module was 3D printed using soft material, and the FBG fiber was attached to the module using epoxy. From evaluation
to verify the performance of our proposed sensor. We also discuss the possible application scenarios and demonstrate how SA
egistered. Furthermore, our study presents a streamlined skin design that simplifies manufacturing processes without sacrificin
he assistive task of limb manipulation. We open source all sensor-related resources on our project website.
nd contact patterns including small contact traces. The proposed SightGAN learns real-to-sim and sim-to-real processes over d
study, we design a novel cylindrical vision-based tactile sensor that enables continuous and swift perception of large-scale obj
nsors. We trained the proposed GAN model with an aligned real tactile and simulated depth image dataset obtained from deform
sired manipulation objectives and outperforming single-shot methods in estimation accuracy. The proposed approach holds pot
nstruction errors across all eight airway models compared to the traditional model, validating the effectiveness and feasibility.
so, we achieve a 10-fold increase in success rate above the baseline control strategy and significantly improve completion time
tial array noise obviously. This work could benefit various scenarios, such as human-robot interaction and physiological informa
iational Autoencoder with contrastive optimization, which jointly learns underlying visual-tactile latent representations, a latent g
y availing both sensory inputs, MViTac leverages intra and inter-modality losses for learning representations, resulting in enhan
reactively handle accidental tactile stimuli.
ot obtainable from static images. We then define the interactive feature plane (IFP), whose spatial property corresponds to the r
 hts applicable to a wide range of existing navigation systems. Our analysis includes 9 degenerate motion identification includin
e tracking. While a tactile-based classification and manipulation process is designed to estimate and align to the contact angle
y, we clarified a way to attach the strain gauge in a position that improves sensitivity. The results of the actual prototype sensor
d on the contact sensing method is proposed to maintain stable contact force and regulate the probe's orientation for robotic ult
and the interactive constraints. Finally, extensive experimental trials conducted on the point-foot biped robot BRAVER validate t
achieve online identification of the noise statistics in the force dynamics. The GPADKF is capable of decoupling the contact forc
al space frame. Second, using the object-centric tactile skill definition, we synthesize unified force-impedance control and forma

d technique to select and improve on partial strategies. We provide conditions under which the algorithm is probabilistically com

ions. In particular, we use Gaussian process regression to learn the robustness based on precomputed predictions so that robu
robot learning has been largely characterized by a mathematically-flawed simplification, hereinafter referred to as the "single ta
f both approaches, while avoiding their shortcomings. Specifically, the relaxed specification automaton explicitly accounts for th
ation and success probabilities and then solves a multi-agent route planning problem. We use an efficient Mixed-Integer Linear
ks with stochastic latency. To do so, we use tools from formal verification to construct a shield, i.e., a run-time monitor, that prov
m over a Varying Weighted Transition System, where navigation durations are deterministic but vary depending on the time of d
again for improved task performance. Particularly, the LTL instruction is specified by co-safe LTL. As a semantics-preserving r
ust be made in each model. In this work, we propose stochastic games as a general model for human-robot interaction, which
y the causal knowledge graph, a proposition graph is developed, based on which the decision module encode the environment
ch significantly reduces the search space. Based on TB-PDT, the search algorithm, namely Intensive Inter- task Relationship Tr
l LTL task specifications. Experiments in Minecraft-inspired domains show that after training on only 50 tasks, LTL-Transfer can
ity of our model, a painting test on an L-shaped surface was conducted, and the measured and predicted values were compare
e probabilistically verifiable. We demonstrate this verifiable behavior generation capacity in both simulation on an exploration ta
posing synergies rife for optimization. We show how to use this representation to leverage both informational overlap between m

by human grasping, a tactile calibration method and a visual-tactile fusion classification method are designed, which improve th
ding the same effect. This paper proposes a new type of soft robotic fingers using a novel method to produce a variable stiffnes
on model. At inference time, this model can be utilized to transfer grasps to novel objects by identifying the most similar-looking
which can be used in an uncertainty-aware planner. An explicit representation of uncertainty in the encoder enables information
a parallel manner and optimizing predicted grasps based on hand-object contacts. Simulation experiments show that our model
correspondences, which are learned in an unsupervised fashion without any expert knowledge. We showcase the effectiveness
zation to obtain attacks that simultaneously minimize the noticeability of the attacks and the chance that selected grasps are su
with multiple reference databases, including images and text references, each with a different coverage and matching challenge
ation Relationship Detection (GSAGED). This algorithm assists robots in detecting targets in complex scenes and determining
espectively and picked the best-performing models for force and location predictions. The best skin configurations yield more th
ngers and a dexterous hand, and tested in the real world. Conducted analysis on the collected measure shows correlations betw
ntation for kinematic grasping motions in synergy space, that allows the generation of novel, human-like grasps under direct con
in it based on past grasp scores. We also leverage a motion vector field estimator to shift the center of our sampling region. W
nd. We articulate a design methodology for low-cost, easy-to-fabricate, large workspace, and compliant Fractal Fingers. The co
h does not capture the composability of the environment. In this paper, we introduce an end-to-end architecture for object-cent
minimal effort. In addition to the physical design and corresponding software, we include a dataset of 1,020 grasps. The grasps
with the ability to monitor and detect errors during runtime task execution. We introduce MoMo, a model-based runtime monitor
ential features, but also the capacity of capturing dynamic interaction details and local geometric features during the grasping p
or and safety faults that are catastrophic. In this study, we isolate the root causes behind these limitations to develop a system,
oyed to assess reachability, providing a feasible and adaptable grasping strategies. Feasible grasp poses are targeted by nove
d 87% with 4500 grasp attempts. Additionally, we achieve a success rate of 84.51% in a dynamic human-robot handover scena
nger-object interactions. In addition, we develop a new dexterous grasping optimization algorithm that solely relies on the point
aling. We categorize PCBs into two recycling levels – with large surface features intact or removed – and conduct experime
partial observations, and manipulating them using a 7-DoF anthropomorphic robot hand. Our approach has significantly improve
or the variability between experiments. The advantages of our approach are demonstrated in the practical comparison of four g

chanisms switch their motions automatically when the motion being generated is prevented. An analysis of the mechanism and
ain finger force. By using a minimum number of optical encoders placed in the drivetrain, accurate position and force sensing c
ectiveness of the proposed design in grasping circular, rectangular, trapezoidal, and concave objects. The results were compar
perimentally verified. A time reduction of 16% was obtained when making contact at 0.65 m/s.
tuator that can accelerate a magnetic ram into various surfaces to generate impacts. The surface was used to transport a part a
ed the overall stiffness and performance of the manipulator, enabling load capacities of up to 3N per finger and more than twice
he external magnetic field is established. The average error between the experimentally observed open angles and the model-p
cal and electronic design and provide qualitative and quantitative experimental validation of the achieved repeatability and the i
assembly state. Viable destructive actions are identified and verified only based on the underlying geometric model, circumven
to a very wide range of planar objects with a single design. A companion paper introduces a design methodology for this new c
n-board camera are used to detect the food item features and plan the grasping strategy that better exploits the possible environ
es with control as well as hardware design changes. This Jacobian exploits the structure induced by environmental constraints f
design of future agile robot hands and feet that include high impact landings on rods with pose inaccuracy.
and orients food through embedded 6-axis force-torque sensors. We derive and validate the kinematic model for HASHI, as w
of the palm as well as the fingertips. We also demonstrate the generalizability of tactile models trained with the fully integrated s
hieved. Here, we developed a cross-structure hand to grasp various objects. We experimentally demonstrated that the integratio
cm adhesive swatch – 2.5x heavier than with suction alone. The adhesive is fabricated on a flexible nylon fabric. A small fan
uations of our approach on four real-life tasks: pick-and-place, water pouring, shooting a hockey puck into a net, and sweeping.
es on dataset with rich spatio-temporal features, a self-designed dataset, ALU-Textures, was constructed with 10 fabrics from e
nt size to human. Here we proposed a new method to reconstruct the human finger joints by using a highly under-actuated han
orkload, enabling a systematic and robust device assessment. The results show that the PULSE device enhanced performance
tion on the board. Specifically, we investigate the vibration characteristics of a single passive coil actuator and construct its vibr
s, and providing a quantitative performance score after the execution of each training exercise. An experimental study shows th
ncrease in applied vacuum load. A psychophysical study is also conducted to obtain the just noticeable difference in stiffness th

pproach that uses a data-driven model's entropy as an uncertainty measure and explore relative to that entropy conditioned on
conditions. Our display is consistent in shape output, with an average standard deviation in height of 0.75 mm or 0.47% of the d
mental mesh rendering algorithm to achieve precise memory reallocation by updating the buffer in a fill-in-the-blank pattern for g
ps through a Series Elastic Actuator (SEA) to approximate the wearer's intended future movements and computes a correspon
ach is evaluated using real-world data, demonstrating its effectiveness in extracting meaningful information and accurately reco
iments involving 10 healthy subjects were carried out, consisting of an offline session for systematic evaluations and compariso
seamless and natural interaction. Human-robot coefficiency is first modelled by identifying implicit indicators of human comfort a
propose utilizing a Human-In-The-Loop(HIL) method to enhance dataset quality. This approach relies on backbone deep learnin
Bellman-like recursion. On discrete navigation problems with both fully- and partially-observable state information, we show that
orkforce capable of consistent, safe operations. We successfully utilized the method developed in cite{kunde2022recognizing} t
s-based digital human simulator to predict the fatigue cost of the possible robot actions. We then use an online planning algorit
y humans, while avoiding excessive uncertainty according to the human preference predictions as determined by reward estim
ased on explicit factors related to state and task understanding that may be difficult to generalize. Addressing these challenges
ta from safety critical scenarios called CARLA-VR. The CARLA-VR dataset addresses the lack of long-tail data samples in com
icromobility vehicles. Specifically, we model human preferences in vehicle selection based on vehicle usage history and curren
sulate pilots' inputs, and the time mapping based trajectory progress is customized to further capture intention. Then, a trajector
n the tendon, and continuously measure it through an encoder and a load cell connected with the cable. Two sets of experimen
mation of the gel to provide the desired contact force and torque depending on contact, but also performs avoidance by followin
roach employs an augmented energy tank with restricted energy flow. The second approach monitors the overall energy flow, r
al world with the tablet and add them to a map, changing the navigational behavior of the robot. The framework is evaluated by
er models trajectory prediction as a Markovian process, where future pedestrian coordinates depend solely on the current state
miliar situations. Specifically, we improve the Socially Attentive Reinforcement Learning (SARL) policy by (1) modifying the train
s that participants exhibit a variety of path preferences: from risky and urgent to safe and relaxed. To model users' decision ma
present a novel machine learning approach for RTP that is additionally scalable to diverse home signals. Trained using curriculu
ated that the manipulation of the LCF induces interpretable changes in pedestrian behaviors, along with a superior performance
he ability of gradient-based planning algorithms, such as model predictive control (MPC), to generate fine-grained high-quality m
generalized path consisting of nodes and edges derives various normal paths efficiently. 2) By multiplying two vectors starting
ed to explain sampling-based motion planners, or even optimization-based motion planners in situations where computation spe
ng abilities and learning to grasp unknown objects within three demonstrations.
al representation to capture the spatial information of EEG, taking into account the topological relationship among electrodes. T
r technique by benchmarking against expertly engineered classifiers for visual-audio-haptic surface recognition due to the indus
ce profiles. Analysis of anonymized unlabelled preferences provides further insight into group preferences. Identification of the m
ches, while also achieving promising performance in simple instructions following. Finally, we share results from a pilot study wi
ble parts, and then an action-aware costmap is constructed for generating a feasible path. The pre-trained large models have g
adhesion, sweating, hair blockage, and skin stretch challenge the reliability and scalability of these modules as the perception
, RNNIK-MKF and CMk=5) with users feeling comfortable reaching upto 60% of the total distance to the target for handover wit
apse when they discover deceptive behavior by the robot. To mitigate this problem, we adopt the robot theory of mind model to
via our method is perceived to be more human-like and have better game performance and experience than self-learning from
robotic system to locate a user and a track, calculate desired moving directions, and provide haptic feedback to the user. We c
while maintaining drivability.
orizon planning through a natural language interface, and a compliant controller for adaptive motion. Our robot-assisted peeling
eraction task and the simulation-to-real generalization performance was evaluated for the proposed network along with alterna
can be used to understand textual path. Therefore, we propose and evaluate two methods, one for storing implicit maps in LLM
ence waypoints, which are probabilistically fused with geographical database information and condensed into a geospatial distri
es of the defect in real-time. Additionally, inspectors can monitor the inspection progress online, which enhances the speed of t
m to pick and place a sample holder at a beamline, recording a 90% repeatability rate. This provides the groundwork for further
ion in the process of carbon fiber draping the production of outer parts. After utilizing the evaluation framework at the two test s
uscle activity and the estimated EoP. Ten subjects participated in this study. The EoP of the wrist was identified through high-fre
s to observe specific parts of the environment via object manipulation. Our method spatially associates the environment with la
d during overground and treadmill walking with nine able-bodied participants. The linearly combined controller with a greater em
he thinking of our community as it grapples with how we want to live with (and “in”) robots. Keywords: Robot surfaces, U
ble, which opens up to using trust as a parameter for informing the interaction with robots.
bling the robot to maintain a respectful and comfortable distance from the target person. Concretely, the social interaction dyna

of partially observable systems. The incorporation of belief predicates in scLDTL enhances its expressiveness while simultane
ditional behavior prediction that allows the robot to infer the human's intentions based on its future plan in data-sparse environm
ounded AI team assistant, that intervenes only when the team's expected likelihood of failure exceeds a predefined risk thresho

as shareable data. Our system aligns and merges the AR-HMD and the robot’s local map automatically, triggered by a hum
eos. They are trained in a semi-supervised manner, with the output classification networks predicting one of possible labels for
w that CoBT achieves approx 93% success rate overall with an average of 7.5s programming time. We conduct a pilot study wi
nt and a set of virtual tubes into which the path of the robot can move is built. Lastly, an optimization problem is solved online t
methods and the higher dexterity of humans during physical interaction. The human motion and strategies are transferred from
y mentioned audio communication systems. Participants were then asked to identify the communication from their system in a p
ynamically allocating authority. Variable impedance control (VIC) and force control are employed, where impedance and wrenc
receiver responses. In our framework, diverse care-receiver responses are autonomously learned through trials and errors. In
ot's strengths by dynamically adjusting control priority based on the deviation distance. We show that using this method in a sup
feedback, distill generalizable knowledge from corrections, and retrieve relevant past experiences based on textual and visual s
th orientation and position in an intuitive way. High operation precision and efficiency can be achieved without interfering with th
city controller, which meets the needs of large-scale HRC. To evaluate the proposed framework, a multi-subject experiment inv
on the prediction of operators’movement from an adaptive filter and fuzzy logic inference. The proposed method is execute
oth switching between different dynamic tasks without gain tuning. We also introduce a telemanipulation architecture that offers
ng problem, targeting human-robot collaborative plans. The LLM generates information that is encoded in the planning problem
e visual explanations through strategy-conditioned landmark states and generate textual explanations by giving the landmarks
to the constraints imposed by ISO/TS 15066. To achieve high-performance outcomes, the control problem is addressed using
ultiple robot links based on ISO/TS 15066, allowing gradual establishing of physical contact. Then, the time-varying zeroing con
ety limits according to the standard. The proposed planner was validated first in simulation, comparing completion times with ot
sNav, can actively discuss with these experts to collect essential information before moving at every step. These discussions c
d with personality-dependent hedonic experiences linked to memories. Within this paper, we propose a task- and platform-inde
ward) within a preferred trajectory. To include the additional information, we design novel regularization methods within a prefe
he constraints and abilities of the robot and (2) cannot perform co-painting without making unrealistic edits to the canvas and ov
datasets, surpassing the previous method with +5.7% and +7.0% gains. Moreover, on the field test with participants, our weara
obile base to monitor the human pose. Additionally, we introduce an adaptive pulling planner responsible for guiding the individu
nifold theory, our approach enables the incremental learning of all parameters of position and orientation VMPs within a library
ss, where the SuperLimbs provide an astronaut with assistance in post-fall recovery while keeping the human-in-the-loop schem
orcement learning baselines via a number of generalization evaluations in simulation and in the real world. We further evaluate
t of these biases within human-robot interactions to inform strategies for risk-aware planning that cultivate appropriate levels of
diction (MoD-LHMP). To address the limitations of prior work, especially regarding accuracy and sensitivity to anomalies in long
continual learning, semantic reasoning, and interactive machine learning literature and develop a novel interactive continual lea
ain and evaluate our inpainting method, we construct a novel large-scale portrait drawing dataset, CelebLine, which composes
e aim to identify the best training method by comparing users trained on a simple task, in a simulated environment on a 2D disp
practical value.
wing smooth and physically consistent strokes to the digital strokes, and two oriental ink painting styles termed as Noutan (shad
e operator can effectively “see through” walls into other rooms. To control robots and send them to new locations, we p
y-states. We evaluated our approach using a four-degree-of-freedom spring-loaded biped robot in simulation and hardware, and
granular terrains. The model can be further integrated with the motion control of bipedal robotic walkers.
th certain existing methods.
tegral Probabilistic Metric (IPM), namely the Wasserstein-1 distance with a novel soft boundary constraint to stabilize the trainin
mplementary constraints, the EBG can search for a near-optimal solution after several iterations within seconds for different beh
n be evoked. Experimental evaluations show that noxious heat stimulation can be detected online and evoke the NWR. The exp
d for estimating lead time. Evaluation metrics, including false positive rate, lead time, and response time, demonstrate the effic
ars with tunable stiffness for controlling the surface curvatures. The outer soft skin layer envelops the patches, forming a human
integrated into a two-level torque control architecture aiming at task-space torque control. The performances of the proposed a
ese two types of domain gaps and propose a novel uncertainty-aware vision transformer to effectively relief the Deployment Ga
sed Alignment (SFA) module to resolve extrinsic conflicts via unifying spatial distribution in BEV space before fusion. Moreover
exploit holistic perspectives from both vehicles and infrastructure, we propose Multi-scale Cross Attention (MCA) and Camera
r that reasons about temporal dependencies among them. We evaluate against different baseline architectures where our prop
his framework helps enhance model generalization across both clean and hazy environments. Comprehensive experiments on
obally generalised representations of morphological information (spatial characteristics) of road defects and colour and depth in
R categories and operating frequencies. In this paper, we first utilize OpenCDA and RLS (Realistic LiDAR Simulation) to constr
om four parts, including the input image resolution, image backbone, view transformation, and occupancy prediction head, it is
d extracts multi-level feature representations, and secondly, it utilizes a transformer-based network to predict human keypoints
e issues, we propose MonoOcc. In particular, we (i) improve the monocular occupancy prediction framework by proposing an a
amework, we propose ShaSTA which learns shape and spatio-temporal affinities between tracks and detections in consecutive
Distribution-aware Aggregation (FDA) framework for cross-domain learning to mitigate the above Distribution Gap in multi-age
and introduce ADVGPS, a method capable of generating adversarial GPS signals which are also stealthy for individual agents
methods were incompatible and their comparison was not possible. In fact, and perhaps surprisingly, conventional forecasting m
re perception as a typical practical application scene, the experimental results on the real-world dataset, DAIR-V2X-Seq, demon
ture. The fused results are input to a statistical filter based on the relative sparsity of noise to achieve defogging. This paper is
aggregation framework called CenterCoop, which first encodes the informative clues from the whole Bird's Eye View (BEV) con
man Filter-based tracking algorithm. In their approach, the measurement uncertainty of different sensors from different connecte
situations arise: 1) Closed-loop testing in-distribution, where the test environment shares properties with those of offline training
n tasks given a human video in a different environment. To allow for generalizing zero-shot across environments and embodime
ns before the task is done. Performing the best action at each time does not mean an optimal trajectory in the end. We need to
os and code are available at https://www.robot-learning.uk/dinobot
ation inputs into a well-shaped and informative reward function. We propose a technique Rank2Reward for learning behaviors fr
s, such as inducing the same ranking over behaviors, distribution over choices, or other related definitions of what makes two re
etwork is introduced, with the aim of addressing the multimodality in human policy. Moreover, task parameter weighting is cons
mized models have been evaluated in an online fashion using the autonomous driving state-of-the-art simulator CARLA and an
actical sampling-based approach that is applicable in higher-dimensional configuration spaces, but only produces a sequence o
al axis is developed. Finally, a multi-fingered hand with a piece of cloth attached to the fingertips is developed, and the effective
hed, and the maximum available workspace index (MAW) of the spherical mechanism is proposed to evaluate the workspace w
we proposed Curriculum-based Sensing Reduction to enable the actor to start with the same rich feature space as the critic an
scovering these behaviors more completely to the agent. Unlike these simpler controllers, geometric fabrics capture a much ric
ducted to illustrate the accuracy of the mechanics model and the effectiveness of the robust planning framework in the presenc
o-explore finger-gaiting in-hand manipulation skills without the use of an exploratory reset distribution.
ot, is applicable to various in-hand object rotation tasks. Comprehensive ablations are performed on how the integration of visio
ereby the fingers move in synchronization using an intermediate dynamic responsible for coordinating fingers. To adapt to chan
on cups during locomotion, as well as the improved gait trajectory, was established to ensure successful climbing. The control
pose to the robot's controller for execution. Grasping efficacy is assessed across both synthetic and real-world settings. Given o
a laser tracker absolute measurement system show a $95%$ improvement in absolute accuracy after two trials. When the lase
ontext of refabricating fiberglass wind turbine blades.
an LLM-driven scheduling module that facilitates efficient allocation considering task scenarios and vehicle information, which a
patial consistency. Specifically, our proposed model incorporates three levels of consistency: channel level, semantic level, and
anual setup, where the microscopic camera was controlled to view a tool moving in a pre-defined trajectory. The proposed met
tions and the prior work.
environment, as well as selecting the area of the asset surface to process. The teleoperator interfaces with the robot platform by
ach, we are able to demonstrate real-time receding-horizon SNMPC with statistical safety guarantees in simulation and on hard
ding modules that first perceive, then predict, and finally plan. Instead, we shift the paradigm to have the planner query occupa
safety filter is experimentally validated using a wheeled ground robot in obstacle-cluttered environments. The result shows that
n mechanism. Based on the online sampled primitives from the past of the current freight trip and anchor-based future data syn
urbances. Both of these contributions are integrated into a robustified motion planning and control pipeline, the efficacy of which
his combined approach enables us to identify multiple distinct local minima and provide an improved initial guess. We validate o
on of long-horizon robot motion paths across different map layouts. Moreover, our approach also demonstrates capabilities for
aller prediction horizon when compared with the standard MPPI for similar success rates.
ed framework not only reduces planning time but also increases success rate in comparison to previous approaches.

nd potential benefit of prioritizing task-relevant data. The experimental analysis shows how algorithmic properties affect perform
ile of the vehicle. The model is trained in a self-supervised fashion using recordings of vehicle trajectories. We then present an
he-art reinforcement learning algorithms by being up to 46 times faster while simultaneously being less dependent on reward sh
We provide extensive experiment results with three different motion planners on three classes of planning problems with over 3
ure the environmental cues, such as route and traffic lights, in vectorized representation, a decoder utilizing iterative route fusion
aper, we propose a Multimodal Neuron Planner (MNP) based on the mixture density networks that explicitly takes into account t
arse-to-fine manner, producing sparse subgoals when the task is easily accomplished by MPC and more dense subgoals when
mple at each timestep. The paper demonstrates the approach on several simulated examples as well as a 7-degree-of-freedom
detects ambiguities related to symmetry and uninformative points and conveys this uncertainty to the user. Perhaps more impo
ore we propose to identify a Hidden Markov Model (HMM) from a fully instrumented measurement set, that is only available for
mances, making comparisons less expensive. Data collected are automatically published on a web dashboard (WD) that provid
iminates the need to use intensive numerical solvers—for example, methods based on finite element analysis— and allows th
ed in this study achieves a spatial resolution of 68 um per pixel. The custom-built sewing machine, controlled by an external com
niform machining strategy based on the exact tool and workpiece geometry. Further, a method for planning coverage paths is p
. Results show that the proposed integrated model outperforms models that consider only system-level parameters, as well as
onnection between the on-board and edge-devices. We show the effectiveness of our approach through extensive testing in thr
utlets along with path finding. Furthermore, we optimize the performance of CBS in dense traffic by defining the concept of follo
tiple network infrastructures, and with different centralized/distributed learning schemes. Our intensive experiments on three da
and the first ODD-based query-time scenario mutation framework to address the limitations. Our case study results show that
man supervision requirements. It further shows that AV cooperation can improve supervision reliability by orders of magnitude
igation. Our method aims to replicate human driving behaviors by learning parameters that emulate human decision-making du
pliant predictions by generating future trajectories in a coarse-to-fine fashion, where the coarser predictions at a lower frame rat
of the reformulated optimization problem is exhibited. Then, a parallel optimization algorithm is presented to solve the reformul
rajectory while ensuring realistic and human-like trajectory generation through learning from real-world driving data. Specifically
and their driving style in terms of minimum acceptable risk. The cognitive digital twin of the driver is then created and updated b
ng strategy is needed to enhance the detection performance for critical pedestrians. In this paper, we propose a novel, safety-a
Cylinder Balanced Random Sampling method that enables the downsampled point clouds to maintain a more balanced distribu
story of detections and is jointly optimized for both data association and state estimation tasks. Since the standard tracking met
anism to facilitate CAV data fusion. Our approach begins with optimizing the connectivity of vehicles while considering commun

cess, we utilize a coordinate transformation strategy that converts the QSLS into a configuration that resembles the dynamics o
tate information in order to determine whether vehicles are autonomous, without requiring any active notification from the vehic
by a motion estimation module. Evaluated on the View-of-Delft dataset, RaTrack showcases superior tracking precision of mov
ehicle positions and velocities. Obtaining this information requires updating existing road infrastructure with vast sensor environ
ure language input, our novelty lies in two aspects. 1) The end-to-end integration of visual and LiDAR sensory input into learnab
ating differentiable joint learning. We validate our proposed approach using the real-world nuPlan dataset and its associated pla
casting (SIMMF) method to capture semantics along with spatial information and optimally select relevant agents for motion pre
et of state-of-the-art deep-learning models on our proposed benchmark and find that all evaluated models exhibit large shifts un

ed by conventional control methods and residual terms which are solved using learning. We employ MRTL for fleet-level emissi
o the large number of taxis required for stability. In this paper, we aim to address the computational bottleneck of multiagent ro
sferable information for current AV improvement. Therefore, we develop a continual driving policy optimization framework featu
SOTA methods.
designed modules named coupled map and reference extractor. A novel multi-task optimization strategy (MTOS) is presented t
low-level trajectory planners. For the numerical evaluation we combine our decision module with a motion-primitive-based and
l by changing the soft-max layer of the deep model to eMem-NDT, for grouping and aligning the memory prototypes of the histo
current state, which lacks the prediction of the potential risk in a sufficiently long future horizon. In this work, we study the probl
’ model parameters. The proposed framework employs a sampling-based approach that is suitable for real-time implementa
y RL models. We successfully demonstrate high completion rates of complex scenarios with recent model-based agents.
sure a high-quality policy. The proposed framework is deployed and evaluated in numerical simulations of dense and dynamic
nal Neural Networks (CNN) and Graph Attention Networks (GAT) to capture spatio-temporal dependencies among agents. Ben
atures exhibited by agents on the scene and the physical environment constraints to generate map-compliant and socially-awa
ion method combining convolutional neural networks (CNN) and long and short term memory (LSTM) is proposed, which is app
where the operator's intent is susceptible to change. We derive a novel shared-control solution that combines RT's predictive ca
uracy and realtime performance. In pre-recorded real-time settings (160 Hz), a peak accuracy of 86.88% is achieved, surpassin
ation, we design an optical flow refiner. In the second stage, we adopt an encoder-decoder network and let it learn to magnify t
s for both a transformer and a long-short-term-memory-based architecture, we empirically show that jointly utilizing in-cabin and
te that a robot using our method CARTIER (Cartographic lAnguage Reasoning Targeted at Instruction Execution for Robots) ca

destrians. To this end, we propose a novel egocentric trajectory prediction model that benefits from multimodal sources of data
fixation data using a weighted Naive Bayes Classifier (NBC). Acknowledging the potential variances in saccadic eye motion du
t tendons. We study the dynamics of impact and use the stiffness of the tendon as our design/control parameter to adjust the m
static actuation from the pair of low-voltage self-sensing electrodes. This approach allows for proprioceptive feedback control o
dering system design considerations, cantilever beam design, cam design, and spiral cam design. The analytical methodology
control scheme and a low-cost elastic robot to perform scrubbing. This study observes the relationship between joint stiffness in
ple solutions to the kinetostatics of cable-driven continuum manipulators with constant cross section or variable cross section. T
and minimum principle, the different transmission path solutions of the minimum tension loss, the minimum tendon deformation
rion is also established which links the critical tendon force, until which the robot is stable, to the design parameters. Finally, the

rque transmission characteristics of tendon driven actuator systems. Additionally, the precise model exhibits excellent conformi
g in a decoded representation of the original data. Beyond offering a comprehensive solution to kinematics challenges, our met
ped relate to multiple factors. Then the jumping performance is improved by optimizing the energy storage modality, rowing traje
i-dimensional path is approximately tracked over time. In order to improve the tracking accuracy, we propose an iterative schem
characteristics of the bisecting plane of a single segment. The degenerate cases for the solutions are also addressed. These ou

nuous and differentiable, characteristics that are essential for modelling smooth and predictable system behavior. We demonstr
roposed approach, we generate an optimal robot design along with the required path to complete the task. The approach could
invariant to the choice of body reference point. This new representation is validated for a self-supervised segmentation approac
h observe this constraint. By composing a wide enough variety of such deformations, many of the most common behaviors obse
cal system (DS) for motion generation and a control system that regulates the directional inertia of the robot is presented. Furth
such as scalability and redundancy. We then demonstrate the viability of our design with a proof-of-concept implementation of
nematics nonlinearity. By introducing the virtual passive joints, both displacement and velocity analyses are readily investigated
ome singularities of the CDPRs and the limits of stability and (ii) show that singularities of the platform wrench system are not s
ntally on two cable types and configurations.
' torque-speed relationship and becomes the wrench-dependent available twist set. Then a new workspace condition and a new
ying Linear Quadratic Regulator. With RicMonk we demonstrate bidirectional brachiation, and provide comparative analysis with
GP-enhanced control experimentally using two examples of underactuated balance robots.
e the conservativeness by adding a magnitude regularization term to the synthesis loss. We propose an algorithm using reinforc
inforcement learning policy. Moreover, to mitigate the myopic nature, we propose a novel multi-step training and single-step ex
by augmenting base skills by learning failure prevention skills with reinforcement learning and forming a skill library to address
applicability of these predictions is an unsolved problem. We present a system that utilizes a Large Language Model (LLM) to i
hile the proposed model achieves the same task success rate as the conventional model, it clarifies that it refers to vision durin
feet. We aim to train this network using the actual robot data and utilize it for tool-tip control. Additionally, we employ Parametric
ness between the variables of two CBNs to generate and transfer informed CBN distribution priors. We evaluate the parameter
to propose three new CMA-MAE variants that scale to high dimensions. Our experiments show that the variants outperform ES
more accurate imitation learning with only monocular images by extracting motion constraints from one exemplary motion of ski
ith an Imitation Learning loss, enforces the aforementioned stability in the learned motions. Differently from previous work, our
nce obstacles are everywhere in daily life. In this paper, we propose a novel transformer and GAN-based Editable Movement P
on approach by looking at the different configurations of the objects and goals present in the scene, learning the association be
and bottle-lid opening task. The experimental results demonstrate the validity of using differentiated forces and torques for mot
ask execution safety. This letter studies the mutation of a robotic skill when it is propagated from one robot to another during a p
the benefits of learning skills from humans. We validate the approach on the wiping task using sponges with different stiffness
codes it into a latent embedding and is trained with sufficient expert demonstrations, predicting states in the potential subspace
supervised fashion. ConBaT is inspired by the concept of control barrier functions in control theory and uses a causal transform
ard, especially for unseen worlds. This work aims to learn generic, sparse, and context-aware rules that govern action primitives
ded joint movement, i.e. the arm falls down on its own, which is unacceptable during direct teaching. Therefore, this paper dem
ning methods. WayEx obtains a higher reward than existing imitation learning methods given only a single demonstration. Furth
umber of surgical demonstrations and interpolate learnt behaviour via a parameterised skill model. This enables a large numbe
cing term. We exploit this increased expressiveness by automating the fitting of the dynamical system parameters through optim
, augmented by diverse and extensive in-the-wild data collected by AirExo, robots can learn a policy that is comparable to or ev
parisons over all demonstrations, and 3) adapts to cases where there are varying levels of constraint violation. Our empirical res
hesised as a generative model of motion trajectories in 3D task space. Additionally, we present the Ray-tracing Probabilistic Tra
hat our method is able to learn precise relative placement tasks with a small number of multimodal demonstrations with no hum
e simulations and deploy the trained policies on a real-world manipulator arm. The results confirm SNDS's ability to address inst
r approach can handle obstacle shapes that fall outside the scope of assumptions typically found in the literature concerning ob
a single demonstration (a prompt) to learn policy that adapts to the testing target environment. Our proposed framework, Prom
e LASA handwriting dataset and complex periodic trajectories. It is also validated on the Franka Emika robot arm to produce st
er employs a simplistic neural network architecture that excels in fulfilling a dual objective: optimizing accuracy while simultane
ms. With the learned neural energy function, a globally asymptotically stable ADS with predefined position constraint is further f
onstraint. Evaluating our method and five baselines in three simulated environments, we show TCL outperforms state-of-the-ar
olution. During the process of ERTC-HD, robot motions generated from extracted comprehensive features of human motion are
we propose to use IQL-TD-MPC as a Manager in a hierarchical setting with any off-the-shelf offline RL algorithm as a Worker. M
pproach for skill discovery with a particular focus on robotic manipulation. Our main insight is that utilizing multiple critics in an a
system without needing large amounts of data from physical cutting trials. The simulation approach was validated on a real rob
richer repertoire of skills that can help an agent generalize to new tasks. Experiments in a household simulator and on a real ro
collaboratively estimate the Q-function. This cooperative process utilizes the TD error to steer updates to the Teacher Encode
viors including sprinting, jumping and sharp turning. Various behaviors can be activated through human interaction using a stick
rivileged information commonly garnered from a simulator to significantly accelerate sim-to-real transfer. Specifically, a teacher
olicy. We deploy our policy on a physical quadruped robot with an RGB-D camera on pursuit-evasion interactions in the wild. D
pose a simple yet effective method, called weighting online decision transformer with episodic memory (WODTEM), to improve
bot and formulate the control of the snake robot as a cooperative Multi-Agent Reinforcement Learning (MARL) problem. Each s
ree Go1 quadruped robot. Our results demonstrate that our method leads to 1.4-2.8 times faster convergence and as low as 50
tive framework designed to selectively identify beneficial data augmentations, thereby achieving superior sample efficiency and
hat satisfies their inference compute constraints. We show that our method scales well - more training resources produce faste
g with APRL results in a policy that is substantially more capable of navigating challenging situations and adapts to changes in
remain upright by compensating for gravity. The paper showcases the first successful sim-to-real transfer of a torque-based de
ups that can be decoupled from each other while preserving essential connections and then learn a decentralized motor control
de the dataset are useful for a variety of tasks in the field of off-road autonomy and, by releasing them, we encourage collabora
seriously deteriorating the UOD performance. To address these challenges, we introduce EnYOLO, an integrated real-time fram
sis to obtain high-fidelity vehicle mobility data on sampled terrains. Following that, the mobility data is utilized to train a PSO-krig
and odometry error. Our approach is label-free, and is able to produce traversability estimates that are robot-agnostic. Addition
at implements geographical localization and can seamlessly integrate with SLAM systems. Compared to GPS, GLN is less sus
age traversability estimation framework for various robots in diverse off-road conditions that integrates robot properties in additio
nt learning is proposed. This function can anticipate potential challenges or opportunities based on semantic information for the
TRajectory generation with dIverse conteXts (MATRIX). MATRIX is capable of generating interactive human behaviors in realis

gh-quality training data through standard GC policies, which results in an inaccurate distance function. Both affect the efficiency
model using contrastive estimation to predict future keypoint states. By learning to perform physical reasoning in the keypoint sp
ons. Results show IGE-LLMs (i) exhibit notably higher performance over related intrinsic methods and the direct use of LLMs in
but this often compromises stability and robustness, especially as the dimensions of observation and action increase. This pape
mulated and real experiments have demonstrated that the proposed method is effective and can generalize to unseen instructio
ng Model for promptable object segmentation. Our trained policies adeptly grasp a wide variety of objects from cluttered scenar
n by sequencing learned skills. We validate our approach on real robot hardware by learning and sequencing two manipulation
of the observation of the world-state, but perform poorly due to the weak reward signal. In this work, we blend the approach of a
ecifically, we parameterize the stochastic policy as the perturbed output of the MPC controller, thus, the learned model represe
across different task scenarios. Then in reinforcement learning, TEP assesses and guides the Actor towards better outputs and
ries consisting of a set of randomized sub-skill segments, enabling the acquisition of rich robot trajectories that capture skill stag
conduct our experiments on diverse real-world objects and show significant improvements over baselines. Our project page is
d an attention-based network to effectively combine them and provide rational task scheduling. Furthermore, a cost predictor is
p of the pre-trained visual representation using a small amount of source task demonstrations. We show that the learned policy
tions such as sensor noise, perception failures, physical perturbations, observations mismatch, etc. To achieve this, we reformu
approach, two deep reinforcement learning (DRL) agents orient and position a robot gripper to deform a DLO into the desired s
a dynamic memory to efficiently learn new observations and 2) a static memory to balance new-old knowledge. When the agen
bouring Space Scoring(PNSS). The PNSS model estimates the explorable space for a given position of interest based on the c
ate the effectiveness of our method on a local navigation task where a quadruped robot needs to reach the target within a finite
Thus a distributed training pattern to mitigate the efficiency reduction was adopted. We evaluated the navigation performance o
ansformation between the radar scan and the map. We tackle the place recognition task by learning a shared embedding spac
ffectively encode the geometry and texture of edges by the hybrid structural grid. We propose a novel local bundle adjustment
e efficiently in new scenarios unseen in the training phase. To equip the navigation agent with sample-efficient learning and {zer
plexity-based navigation method based on hierarchical deep RL with the proposed metrics. Experimental results show that the p
e-agent topological exploration, and multi-agent exploration for indoor mapping, across different input modalities. Our work moti
diverse nature of the tasks offered by human-robot gym creates a challenging benchmark for state-of-the-art RL methods, (b) by
s each agent's experiences, improving its adaptability to unseen crowd behaviors. In assessing an agent's robustness against u
the pipeline against variants of an occluded pedestrian crossing scenario in the CARLA Urban Driving simulator. Specifically, t
xisting works that require convex approximations for the objectives and constraints, this work exploits a first- order method to a
have studied this topic mainly in a single-task setting, where symmetry usually refers to symmetry in the motion, such as the ga
ity of reinforcement learning (RL) in long-sequence task decisions, we propose an online task assignment decision method bas
 (e.g., collisions), which makes it prohibitive for the policies to learn safety concepts. Besides, these bulk policy networks canno
y, we exploit the system's inherent symmetries to augment the training data. Consequently, our approach learns to successfully
which are time-consuming and sub-optimal due to their handcrafted design. Deep reinforcement learning (DRL) has shown gre
cts a binary segmentation mask that identifies the specific points requiring adjustment, along with estimating their correspondin
edom and gradually releases them, using behaviour cloning (BC) from an early learning procedure to guide optimization in late
onally, our method simplifies the traditionally labour-intensive task of reward shaping, boosting the efficiency of the learning pro
gged locomotion application. We show the efficacy of the algorithm in tuning the high-dimensional controller parameters and als
e measured the transition success rate of 72 transition pairs and achieved an average success rate of 99.99%, which is over 1
iple modalities of human input. We for the first time demonstrate various bipedal motions on a quadrupedal robot, and showcas
ON first leverages a pre-trained model to generate a human motion from a language command. Then an RL phase trains a con
um learning to allow a single network to conduct both normal gait and various joint-impaired gaits. We verify that our framework
nt faults. The architecture simultaneously trains three joint task policies for health, power loss, and locking scenarios in parallel,
ous skills, each with its unique goals and observation space. Finally, we develop residual regularization for learning policies tha
tempting to carry out their main tasks. Through an in-depth analysis, we highlight both the compliant nature of the resulting con
conducting comprehensive simulation and hardware experiments, we demonstrate that the proposed IFM framework can signifi
ve approach unifies them, addressing their lack of synchronization. At the heart of our contribution is the synthesis of stance fo
pth camera for perception which is low-frequency, jittery, and prone to artifacts. We show how a single neural net policy operati
mal Policy Optimization (PPO) to derive our method, Distributional Proximal Policy Optimization (DPPO). The risk preference, ra
ncertainty of environments, and perform risk-averse policy learning by optimizing the worst-case scenarios via a risk distortion m
the secondary objective of minimizing energy consumption. It seeks a policy that minimizes the energy consumption objective
dequate to navigate robots through challenging confined 3D spaces, requiring both agile 3D obstacle avoidance and robust leg
tories that simultaneously satisfy the task specification and maximize the locomotion robustness. Our framework outperforms a
azelle, which confirms its effectiveness in real-time.
ally forward invariant. Lastly, we experimentally demonstrate that this approach results in successful locomotion for both flat-foo
n and on hardware. These results include validation of this controller’s ability to perform fast, reliable forward walking at 0.7
motion phases that can be of different types: stance, flight, transition phases, etc. We introduce a unified formulation of the CoM
moving forward, backward, or walking in place. Compared to traditional template model-based approaches, our framework exh
ent unit (IMU), joint encoders, force/torque (F/T) sensors, and motion capture ground truth to evaluate the proposed method. Th
tests, with a maximum ratio of 2 on flat terrain. Experiments also demonstrate Barry’s power consumption and cost of tran

outu.be/1NKQ0rSV3jU.

balance during trotting. This paper explores this phenomenon using a rat robot equipped with a soft actuated spine. Based on t
ions, and/or expensive motion- capture equipment, all of which limits scalability. Instead, SLoMo only relies on easy-to-obtain v
motion control framework. The Carpal-Claw functionality is validated through various experiments on a very high obstacle and st
eal-world experiments show that SpaceHopper can successfully reorient to a safe landing orientation within 9.7 degree inside a
o effectively train a single locomotion policy capable of controlling a diverse range of quadruped robots. The robot differences e
arameters and gait configuration on its motion stability. These results provide a theoretical foundation for the design and stabilit
ipulation. By deploying our controller on a quadrupedal robot using teleoperation, we demonstrate various real-world tasks suc
ntly allowing for rapid policy updates through supervised learning. We also develop a high-level network to track diverse comma
enerated by MPC. By such an approach, our proposed controller can provide more accurate inputs to the system, enabling it to
y and an agreement protocol. At the lower level of the control scheme, distributed nonlinear controllers are developed to impose
mplifications. We validate the robustness and controllability of the RL training process within our framework through a series of e
inear model is discussed. Although the nonlinear model has only one nonlinear term in the gravity term, numerical simulations s
sumption. The performance of the proposed limb trajectory planning strategy is experimentally validated using the same leg-wh
of each iteration. Simulations and experiments on a quadruped with parallel springs demonstrate that continuous jumping can b

within a factor graph framework. The generated map is not only memory efficient, but also offers higher reliability and speed in b
ntal solution matrix that characterizes the evolution of perturbations through a hybrid trajectory and its 2-norm, which represents
acceleration trajectories, creating a motion library for the system. An online linear model predictive control with whole body con
ly, we introduce a guiding torso path heuristic at the global planning level to enhance the navigation success rate in the presen
r failures based on historical information, showing that the framework is adaptable to changing environments and is suitable for
everages an asymmetric actor-critic structure, enabling the utilization of privileged information from simulated environments dur
well-demonstrated for observing the inherent characteristics, exhibited the locomotion characteristic of the model proposed. By
ork to the popular nuScenes dataset to provide efficient and highly accurate annotations. Compared with the original nuScenes
car-based data collections, our setup can simulate any robot (quadrupeds, quadrotors, autonomous vehicles). The dataset inc
l's adaptability to complex scenes. Meanwhile, we propose a multi-scale sub-pixel lightweight decoder that generates high-qua
and granular feature representation derived from a general-purpose model. Moreover, our representation is based on tracking

lign directly with graded similarity labels, this approach enhances ranking capabilities without expensive re-ranking, offering dat
To achieve this objective, this study compares VPR learning approaches, conducts a comprehensive empirical analysis, and p
on. Furthermore, we introduce a time-constrained graph attention aggregation(TC-GAT) that propagates node information acros
n Transform(CIT) module to make these two feature spaces interact knowledge with each other to enhance the feature represe
al measurements in an intuitive way with a given small number of bits and then perform an EKF update with these quantized me
ntic Label Filter, Dynamic Object Filter, and finally, utilizes Voxel PFilter for maintaining a consistent global semantic map. Evalu
g the map using this structure allows for end-to-end solving of maximum a posteriori (MAP) based probabilistic mapping. Globa
umption from negatively affecting the depth predictions, we propose a novel masking approach designed to filter out such pixels
ured over long trajectories. We demonstrate the reconstruction system with data from a multi-camera, lidar sensor suite onboar
nting high-quality data association for parameterized semantic features, allowing continuous map update and refinement during
object and the ground where it is located, dynamic objects are classified into two types: suddenly appear and suddenly disapp
l as new observations are added. Experiments on synthetic and real-world datasets demonstrate that our method can generate
e voxel grids and to update their occupancy states in parallel, and it exploits CPUs to restructure the PVM using the updated vo
ection based on the pose estimation of stairs is introduced to calibrate the output of raw modeling that is influenced by tilted per
astically outperforms recent methods in predicting traversability for both on- and off-trail driving scenarios. We compare our me
er layer. The GFN interprets the surface and semantic information from the input data and integrates them to extract features o
able our model to predict an unbiased local directional soft lane probability (DSLP) field in the limit of infinite data. We demonstr
and student models, maintaining superior performance while enabling minimal computational load during deployment. We condu
map. This method required only 878 msec on average to perform global localization and outperformed state-of-the-art global re
that our method achieves very promising results in dynamic environments. The implementation of our method is available as o
ating learning representations with the above-desired characteristics. With these two auxiliary tasks, the agent learns a spatially
on masks of the image sequence generated by a frontend instance segmentation model, and associates corresponding masks
PM (MC-IPM) technique is implemented, capitalizing on motion for distortion compensation and subsequently selecting the mos
 to Branch-and-Bound Search (BBS) in the 3D map is introduced to provide the initial pose. Finally, a graph-based optimization
eatures from a 2D radargram. With DECs features obtained, a metric subsurface feature map is constructed. Finally, we perform
are subject to extrinsic calibration and time synchronization precision. In this paper, we propose a novel cost-effective reconstru
elects patches with complementary image information. Photometric error minimization in the image patches is then fused with i
method is capable of evaluating one million particles in real-time on a single GPU and enables robust pose initialization and re-l
or system initialization without prior position information. Experimental results show that the proposed method outperforms exist
dataset with 6.7-7.6 km routes driven in similar and opposing directions under various lighting conditions. The proposed algorith
show that the proposed method achieves the state-of-the-art on the tasks of homography estimation, pose estimation and visu
we show that our approach consistently outperforms state-of-the-art methods in various real-world settings. Our code is availa
h the SLAM system. To address these issues, we present a semantic-based cross-view localization method. This approach com
er. Additionally, we propose a learning method to estimate the photometric-sensitive depth uncertainty for guiding the scale rec
allenging aliased environments. Since our work aims to enhance the reliability of SLAM in all situations, we also explore its perf
nt GNC-based optimization. RELEAD has undergone extensive evaluation in degenerate scenarios and has outperformed exist
spatial partition, which efficiently captures important semantic information within the channels and preserve spatial relationships
ntegration measurements are tightly integrated using the factor graph optimization, with online calibration of the LiDAR-IMU ext
se registration. We evaluate VioLA on two real-world RGB-D benchmarks, as well as a self-captured dataset of a large office s
tly learn how to effectively observe the first-person-view (FPV) world to interpret bird-eye-view (BEV) maps. Providing explicit g
n (HFA) module is further designed to aggregate multi-scale and multi-level features in the pose regressor. Benefiting from thes
ational efficiency. In this paper, we propose a novel approach that achieves state-of-the-art inference accuracy with significantl
per: (1) A NeRF is designed to separately predict uncertainties for the rendered color and depth images, which reveal data relia
oduce SeqGeM, an aggregation layer that revisits the popular GeM pooling to produce a single robust and compact embedding
sferable representations of places. When transferred into new environments, our CCL continuously reviews the contrastive mem
nsformer autoencoder applied on depth images. This framework not only furnishes accurate robot state estimates, including un
esults on the large-scale publically available datasets demonstrate the effectiveness of our approach in scenes lacking distinct f
cant changes in the scene. The addition of a Bayesian filter boosts the accuracy of long-term place recognition, enabling reloca
e localized. The method is evaluated on a three-coil system, and simultaneous actuation and localization is demonstrated with
pairwise overlap among the narrowed-down place candidates to determine the final match. Experimental results on the KITTI a
nd online pose regression network is designed based on optical flow estimation and spatial attention to achieve real-time monoc
istance field. Additionally, we show how to perform scan registration and global localization directly within the neural distance fi
ound place recognition. We demonstrate the superiority of our proposed method by comparing it against several recent state-o
LAM measurement by solving a scaled pose graph problem. To further increase the localization accuracy, we provide an iterat
odel capacity. EffLoc excels in efficiency and accuracy, outperforming prior methods, such as AtLoc and MapNet. It thrives on l
n improve localization accuracy in large-scale scenes even if the semantic information has certain errors. Experimental evaluat
e, which is flexible and computationally efficient. In addition, we take advantage of the uncertainty for pose refinement to enhan
underlying 3D geometry of the scene and use the geometry, in return, to estimate a 6 DoF pose in a geometric manner. We pre
the query image and resolves spatial symmetric ambiguity. The similarity scoring module sorts out the top-scored database im
m 6DoF pose estimation. In the online phase, VOLoc proposes an online Geometric Recovery Module (GRM), which is compo
nes the timing of a single spike based on a pixel’s intensity, as opposed to prior SNNs relying on rate coding that determine
monstrate the proposed one outperforms the traditional 17-point algorithm in term of accuracy.
ns derived from off-the-shelf self-supervised models with no VPR-specific training are the right substrate upon which to build su
databases are tuned for lightweight applications. We demonstrate the approach by running the whole system on a low-cost sing
mages within an efficient filter-based framework. This tightly-coupled fusion enables efficient 3D motion tracking with bounded e
nsistent state estimation at night. Furthermore, a tracking recovery module is also designed for tracking failures. Experimental
alysis and show increased localization accuracy in comparison to state-of-the-art (SOTA) tightly-coupled algorithms.
nd semantic cues. The semantic information is extracted from images by deploying a state-of-the-art object detection model on
Each trajectory includes 20 to 50 seconds of recorded sensor data and ground truth labels. Data from different sensors can be u
ndings and examine their impact on the estimation accuracy and consistency. Finally, the system is implemented to achieve eff
low-cost robot system deployment in real-world applications.
ed method has only one insensitive tuning parameter and is agnostic to the configuration of the IMU. The system was validated
onally constrained micro-aerial vehicle (MAV). We validate our approach through closed-loop flight tests involving dynamic traje
nd shifts, and robustness to faulty inertial measurement units (IMUs). We propose a novel equivariant filter (EqF) formulation for
t a novel Levenberg-Marquardt (LM)-family trust-region adaptation of the iSAM2 fusion algorithm, which is optimized for robust
of leveraging these radio maps as a substitute for floor plans in multimodal IPS. We introduce a new framework to address the
tions using a Vision Language Model (VLM). By leveraging detailed text information, our approach efficiently extracts correspon
lihood functions obtained from multiple observations. The proposed method, termed as the multiple update PF, leverages prior
point clouds, lack these features and require frequent updates in response to environmental changes (e.g. moved furniture), un
ilable, we integrate human pose data using a temporal convolutional network (TCN) to compute spin priors within the factor gra
biases. Its short-term accuracy and fast operation (~5ms) make it a proper candidate to supply the initialization to more comple
ta and Kalman filter sensor fusion to 3D scan-to-scan and scan-to-map registration, on three datasets with data from two recen
f 3D scene graphs for locally consistent correspondence searching and global substructure-wise outlier pruning. Such a hierarc
ased VINS, which seems counter-intuitive but improves the estimation performance in both consistency and accuracy. Capitaliz
ments with participants on two handover tasks, with results indicating that the proposed method leads to seamless and pleasura
o obtain the data-driven ability. It also integrates rule-based verification for safety insurance. We conduct a series of experiment
rned hyperparameters. In the second phase, we propose a safety filter based on high order control barrier functions (HOCBFs)
esentation over time. Extensive simulation experiments with diverse objectives demonstrate the generalization capabilities of D
m decisions through the MPC-based actor and long-term prediction via the critic network, unifying the benefits of both model-ba
pose to perform fully reactive closed-loop planning over these discrete latent modes, allowing us to tractably model the causal in
ject parameters. At the time of inference, given a new object, GenDOM can estimate the deformable object parameters with on
er with a variety of sizes, weights, and initial configurations. In addition to quantitative simulation results, we demonstrate succe
ss that relates motion prediction from the dynamics model with motion prediction from point-cloud registration to train a Hamilto
get stuck in bad local optima, depending heavily on the quality of the optimization algorithm. To this end, we propose Deep Mod
former (SIT) and an Adaptive Dynamics Model (ADM) under a variety of simulated dynamics. The SIT uses attention mechanis
ment of robotic RL methods, is comparative inaccessibility of such methods. To address this challenge, we developed a careful
raint manifold by approximating the normal space of possible motions from the data, which we use to train a Gaussian process
L, based entirely on learning value functions via temporal-difference learning. We show that value learning on video datasets le
ysical rover model and performed maneuvering experiments on a steep granular slope of poppy seeds to explore the rover's tur
t Learning, we utilize a decoupled RL framework to learn two policies, (1) a task policy that only considers improving the task pe
rearrangement (PMMR) toward optimally resolving the challenge. Extensive simulation and real robot experiments demonstrate
ation poses. Specifically, the scene and language-task information are mapped into a joint scene-task representation feature sp
-hand RGBD camera, ForceSight performed tasks such as precision grasps, drawer opening, and object handovers with an 81
ng, a focused training strategy is adopted with the hypothesis that an agent trained on a small set of representative shapes can
ent, as many grasps will not lead to progress on the task due to topological constraints. Therefore, we propose a grasp loop sig
n the camera is closer to the object. Therefore, we propose a novel coarse-to-fine affordance learning pipeline to mitigate the ef
th-first search improves performance; we show how the search can be run ``eagerly lazily'' to only simulate actions in a physics

line map prediction methods without bells and whistles and can be immediately incorporated into any Transformer-based lane-
The combined features are then decoded into signed distance values through a small multi-layer perceptron. We demonstrate
ocal map to boost model training efficiency and curb the catastrophic forgetting issue. A decoupled scalable global map is furth
the noisy image gradient computing problem in grid-based NeRFs, we further propose a re-devised truncated dynamic low-pas
ve dataset with quadrics segments and objects is collected for training and evaluation. Experiments demonstrate the effectiven
s to precisely and efficiently identify the dynamic bins within point cloud bins, thus overcoming the limitations of prior approache
g keyframe selection strategy to address the forgetting issue and enhance mapping quality, particularly in marginal areas. To th
loss functions. Unlike current state-of-the-art approaches, we explicitly model the uncertainty of the object shapes and poses d
ween pairs of consecutive images and b) connecting neighboring segments within an image using their pixel centroids. This unv
AMs. Map-reading extracts salient information about navigational behaviours from the overlooked wealth of pre-existing, abstra
he map efficiently with the consideration of occlusions. The occupancy status of an arbitrary point in the map space can then be
ation of the dynamic environment using a Bayes filter. This volumetric belief models, which parts of the environment can be occ
nown firing pattern of the scanner. We use this knowledge to connect these points to their neighbors and label them using the a
ve is to jointly optimize the poses of all frames for mapping images onto geometric reconstructions. Our pipeline works in an off
a Gaussian Mixture Model (GMM) are combined with an implicit GP map, also employing local GP-block techniques. The prop
pancy updates. Furthermore, we conduct extensive benchmark experiments on various LiDAR sensors in both public and priva
R-GPE) is proposed for incrementally estimating and preserving ground information, and Scan-to-Map Removal (S2M-R) is pro
e conducted different experiments using these two depth sensors on public benchmarks. Our results show that our system per

havior. The impact of the tunable bistability on the swimming and turning performance is investigated through extensive experim
links connected by angled swivel joints. Each segment is actuated by tendons, which allows for continuous bending. The syste
ater ArmVehicle Emulator (WAVE) is a 10-degree of freedom system designed to replicate an inspection-class UVMS. WAVE in
d modeled. The originally nonlinear actuator performance is linearized by the post-fitting mapping. Asymmetric dead zones of t
itch and roll control. We evaluated Snapp in open water environments and demonstrated significant improvements in speed an
extensive experiments are carried out to examine its thrust generation and speed. It is found that the omnidirectional robot has
diment collection and extensively tested the system both in a tank and in field experiments in different spots along the Tyrrhenia
ain interaction force, removing the effects caused by gravity, buoyancy, and thrust. Second, we gather a walking cycle's discrete
s modeled using path-optimization and can be solved in real-time with a common non-linear solver. The constructed objective f
ater dataset, FLSea. Evaluation results on this dataset demonstrate significant improvement in depth prediction accuracy by th
ith 4 objects with symmetrical shapes and poor texture spanning across 33, 920 synthetic and 10 real scenes. All objects and s
ce for perception sensors. However, they too are not without their limitations. While imaging sonars offer superior long-range vi
d uncertainty quantification in image segmentation. Our method, enhanced by a style modulator, merges content features, and
ter from the scene and apply distinct light physics during training. Our method avoids the need to estimate complex backscatte
c scene parsing of underwater cave environments. In particular, we formulate a novel transformer-based model that is computa
activity within each habitat. We demonstrated the system in field trials on reefs in the U.S. Virgin Islands, where it was able to a
e introduce the untied cross-aware position encoding with principal component analysis and different projection matrices, which

y submerged UVs and to the experimentally observed behavior of an actual vehicle while fully submerged and while interacting
 ealthy initial attitude and inaccessible wave disturbance during takeoff, thus alleviating the thrust distraction from stability recov
ons that do not require precision maneuvering. Furthermore, since gliders only activate the buoyancy engine for small time inte
challenge, we present a robust model predictive control strategy utilizing Control Barrier Functions (CBFs), which increases the
thrust in the water before and after optimization differs by 400% according to the test results. We optimized the minimum differ
er of effects as a ‘sphere’ that a 3D point will project on. Furthermore, the S-NS model includes additional parameters to
orms consistently at high overlap (20-50%). In comparison, GICP (a variant of ICP) performs well for fine alignment and is bett
h support for sim-to-real transfer. We validate this framework by implementing a state-of-the-art control architecture and demon
using a hypothetical `explore-exploit-migrate' scenario in a two hour field demonstration with eight unmanned surface vessels (U

orporate the charge availability constraint at the recharging station in addition to the maximum charge capacity of each AUV. Th
sed Distributional RL based policy against a traditional RL-based policy and two classical methods, Artificial Potential Fields (AP
d occupancy map data obtained from a prior exploration phase to create a virtual map, describing the uncertainty of each map
mong previous studies. In addition, the performance of Sea-U-Foil in foilborne mode outperforms those of most current unman
best view to capture based on the collected data. We evaluate our method using simulated surveys in a high-fidelity side scan s
that are present in the complex underwater environment, a trajectory-tracking control strategy based on a disturbance observe
can be optimized by online Simultaneous Localization and Mapping (SLAM) and result in improved scans, based on the curren
ntrinsic to any dynamic docking scenario. We benchmark the performance of our approach using the Stonefish simulated envir
determine the water clearance on the sea surface in world units. We validate the proposed approach on real data collected fro
der to calculate the desired marker color. No communication between AUV and DS is needed to determine marker color. Exper
ble regions using model predictive control-based local path planning and control algorithms. The performance and practical util
nd its benefits for robust odometry estimation in an extended envelope of conditions. Lastly, we show the transition between m
with respect to the magnetic north. The proposed framework operates in all environments with slow-varying magnetic fields, ma
extend the sensor data of the real-world AQUALOC dataset to obtain the LBL-AQUALOC dataset. Experimental results on the

tially. Our approach significantly reduces the problem's complexity: in a serial n-link EAMM with m joint-lock mechanisms, wher
e planar surfaces and even small beam-width structures. The paper concludes with discussions on design simulations and drill
n to present this novel control strategy that harnesses the benefits of dual quaternions for cargo UAVs. Our work also holds prom
o unwanted torque generation. To address this problem, we introduce a new model-free center-of-mass position estimation me
all texture detection during aerial interactions. Our experiments demonstrate that our system can effectively replace or complem
ty, we eliminate the design of traditional robotic legs. Instead, we realize steady standing through a tripod-like structure that con
s then mounted on the tiltrotor. A switching controller including exclusive control modes for transitions between free-flight and p
e first aerial transportation and manipulation simulator incorporating different payloads and passive connection mechanisms wi
n simulation and experimental data demonstrate that the simulation is accurate and can be used to continue the development o
consumption during descending. Existing research mainly focused on how to avoid entering the vortex ring instead of offering s
e formally study the sensitivity of the attitude error to such parametric variations, and we provide a method to act on the load po
propellers allows the tether to be wrapped around branches for perching or disentangling. We experimentally determined the m
atics. Intuitively, it is found that if the further motion of the UAV platform is well predicted, the manipulator can directly countera
d, therefore, will be useful to distinguish them from external perturbations, such as wind or physical contact with the environmen
ht envelope figure suggests a novel, natural and intuitive remote piloting interface that we validate in flight tests. Furthermore, w
t focuses on multi-stage ducted centimeter-scale actuators suitable for use on small robots envisioned for deployment in huma
ounds and costs for the sinistral and dextral turns due to hardware failures. Using optimal control theory, the main result of this
 decreasing future UAS operating costs.
ng surface are alleviated. To further adapt to the residual errors, we design a compliant gripper using self-sealing cups. Multiple
on that allows for a pitching-up moment regardless of the wing side facing up. The feed-forward terms computed are fused into
ncertainties. Even small attitude errors can cause a loss of contact points between the end-effector tip and the work surface. To
ane, with no need for complex optimization. We validate our approach through three simulation experiments conducted in the G
ers in Euclidean spaces, to the manifold system, as is, without any modification. The accuracy and the efficiency of our particle
or modeling errors in thrust and moments. We compare our E2E approach against a state-of-the-art network that commands th
ruct a panoramic image. Besides, the results of the power measurements show a significant reduction in the cost of transport b
ody winged drone with trajectory optimisation to solve the motion intelligence problem under diverse flight mission requirements
is introduced into the control framework by virtue of the quadrotor’s bidirectional thrust capabilities. The ADRC scheme is p
s. This problem is relevant because the selection of the stable CL equilibrium AEQ significantly impacts the performance of a U

e pushing on these surfaces while managing platform saturation, this work establishes a safety assessment process. This proce
ve forces to maintain the grasp. The UAV utilizes four main propellers to counteract gravity and four auxiliary ones for enhanced
tory planning approach integrates coupled dynamics model with a single rigid body to generate the energy-efficient trajectory fo
ds lifted cooperatively by multiple drones, and the corresponding coordination. Algorithmically, we provide a solution encoding
he same time allows task-specific assembly, increasing the controllable DOF of the system. Each module is interchangeable wi
form is found to be more energy-efficient than an omnidirectional platform and requires fewer actuators. The effectiveness of o
n a graph using a discrete representation of the state and action spaces. The method is evaluated through simulations and real
s elastic actuation enabling sensorless torque control. The design of the magnetic gear was optimized using a differential evolu
2.5 mW/cm2 at 500 V. We demonstrate a theoretical model of an electrostatic adhesive capstan clutch and demonstrate how l
c performance is basically close to that of our optimized non-foldable wings. Moreover, the vehicle with foldable wings has been
n external magnetic field. An AC-powered Helmholtz coil generated the bristle-bot’s driving field, providing 2D input control,
This approach enables to modify a minimal set of design features, namely the spring diameter and the coil diameter, to achieve
ansferability for future deployment in distributed growing systems capable of acting in various scenarios.
dynamic performance of the PULSE115-60 actuator are discussed. The PULSE115-60 actuator has a highly dynamic response
on motor driving torque and energy consumption is analyzed. Finally, the results of experiment show that, in the best case, the
wide range of curvatures.
mechanism and model constituting the gripper is described. Afterward, a quantitative evaluation of the performance of this grip
ing assembly experiments of a small robot with a life-size arm and experiments with environmental contacts by the small robot
nd synchronism mechanisms. Besides, position accuracy, trajectory tracking ability, stiffness, and load capacity experiments we
modes of the robot were explored in detail. The kinematics and control of the different operational modes were analyzed and ex
s are utilized in the transmission to achieve high backdrivability and high joint stiffness. Actuation torques are intentionally coup
force. The main part of the article introduces a new design of multi sailboats towing system which can tow floating objects to sa
% for detecting obstacles even when the acrylic cover rotates. Finally, a contamination protection experiment was conducted fo
hardware-software remotization architecture.
reedom and features pivots, making it easy to actuate with motors. The system design of the prototype is also presented, inclu
m can be controlled in the same manner as the Lissajous curve. Finally, the proposed method is experimentally validated. The L
n and eversion of the film rolls. Such a bi-cavity design enables the BIBOT to navigate and grasp without relative movements be
model predictive controller to perform WAIR to showcase the unique capabilities of our robot. We implemented the model and c
utrients and oxygen, leading to improved fusion of myoblasts and the directional arrangement of myotubes, thus improving the
y control of the two exoskeletons transmits online feedback between the two users, but at the same time assists in movement a
ntensity-level iterative Hough-transform with Iterative End-Ponit Fitting (IEPF). The effectiveness of the proposed control and tra
velocity of the target object is commonly unknown and difficult to analyze. To deal with it, this paper proposes a data-driven-ba
ively established for further closed-loop control scheme, achieving adaptive positioning and safe surgery. Experiments first dem
ry time. Moreover, misalignment of the robot's remote center of motion (RCM) and trocar position during the procedure can lea
but the average deviation between the estimated and ground truth gait percentages were 1.6%, 1.8%, and 2.1%, respectively, f
necessary, stop the robot or take control to avoid trauma to the patient. In this paper, we propose a novel visualization concep
d at the corners of its body. This innovative design maximizes the contact between the tracks and the colon wall, enhancing ma
dVRK has been carried out across a vast range of research topics. During this extensive and distributed process, common tec
a quadratic programming problem and incorporates physical constraints derived from viscoelastic parameter estimations. Safe
its potential for advanced TEM applications with improved operational dexterity and reduced view obstruction.
d reconstruction of abdominal images can be affected by the patient's breathing motion. This study introduces a method to com
ith on-the-retinal objects based on the intraocular shadowing principle. We validate this pose-estimation method in a Unity simu
to unify the input point cloud and extract the key points, respectively. Afterward, the minimal spanning tree is employed to gene
e in the yaw direction while being fixed in that position. The yaw angle is estimated at a point parallel to the left ventricular longi
ential of the Doppler signal and recurrent neural network modules to preserve sequential information. Input to the first encoder i
registration errors for human volunteers and phantoms were consistently less than 1 mm. In robot osteotomy experiments bas
based 6-A fiber optic F/M sensor and its in-situ calibration method for orthopedic surgical robot force sensing are proposed. Fur
n significant stress on the delicate tissue of the fundus, potentially causing irreversible damage to the eye. To address these iss
y adaptively adjusting the milling cutter's velocity and position. 42 ex vivo animal skull-milling experiments conducted by the aut
same applies for robotic microsurgery. The only space available to manipulate controllers is a narrow space between the patien
s on the surgeon. The paper describes the intricate system development work required to enable these virtual fixtures through h
efficiency. In addition, we propose a dynamic pedestrian avoidance method, namely Socially Aware APF (SA-APF), which can
and virtual fixtures are used to provide force feedback to surgeons to avoid possible collisions when they operate the surgical
ges both force and vision inputs for laparoscopic FOV adjustments. This approach integrates a virtual wrench, generated from v
t polynomial curve. Then, the position of the distal wrist tip can be calculated based on the curve shape. To verify the accuracy
ntrol strategy was proposed. Based on this strategy, on the one hand, the interventionist perceives the multi-modal information f
merous joints consisting of pin-and-hole mating mechanisms. This complexity can increase the overall size of the robot and com
nalysis and prototype fabrication. We achieved the amplification ratio of ×8.44, resulting in the axial displacement up to 86 µ
MRM). Actuated by shape memory alloys, the magnetic moment direction of MRMs can be selectively reprogrammed in real-tim
ions against the constant peristalsis of human intestine. The proposed robot is designed to finish the drug treatment concentrat
e. Experimental results underscore that the proposed control system enhances motion stability and precision. Phantom testing
ngths using a deep neural network. In addition, we propose the integration of uncertainty estimation to address the critical issue
alities into independent box constraints. Further investigations are made for sampling, control, and workspace estimation. Appr
nel for navigating the robot. The robot was firstly examined in a large intestine simulator modelled based on the colon-rectal mo
ontrol of a magnetic swimmer using a 2D ultrasonography device attached to a robotic arm to sense the swimmer's position. W
P) and modified YOLO-V8 vision model is proposed. This proposed system can analyze the current surgical workflow and gener
upport Vector Machine (SVM), Decision Tree, Naive Bayes and Random Forest. Cross-validation on simulated test data showe
of-the-art machine learning architecture (Transformers) to detect and segment catheters in axial interventional Ultrasound imag
rmance is better when working near the center workspace and away from the servos and the average error of the robot is 1.39m
employs a grid layout with distributed sensing nodes throughout the environment. This system effectively tracks undesired mov
or avoiding tissue damage cannot be met in this control mode. Also, other capabilities, such as hand motion scaling and haptic
uring offline training. This method can take user-specific characteristics into consideration and help them achieve better teleope
in the flexible material. In vitro experiments validated the passive navigation ability of the FCE in rugged intestinal tracts. Furthe
ed and demonstrated in experiment. The bendability was tested showing the camera could reach 20 degrees in bending angle
apabilities, producing an optimal power output that can concurrently illuminate up to 15 LEDs at their maximum intensity. Crucia
urgical risk. This method protects the surgeon's behavior and controls the catheter force at the same time, which is named exce
ngle-core FBG fiber. A thin-walled soft sensing tube helically embedded with FBG sensors is designed for a robotic-assisted fle
ter introduces an ontology-based research framework for creating and implementing a complex autonomous surgical workflow,
nts on custom-designed silicon phantoms and defining a set of success/failure metrics, we have put forward findings that estab
y eleven participants, who could successfully control the microsurgery tools to perform basic surgical movements. Furthermore,
At the same time, the accessible-visible area is better and faster matched to the working volume of the robot. This paper presen
ewire navigation times, while successfully reaching all predetermined luminal targets. The clinical relevance, usability, and oper

high dexterity after mounting on a rigid robot arm. Experiments demonstrates that the proposed the slit design doubles the stiff
e-guided methods successfully completed navigation tasks in a static environment, outperforming human navigation performan
reach desired aneurysm targets and then perform the delivery of embolic materials and tools. To enhance the control and prec

edle with full three-degree-of-freedom (DOF) planar needle motions. Cross-validation studies using custom-designed multi-laye
from real robots. The excitation trajectory was evaluated and achieved on par criteria when compared to state-of-the-art report
ble material selection, arbitrary geometry design, and programable magnetization profile. We can make full use of this system to
s mechanism was used to launch a microscale projectile (a salt grain) thereby demonstrating the relative high power actuation a
essentially Pickering bubbles composed of magnetic particles and air bubbles. Their hollow structures help produce lighter mic
This work advances magnetic field-controlled microrobotics, bridging the millimeter-to-micrometer gap. It holds promise for app
electroosmotic flows produced by the OCP in the channels arranged between the electrodes. The MS works by aligning the m
logy and fabrication, followed by in vitro characterization and testing within brain tissue phantoms.
y driven within a specific narrow frequency range. This type of micromotor shows bi-direction motion capability and has the pote
ed platform is more precise, accurate, and efficient than teleoperated and manual microsurgery conducted by experienced scie
dered. For human demonstrations, we collected data from repeated tests performed by ten operators following three motion stra
ocol is also presented in order to support tying knots with shorter and thinner suture. Leveraging the natural advantage of nano
obotic cell micromanipulation. A weakly-supervised depth completion network is proposed to take cell images and sparse depth
ation shadow compensation across an expanded field of view. At the same time, the system can also dynamically adjust comp
nally, the end-effector is controlled to obtain intact single cells from tissues by model predictive control (MPC). The performance
nsions of 16 mm × 16 mm × 22 mm, and workspace of 0.7 mm3. The end platform of the micromanipulator can be equipp
he insect-scale robot with high mobility, that is, applying Grassmann-Cayley Algebra to avoid the singularity of the transmission
miniature robots with a maximum characteristic length of 10^-2 m can exert a force and a torque up to the order of 10^-1 N and
r fluorescence techniques to enhance visual contrast between sperm and culture medium, leading to clinical impracticality. To a
d to execute a SIFT matching-based visual servoing control at progressively increasing magnifications, using only a single high
ct to the cell membrane, whereas the QCR force sensor was incorporated to detect the force interaction between the sample an
m based on feature matching. The overlap between two adjacent patches is 50 % and 33 % in the X and Y axes respectively. T
ensional positioning stage for sperm immobilization that can be seamlessly integrated into standard clinical platforms. Based on
sperm morphology parameters automatically and quantitatively. A novel attention-based instance-aware part segmentation net
 -controlled microtrajectory optimization. The proposed method is adaptive to imaging details of different regions of interest (ROI
erate acoustic radiation force and excite microstreaming. This method can accomplish rotational and translational operations pr
dvancement. Our methodology is demonstrated by teleoperating a 12-mm-long screw-shaped UMR (5 mm in diameter) within a
ances from the Sim-Suction-Dataset, leveraging the synergy of zero-shot text-to-segmentation. Real-world experiments for pick
ed via repeated observation. Our model results in a three-stage planning process: first, the robot plans using only observed obje
ly, we propose a real-time spatial-temporal trajectory optimization method that takes into account the whole-body safety, agility
-level action space around a semantic map memory and leverage the explored environment as instance navigation points. We
while also substantially reducing the compute time required to update the MPC policy. Simulations and hardware experiments
formulation. In this work, we introduce a hierarchical approach that uses Gaussian Mixture Models (GMMs) and Gaussian Mixt
oses enables a direct assessment of graspability, encompassing both the quantity and quality of grasping poses. This assessme
impedance controller to realize the end-effector's force control, and the introduction of impedance control has improved the saf
omputing path-wise utilities. These utilities trade-off maximizing the visual Information Gain (IG) for scene reconstruction and th
al-time even in operating regimes that greatly differ from the training distribution. Moreover, we design an uncertainty-aware mo
gn a novel action sampling algorithm that reasons about geometrical differences between point clouds to further improve the ef
andard Model Predictive Control, MPC) is formulated, and a Neural Network (NN) that leverages the use of Feedforward and R
g training. This enables the design of a safe and robust model predictive controller through appropriate cost function design whi
n enhancing the performance of principally kinematic locomotors (those governed by Rayleigh dissipation or zero momentum c
e-variation of parameters by introducing an optional forgetting mechanism. The characteristics of the recursive regularised least
mation, direct input-output dynamics are shown to exist by training neural network models on simulated trajectories. Output dyna
ed to capture the dynamics of the system. To guarantee the cohesiveness between the multiple predicted states of the system
otion model generalizes to previously unseen terrain conditions. The output distribution from the proposed motion model can be
ss criteria and is applicable across various tasks and robot models. We evaluate its performance in the context of motion retarg
study and assessed subjects’ blood pressure and heart rate levels while applying MiBOT. We found that our mechanical s
y. Furthermore, a new metric named clamped temporal error (CTE) is proposed to give a more comprehensive evaluation of pr
ynthAct datasets: AMARV, a large synthetic collection with over 800k multi-view action clips, and Synthetic Smarthome, mirrori
cteristics were then utilized in the robot skill modelling to achieve a more human-like grasp. Since no force sensors are fixed to
cenario, we demonstrate how our framework can help ergonomists improve workstation designs thanks to the resulting suitabili
esentation and hierarchical deep reinforcement learning to achieve state-of-the-art full-body control. We validate the effectivene
esigned for 2D encoding, where some operations of multi-scale fusion and frequency domain filtering are performed to extract m
were equipped with EMG sensors and skin markers that were streamed in real time and used as targets for the MHE. The upp
continuum robots: shape locking and adjustable stiffness. To validate the modeling approach and theoretical results, a series o
on experimental results compare well with simulation exhibiting an error less than 4%. Only one region of operation demonstra
rforms its conventional counterparts regarding tracking accuracy and response speed. Our work reveals a new direction for adv
l robot attest to the superior modeling accuracy of our approach in 3D motions with unknown load. The maximum position/rotat
t solves these nonconvex feasibility problems with a sequence of semidefinite programs whose objectives are designed to enco
C’s capabilities to handle multi-task tuning and we add a barrier cost function to enforce actuator constraints. Our experime
econstructs the shape using piecewise spline interpolation. The approach only relies on knowing the number of the key points a
n dynamics and external disturbances, it thus can guarantee desirable tracking accuracy and control adaptability. The learned
g., input-to-state stability, and we prove the convergence of the controller coping with locally Lipschitz nonlinearities. Finally, we
ency. A combination neural network of GRU-MLP (Gated Recurrent Unit-Multilayer Perceptron) was proposed to identify and c
orces and torques at its base. For the first time, this is done using an unscented Kalman filter without restraining the dynamics
 sing four objects of different shapes and weights in 140-mm squared target boxes. We registered a success rate of almost 65%
ng a way toward Mixed Reality Soft Robots. This full Skinned VR-Soft IK pipeline uses skeletal animation and GPU picking. We
ed model of the robot that enables frugal learning of a motion strategy. This strategy is then transferred to the complete model
y-state error. Feedback is visually provided from a calibrated stereo vision system. Target-positioning and trajectory-tracking ex
sent POE-M. POE-M predicts key point positions from acoustic signal observations and uses an energy-minimization method t
-pass actuator dynamics, including swelling-driven actuators used in hydrogel crawlers. These models accurately capture the d
v methods without relying on the elastic dominance assumption. This controller is applicable to the general class of underactuat
us between biomechanical and robotics-focused research, providing critical insights for the advancement of robotic hand actua
y weight and weight increase for the center of mass (CoM) acceleration and stability as confirmed by several experiments. Furt
otype has demonstrated vertical climbing on both flat cinder block walls and uneven rock surfaces in full Earth gravity.
so demonstrated the ability to traverse obstacles with a payload in the inverted climbing tasks.
to do less high bandwidth perturbation correction by planning arm motions that induce zero perturbation on the base. We are a
ems. We test our method in two classes of high-dimensional systems - low Reynolds number swimmers and free-falling Cassie
egged robots. The framework is handled by a Behavior Tree (BT) and it takes into account recovery methods to deal with possib
etwork controller to satisfy the requirements specified in STL. Our controller learns to roll out trajectories to maximize the STL r
f sub-problem solutions in the absence of an optimal substructure, and without uniform time discretisation. We present a framew
ironments with multiple narrow passages require a densely sampled roadmap to capture the connectivity of the environment. S
stimated using a modified Stein Variational Gradient Descent (SVGD) method and embedded into the MPPI algorithm to find a
o the true visibility polygon as the step size goes to zero. This dynamic-programming approach allows the evaluation of visibilit
with numerical examples. The clothoid approximation ensures an accurate and smooth representation of the curve, and the cloth
poses. The effectiveness of our methods is thoroughly evaluated and confirmed using both simulation and real-world experime
rade off battery usage and coverage quality, maintaining uninterrupted exploration by at least one agent. Our approach derives
he algorithm acts as RRT-Connect to rapidly find a feasible solution, which helps compress the sampling space as Informed RR
he extremal trajectory of an optimal control problem. We show that the set of stable robot configurations is a smooth manifold d

form Neural Focus, which constrains the point cloud within the admissible ellipsoidal subset from Informed RRT*, and feeds int
ver a diverse set of challenging, realistic problems for complex robots ranging from 7 to 14 degrees-of-freedom. Moreover, we s
mly-generated environments.
el trajectory optimization framework called RETRO. RETRO employs adaptive optimization techniques that span both spatial an

rch space is far more non-convex than was previously considered in the literature. The wide range of applications for boustroph
strated to be effective in a number of time-critical tasks. However, robotic systems often have more time allotted for planning th
ted from a novel context-aware planner. Our planner adapts the robot's velocity space based on the presence of entrapment in
mble-of-costs-guided Diffusion for Motion Planning that aims to combine the strengths of classical and deep-learning-based mo
ariety of robotic systems, and show that it consistently covers larger portions of free configuration space, with fewer polytopes, a
king information (evaluating where an evader may be). The algorithm is inspired by well-known methods in the motion planning
st, a bidirectional vertices reduction algorithm is proposed to tackle the asymmetry of the path and the environments. During th
riangulation when part is in the free space. We perform experiments on 4-DOF, 5-DOF, and 6-DOF serial manipulators. Infeasi

hecking scheme, which leverages robot-specific traversability learned in simulation. We use LazyPRM* with an informed sample
mulation scenarios, varying in people density, number, and movement.Simulations were executed with an Nvidia Jetson Nano
al efficiency which only takes a few seconds to compute the full state and control trajectories of the drone through tracks with do
nts, (ii) metrics that characterize the difficulty of motion planning in these environments, and (iii) an open-source software stack
to overcome unknown external perturbations and inaccurate model parameters. Extensive experiments show that our method
ose a two-stage planning approach, including transition point design and hitting primitive generation, with a simplified expressio
ves in an arbitrary way. The ultimate goal of the proposed framework is to release the suspended load to the ground target and
multi-vehicle planning approach is to minimize the collective value of the dynamic model reward states. The purpose is to addre
stimate human commands and state feedback reliably. A control strategy for the quadcopter is designed to interact naturallyÂw
UAV, the framework enables the provision of integrating the overall autonomy stack, potentially opening up an entirely new fro
, an admittance control-based position compensator is proposed. Then, after the backlash elimination, a dual-motor linear activ

lure. To avoid the control allocation switch from the fault-free control to the FTC, a dynamic control allocation is used. In additio
ascaded proportional-derivative (PD) controller, this slosh-free reference is transformed into task space acceleration commands
e JetRacer autonomous vehicle with additional modifications to support four-wheel steering, model predictive control-based pa
hrough considering the worst-case distribution within the ambiguity set. To facilitate solving the OL-DR-NMPC problem, we refor
ree largely employed control approaches. The proposed prediction method is then validated using simulated and real robot exp
tracking errors over repetitive operations. Additionally, hardware experiments are conducted to validate the effectiveness and r
nce, a high-pass filter (HPF) is integrated with the F/T sensor and its impact is investigated. Experimental validation is conducte
mics to generate optimal ground reaction forces (GRFs) to track the safety-ensured trajectories from the top-level planner. The
s based on safety levels, and 5) A low-level controller. Our comprehensive framework, comprising these components, enables a
he acceleration level. Additionally, a parameter calculation method is proposed to compute unknown parameters of the end-effe
ormance to established BO-based tuning methods. We further evaluate RaGoose performance on a real precision-motion syste
periods, we classify the movement into short-term movements and long-term movements and compress their patterns in the fr
gence of discontinuous switching terms. We analyze the boundedness of control signals and the stability of the closed-loop sys
high inertial forces involved. Unfortunately, hard nonlinear behavior is associated with piezo- electric actuators, which adds to th
ables a robot to react to unknown interactions while accurately tracking the desired trajectory in free motion. The effectiveness
arios, precise tracking of the human's operation is achieved by combining MP and MV controllers, while a combination of MV a
s in the motion subspace. The experimental validation with a torque-controlled robotic manipulator on both flat and curved surf
an online learning policy allows to properly estimate the requirements of the task and adapt the controller accordingly, thus simu
ybrid control in task space, ensuring accurate control of end-effector force while achieving precise position control in other direc
g velocity, we must retrain a new NN policy. In this work, we attempt to bridge the gap between these two bodies of work on a b
is proposed to solve the trajectory tracking problem. We linearize the bilinear model around the estimation of the lifted state an
rface-deviation model, based on the assumption that the robot action can specify the cutting surface made by the tool. Furtherm
d model to minimize the gap between the prior model and reality, and a real-time multi-contact MPC controller. We demonstrate
ng data over three distinct UGVs and four terrain types. We show that our protocol offers increased predictive performance ove
orthogonal foliations because they are, in a certain sense, orthogonal to the task self-motion manifolds. The approach provides
nt scenarios with a Fetch robot, involving planar table puzzles and an escape room scenario. LA-RRT significantly outperforms
ity and manipulator postural control domains, contrasting the performance of the opportunistic approach with state-of-the-art RR
ces with up to 12 cubes in an average time of about 200 seconds.
ensures that the remaining segments during following strictly remain within the collision-free regions defined by the base and p
sent a road demonstration result for urban city driving.
-effector trajectories while executing the minimum number of reconfigurations. Our graph-based method generates a diverse se
 ning. Inspired by these approaches, we propose the first physics-informed CMP framework that solves the Eikonal equation on
lt. We demonstrate the approach with extensive simulations and experiments using a Crazyflie 2.1.
ollision-free, but is actually in collision with unperceived parts of the environment. Our experimental results show that successfu
ss these problems, we propose a framework called Eclares, in which our contribution is two-fold. 1) First, we propose a method
es also present computational challenges. Our contributions are threefold: First, we solve previously unsolved problems; secon
e SEF constraint. As a result, even with an SEF path, the robot still encounters self-entanglement issues during execution. This
ural Networks and has two variants handling different inputs, respectively: state (signed distance) input and point-cloud input fro
imal base trajectory subject to geometric task constraints, end-effector trajectory continuity constraint, collision avoidance, and
and real-world environments. Unlike previous learning methods, which require task-related data, our method uses pre-trained n
rsing challenging terrain that requires discovering a contact schedule, navigating non-convex obstacles, and coordinating many
nt cloud in real time. Based on the risk mapping result, a risk inspired active exploration method is proposed to actively search

e achieved improves from 2.4 m to 3.2 m. We also quantify the system’s ability to adapt to new environments without exter
eactance without diminishing compliance to the verbal requests issued by the robot. This phenomenon has been verified before
complex setup. Specifically, we target robot-to-human box transfer while the human partner is on a ladder, and ensure that obje
ng a multimodal interaction framework where language and hand-drawn inputs can serve as navigation constraints and control
, employing deep learning for accurate tactile perception. Simultaneously, the actuator array generates mechanical vibrations to

To address these challenges, we propose CAMInterHand, a cooperative attention-based method for multi-view interactive hand
ove the performances further by using both 2D skeleton heatmaps and RGB videos as inputs. To our best knowledge, this is th
ose, and gaze direction, are used to train a supervised machine learning algorithm to classify whether a touch is intentional or n
ld scenarios. In the experiments, we validate TiO on GuessWhat?! and InViG benchmarks, setting new state-of-the-art perform
e Models (LLMs) to make sequential decisions to manipulate different modules for perception, navigation and communication. E
we propose a new robotic system that can interpret the user's intention and pick up the target object, Pragmatic Object Graspin
ry voltage into a noise-reduced 2D image. The image reconstruction network uses an attention mechanism to focus on salient
ion of the given commands could lead to a decrease in malfunction and undesired actions of the robot, enhancing the reliability
w politeness during group discussion, e.g., gazing to the speaking member, looking to the middle when both members are talkin
. The evaluation of the novel interface was performed by 17 participants who completed the setup and operational activities wh
nalyze the performances of these observation models and propose an optimal observation model combination scheme. Moreov
ships in the IAP, we propose corresponding methods that tightly fuse visual-inertial-range odometry and kinematic constraints t
ctiveness of our method on three different datasets, indicating a significant improvement in robustness compared to previous wo
t with three robots communicating through an ad-hoc network. Our code is publicly available: https://github.com/MISTLab/Swar
ility. A method for Visual Geolocation and Matching Network (VGMN) is suggested by constructing a graph convolutional neura
eï¬ ning the joint pose graph optimization through place recognition, global optimization, and the removal of redundant data to
S-denied environments without range restrictions. The proposed approach in this letter is also tested in simulations and experim
ariance with the acquisition time and point range, allows us to alleviate the strict time synchronization and to overcome FOV diff
iii) effectively associates and optimizes large-scale and numerous map segments in the real-world scenario. AutoMerge utilizes
els and the observers into a distributed Model Predictive Control (MPC) framework with kinematics limitations and collision avoi
ion transformer network (GATN) that represents the problem of collaborative scheduling as a bipartite graph, and integrates bo
ed estimation uncertainty for the target states, while minimizing the average number of active robots. The robots have a limited-
 separates into two phases: nearly optimal and minimal performance. We speculate that in real-world applications that are conf
nce rates and accuracy compare to existing state-of-the art CSLAM back-end optimizers.
g poor performance diversification through determinants does not replace the better agents in the archive. Rewards and divers
d punishment. To aid robots in distinguishing between perception-friendly areas and unfavoured zones, we instead provide sem
ompanied by a text-only dataset that evaluates LLMs’ agent representation and reasoning capability. We experimentally de
sion episodes to improve performance in obstacle-rich environments. We implement an attention mechanism to attend to the n
nimals' instincts, and propose a robust multi-robotic collaboration and communication framework, C3F. It introduces graph-bas
Rollout (MLAT-R) algorithm. MLAT-R provides 1) provable improvement over a base policy, 2) policy improvement with respect
nsive experiments are conducted, demonstrating that robots trained in competitive environments outperform ones that are train
e propose a reinforcement learning method that learns visual coordination behavior using OWC. Our proposed visually coopera
ion-exploitation dilemma that arises when robots must balance the mission goal with seeking information on undetected neighb
ssigned region. As the team serves events, we assign each robot a reputation, which is then used to adjust the size of a robot's
bitrary patterns that our solution convergently achieves global accuracy as accurate as a centralised non-linear factor graph so
pansions. The spatial expansion collects delayed states from distant neighbors, while the temporal expansion incorporates prev
n (SDR) to furnish a lossless relaxation. By solving the relaxed problem, time synchronization and relative pose estimation can
We demonstrate the performance of MOCHA through real-world experiments with commercial-off-the-shelf (COTS) communica
and employ graph matching to assign UAV clusters to UGV clusters for efficient recharging management. We perform numeric
and other space-filling machines, we manufactured 48 rhombic dodecahedral cells and used them to build various superstructu

o statically model an individual robot's capabilities within our ros_bt_py BT framework. It offers a runtime auction system to dyna
ptimization problems in a centralized and distributed fashion. Finally, we demonstrate the effectiveness of coordination among
e coordination process. A novel task allocation protocol is embedded within the framework to guide this process. We substantia
ere vertices are full composite spaces and edges are transitions between these. Using the hypergraph reduces the representa
he formulation of factor graphs, along with an algorithm for the efficient, real-time solution of such factor graphs, iSAM2. This es
e also supply complex dynamic utility models that are applicable to many missions and new algorithms for intelligently scheduli
sharing among robots to reconstruct their collective maps. The evaluation of RecNet encompasses an array of metrics, includin
time, the navigation behaviour of the robots, just by adapting the semantic labels; and (iii) the robots can reason about their nee
ore, figure out the size and shape of 3D structures and then plan surface-inspection path. To overcome these challenges, we p
SOGM. The simulation results show that our method achieves competitive performance against state-of-the-art methods in dyn
propose this intent representation in the context of the Dec-MCTS online planning algorithm for decentralised coordination. We
. We validate our findings through simulations and hardware implementation.
nstead, we focus on the setting of decentralized multi-agent, multi-target, simultaneous active search-and-tracking with asynch
between curbing map uncertainty and achieving efficient task allocation among robots.
ures. Then we use an off-the-shelf solver to show the type of optimal solutions that our planner can produce and assess its per
entives for a bipartite graph matching approach to MRTA. Specifically a Capsule Attention policy model is used to learn how to w
operations. From mineral exploration to ore shipment, there are roughly nine steps in between. These include: geological asses
trategy we propose, which is executed once the routes are computed and has the aim to prevent congestions, collisions and de
n efficiently converge to a Nash-stable coalition partition even in conditions of insufficient communication distance.
st while maintaining a collective observability of the target(s). In order to enforce satisfaction of the observability constraint, we
eometric conditions for the degeneration of bearing rigidity (singularities) and the resulting uncontrollable motions. A classificat
mulate the multi-robot patrolling problem in discrete environments as an 'unnormalized' joint steady state distribution entropy ma
cal information. Practical applications include defensive structures or escorting missions with the added resiliency of a swarm w
y a noise-resistant heuristic optimization scheme based on the Particle Swarm Optimization (PSO) using a fitness evaluation th
nt order and inter-robot collisions. Larger structures are either intractable or yield high-cost solutions. A prior approach mitigates
ce. Furthermore, a fast optimization algorithm is presented to find finger configurations for force-closure precision grasp based o
nd highly modular. This paper presents the details of the design, the actuation principles, and the experimental validation of the
as crafted from TPU and conductive TPU, and their mechanical and electrical properties enable easy implementation of tactile
erization, where the prior constraint provided by a hand poser network, is introduced to ensure that hand pose is natural and rea
nts a novel solution in the form of a wearable dexterous hand, namely Hand-over-hand Imitation learning wearable Robotic Han
riments in which a humanoid equipped with WARABI Hand grasped forearm, held hands, and interlocked fingers with a person.
otion and speed. We demonstrate that a soft skin enables firmer grasps and piezoresistive sensor integration enhances the ha
BC deteriorates; the performance of offline RL algorithms is also less than satisfactory. Upon examining the provided datasets

onstraints with the assistance of both control Lyapunov functions (CLFs) and control barrier functions (CBFs), the multiple UAV
g connectivity for a group of quadrotors with onboard sensors under distance and field of view constraints. Leveraging distribute
he leader must be limited so that the curvature is first (second) order differentiable. We further show that these constraints are
allel, on-demand or asynchronous computation for efficiency when different aspects are more or less important. This is the first
d analysis on the impact of the 5G QoS functionality on the performance of 5G-enabled UAVs. Furthermore, we introduce a nov
arning and cloud and edge-computing in the context of robotic manipulation. In this work, we describe the mechanical design of
tier for cost and latency by running the application task on all available server configurations, we evaluate cost and latency mod
and operate on robots equipped with off-the-shelf WiFi radios. We demonstrate this through a set of experiments showing our M
with a position drift compensation algorithm, for which convergence guarantees are provided. The experimental validation of th
ilitate this process through a learned attention mechanism, bridging the query image embedding with the embedding of parame
e aggregated feature also benefits a more accurate proposal refinement. To demonstrate the performance, comparison of VeloV
and narrowed the difference of detection performance across object categories. Meanwhile, our MTRadSSD outperforms thos
spaces. Furthermore, a depth calibration (DC) scheme introduces depth reconstruction to further adjust the 3D object localizati
for object detection that can more effectively adapt to real-world environments with adverse weather conditions by addressing t
e ambiguity issues for symmetric objects, we propose a novel training procedure for symmetry-aware keypoint detection includ
ta are fused to improve the prediction accuracy. Normalized mutual information (NMI) is leveraged as a modulator for calibratin
cover the metric scale of the object based on category-level statistics. Finally, we advocate using the RANSAC-PnP algorithm t
equently refines the features based on pose and size constraints. Experimental results on REAL275 and CAMERA25 demonstr
to images with many unknowns and complex components, such as driving scenes, these methods often exhibit unstable perfor
our PVTransformer achieves much better performance compared to the latest 3D object detectors. On the widely used Waymo
atures used for distillation. We evaluate our novel distillation approach on a synthetic and a real event dataset with aligned gray
A) that builds multi-scale tokens from a single-scale input feature to enable more fine-grained feature learning. Second, we prop
e three motion models in our work and show how the unicycle model and the bicycle model improve turning cases. On Waymo
that there exist a room for improvement for the Pillar Feature Encoding (PFE) stage. We demonstrate that with sufficient repres
ow-resolution image. To facilitate this purpose, we propose an efficient axial subband approximation module for extracting axial
nly a negligible overhead. Concretely, the Cross-Cluster Shifting operation enhances the conventional design by shifting partial
dly exploit geometric constraints, to boost the segmentation quality of challenging edges. Experimental results on the Cityscape
better, even though only single-modal data is used as the input during the inference stage. Extensive experiments on the View
model to predict the class distribution of unlabelled data. By aligning the class distribution of pseudo-labels with this prediction,
tions between perspective pixels and BEV grids. It utilizes the attention mechanism to compute similarity scores for each assoc
ing the current objects' poses given past observations and (ii) another branch that detects objects based on the current and pas
nd neck to increase channel information of the feature map. Moreover, an attention mechanism named CBAM is applied at the
P's strong vision-language representations to segment regions from referring expressions, while utilizing its ``U-shaped'' encod
not be sufficiently equipped to handle extreme blur and minuscule objects. Instead, we put forth a novel coarse-to-fine detection
high-end vehicles equipped with LiDAR sensors). Under this setup, we proposed a novel, simple, and end-to-end trainable fram
h maximum entropy training. This further improves the metacognitive module’s performance without requiring access to OO
linear and rotational velocities of frames randomly attached to rigid bodies due to robot interactions can be used to identify obje
e planner effectively navigates around obstacles. Unlike conventional methods that emphasize explicit obstacle tracking, our ap
semantic links between point cloud features and open-vocabulary labels. The intensive experiments show that our approach ou
(EDL) based uncertainty estimation framework. Specifically, we propose MEDL-U, an EDL framework based on MTrans, which
timization-based pose estimation achieves higher accuracy, while the filtering-based tracking, which utilizes the differentiable p
te its effectiveness and establish the first benchmark for radar-based dynamic occupancy grid mapping using the publicly availa
ethod of low-resolution data from the perspective of time-series fusion. By modeling time series of sparse data, we indicate chan
uch as multi-storied buildings or staircases, and for diverse moving objects such as people carrying various items, doors swingi
dense RGB-D point clouds, as well as the discrepancies between VI-SLAM point clouds generated with different object and env
on. Finally, in order to address the limited resolution of RV and its insufficiency of 3D topology, a new point refinement scheme
ance. The final association combines geometric and appearance features to overcome the limitations of center-based tracking
gle-sensor or multi-sensor fusion-based BEV object detection methods have not yet taken into account capturing such multi-lev

from dense urban, rural and off-road driving domains. The proposed method consistently outperforms uncertainty estimation a
an end-to-end manner via bipartite matching and complex post-process is not needed at inference. To enhance the supervisio
emporal Information Disturbance, which predicts perception information when disturbance occurs. Specifically, the Historical Fr
recent supervised deep learning approaches, on S3DIS and ScanNet datasets. Moreover, CID can be used to group points in
eep appearance descriptor, surpassing existing MOT methods which even use the deep appearance descriptor on a 2.5 fps da
generate 3D pseudo labels from 2D bounding boxes, into the existing self-training pipeline, our method is able to generate mor
bust. Through extensive experiments, we demonstrate that our method performs on par with state-of-the-art approaches on no
etrics. Notably, it outperforms the other state-of-the-art methods by 9.32% in precisely predicting the path with 18.28% faster in
ures. For instance-level fusion, we propose a Radar Grid Point Refinement module that reduces localization error by accountin
. This paper explains how OVAAL can shift more attention towards the minority classes and thoroughly analyzes its contributio

multaneously. We then tackle the two major challenges that emerge with discwise AL. Firstly we devise a new acquisition functi
SAL, a size-balanced warm start active learning model, based on the observation that each object class has a characteristic size
control. Subsequently, we iteratively solve prediction anchor points based on two motion models to robustly predict the target t
special heading requirements for agricultural applications. The controller is implemented as a Model Predictive Control (MPC)
vexity within the proposed formulation to form conservative inequality constraints. We then obtain the desired trajectory by solv
retical properties when used in conjunction with iterative methods for trajectory optimization such as a more clustered eigenvalu
PC, solving larger problems, at faster rates. In particular, for tracking tasks using the Kuka IIWA manipulator, MPCGPU is able
points, which significantly accelerates trajectory planning. Furthermore, the planned paths are tracked via a time-adaptive mode
contour following and peg-on-hole alignment tasks, invariant descriptors are shown to result in task representations that do not
etiming problem with quadratic objectives as well (e.g. minimize control effort or match a nominal speed by minimizing squared
computational costs. Here, we employ a distributed processing approach that enables not only the minimization of computationa
speed and reliability of the approach on problems with stochasticity in nonlinear dynamics, obstacle fields, interactions, and terr
of CERs is the automatic discovery of clutch sequences. Due to complexity, many methods still rely on pre-defined modes. In
hat can optimize the state trajectory, input torques, and stiffness profile. We provide an efficient method to compute the forward
e tracking is necessary. We show that a simple reformulation of the optimal control problem combined with standard estimation
ss. We demonstrate the usefulness, simplicity and computational efficiency of geometric algebra in several experiments using a
objective, we also provide conditions on when an additional observation of the unknown disturbance behavior needs to be coll
ning the value function of a nominal policy, and show that the value function of the maximum-over-time cost is a CBF. We dem
m of mobile robots in a human-multi-robot teleoperation scenario.
handle hybrid systems with contacts. We demonstrate that our method can generate tight under- and over- approximations of
We achieved 6.8kHz, 1.9kHz, and 1.8kHz update rates of the MPC for 0, 1, and 2 contacts, respectively. This allows the robot
CA). We use Differential Dynamic Programming (DDP) to obtain optimal control commands that can form granular particles into
ntrol policy via TO-guided RL policy search that, when used as initial guess provider for TO, makes the trajectory optimization p
e in Indiana, USA. The results show that the proposed control policy exhibits improved robustness to parameter tuning and data
velop the algorithm, we capture our problem as a sequential game between a controller and an adversary, where the controller
ve, a robust controller is designed through the linear matrix inequality (LMI)-based approach. More importantly, a necessary and
el data-driven robust policy iteration method is proposed to solve the robust optimal control problems. Without relying on the kn
egulator (TVLQR). RTC-D augments RTC with a design optimization layer, maximizing the system’s robustness through a
area coverage (HEDAC) and slow the diffusion using the non-stationary heat equation. Our approach extends HEDAC to applic
d performing a whole-body pick-up motion on a quadruped equipped with a six-degree-of-freedom arm. These benchmarks indi
etwork to anticipate collisions from externally observable metrics. Anticipated collisions are avoided using a simple, but yet effe
es the follower's feedback dynamics. Based on this learned model, the leader devises a collision-free trajectory to guide the fol
ose edges. The convoy is permitted to wait at any vertex to allow the service vehicle to repair an edge. The service vehicle is pe
method maps agents to pre-defined sets of spectral coefficients. In the second method, each agent learns a weight distribution
ral transformer combined with multi-agent reinforcement learning. This approach effectively captures the complex interactions b
h significantly outperforms existing algorithms, fixed heuristics, and prioritized planning with KATS. The performance is maintain
coder that utilizes the decoder to select the optimal compression. Numerical simulations are conducted in a large close-to-real m
and then use the reservation grid to prevent collision. Based on the reservation system, we develop a new formation planning
ch can be used for IPP with both discrete and continuous sensing robots, with point and non-point field-of-view sensing shapes,
unstructured and human-shared environments. To this end, we introduce a multi-robot benchmark framework that is based on s
g computational demands compared to existing approaches.
based on shared position trajectories of the quadrotors. Static obstacle avoidance is coupled with distance queries from the Oc
monstrate that CB-MPC improves the executability and success rate, allows for closer robot-robot interactions, and scales bette
evel uses the resulting solution with discontinuities as an initial guess for a joint space trajectory optimization. The procedure is
combining the use of ground truth proximal (local) information and fuzzy distal (global) information to let agents sequence local
e of the unknown workspace. Experimental results on several standard benchmark workspaces show that our algorithm scales
d’s ability to efficiently plan paths in complex environments with many narrow passages, scaling to robot teams of size up t
or large-scale MCPP problems. We show that both heuristics result in reduced-size MIP models that remain complete for RMM
collision classifier network, as well as Gazebo simulations and real experiments to assess the resulting capabilities of the N-MP
eves the fast detection of the cable while the respective measurements are afterward fitted to a Bézier curve in order to appro
osed of several cycles of flying, landing, recharging, and takeoff, validating the capability of extended, essentially unlimited, ope
 o provide demonstrations for imitation learning. Finally, a series of experiments are conducted to demonstrate the effectiveness
during exploration and selects the next best frontier to navigate to. The collector strategy ensures global exploration by balanci
g diverse slant-angle drone datasets with various weather conditions and backgrounds. The model leverages two distinct featur
rther fused with the detection results based on an extended Kalman Filter (EKF). A perception-aware model predictive control (
ation Per Thread culling strategy (MLPT) and Thread-Efficient Warp-Allocation (TEWA) scheme for GPU to enable Jetson-SLAM
her design a collision recovery control for rapid recovery and collision-aware mapping to integrate collision information into gen
ding the global optimizer for Gamma, we identify the value of psi_d which maximizes the perception of risk within the FOV. We
for applications where high-fidelity modeling is required. To bridge this gap, this letter introduces a spatial hash map for rapid G
y reached to provide the most information. The two modules are tightly coupled and run onboard allowing the vehicle to safely
lete cover age. Based on SSD, a hierarchical planner effectively divides the large planning problem into independent sub-probl
d for target assignment. Using these two components, our pipeline effectively assigns tasks among robots, and enables a meth
stance wireless communication in such scenes can become a limiting factor, especially when automating the navigation of a UA
nd robust sim2real performance. A set of simulation and experimental studies in diverse environments are conducted and demo
uccessfully trains the vehicle to traverse through the center of the gap at an approximate attitude to the gap with aggressive tilte
ch time step, an occupancy map of the environment is updated by the VI-SLAM by fusing the onboard measurements, and a C
(or occluded) by an upstream whisker receives less incident flow. We show that this relationship is predictable for two whiskers
s with limited computing power. Our method adopts a novel ensemble detection strategy, combining multiple computationally e
mber of covisible features while ensuring small enough parallax angles. Additionally, we propose a differentiable and accurate
nt farming robots in scalability and versatility. We showcase compliance control algorithms combined with artificial intelligence w
ngth, and spread of the encountered updrafts. We demonstrated in simulations that our approach can reliably detect updrafts s
hancing the speed and accuracy. Eventually, we develop a neural network with an impressively small model size of only 4.5M.
g plants must be matched across many robots. We demonstrate superior performance compared to several baselines, obtaining
e scanning as input to register each target object with mesh reconstruction and novel object pose tracking. In the second stage,
to curvilinear surfaces. The performance, repeatability and uniformity of the sensors are characterized in detail. We also presen
ect or the arm. In this paper, we build upon the advanced sensing and attaching capabilities of our suction cup MIGHTY, and in
itate fine-grained local feature alignment by dynamically updating and matching the grasp prototypes from the simulation and re
depth perception and training speed for transparent objects. Robots often operate in the same area, such as a kitchen. By first
racking feature points, while simultaneously refining the camera poses via a robust maximum a posteriori (MAP) inference. The
ven without sensor feedback. Our experiments demonstrate accurate stroke size control across different liquids and fill levels.
-world datasets, we show that HandNeRF is able to reconstruct hand-object scenes of novel grasp configurations more accura
t executes a pinch grasp without needing touch sensors or geometric models. Lab experiments with a robotic manipulator equip

integrates residual blocks with coordinate attention to re-optimize grasping features. Then, the unsupervised low-light feature e
he reconstruction efficiency. We evaluate the proposed method in real-world scenarios with comparison to existing priori-inform
n analytical model. Our work has the benefit of using vision to predict an articulation model before touching the object, while also
ate to perturbations. The object slips, inducing a failure of the grasp and damage. In this study, we introduce a method that uses
a set label for each point such that such that points that share a set label form a cross object correspondence. The alignment m
s refinement method employs a differentiable renderer to construct joint optimization constraints. The experimental results dem
and is dynamically updated throughout the manipulation process, built by sequencing a set of primitives. The estimation strateg
ables dynamic grasp tracking. Our model can efficiently generate accurate, 7-DoF, dense, and temporally-smooth grasp poses
get contact, and facilitate the use of our specialized end effector for harvesting. In autonomous tomato harvesting experiments c
emisphere centered around the object to their desired approach direction. A physics-based grasp planner finds the most stable
ing electrical impedance sensing. We design and mount a digitally embroidered electrode array to a commercial robot gripper.
erification of grasps during execution to avoid costly downstream failures as well as autonomous reward assignment, providing
we show that a classifier can recognize an object from a set of trained ones with a high success rate of almost 95%. In addition
such as bolts and nuts are estimated with accuracy on order of 1mm.
y adjust an object's pose for a better perspective. Addressing this inefficiency, we introduce a novel object rearrangement syste
rst method for 6-DoF object pose estimation with only the event-camera. The proposed method achieves comparable results to
r in the real world with imitation learning using the cloth point cloud as decision support, which is captured from a single RGBD
ospectral parameterization, we can accurately estimate the ground truth linear velocity of the trajectory in an optimal fashion wi
rent scenarios, serving as a reference framework in algorithm design. Our research lays the groundwork for further collaboratio
ed and adopted in the last few years and presents a novel benchmark with protocols for different tasks that evaluate both the s
ark will be a valuable tool for advancing the field of robot manipulation. By providing a standardized evaluation framework, rese
benChmarking paRadIgm for evaluaTing trajEctoRy predIction Approaches (CRITERIA). Particularly, we propose 1) a method
et, tailored to camera-only 3D detection methods. Surprisingly, we find that state-of-the-art camera-based detectors can outperf

tional graphs. Drawing from the knowledge of leading robot architects and system architecture experts, RobotPerf establishes
eveal the influence that different properties have on the outcome of five manipulation primitives. We believe that, in general, res
m by simultaneously adapting online the admittance and the power limits in the controller, ensuring both safety and task perform
m comprises of the Sawyer robot, our specially developed IoT katana and a skill-training program for the Martial Art of Iaido. We
nees. The 3D path extracted from the generated trajectory is then projected onto a 2D plane with respect to the direction of the
udies, simulations and robot experiments to analyse and evaluate the proposed method in several table top tasks where a robo
nd designs controller at the subsystem level. The stability of controller in the presence of contact with the virtual environment an
ntirely in simulation and validate our method in four real-world dressing scenarios of a medical training manikin. We show that o
ding effects of exergaming sessions at the onset and during the squat phase of the squat cycle. We acquired surface electromy
ms at showcasing the external dynamics dependent gait adaptation using a Cable-Driven Leg Exoskeleton (CDLE). A swing ph
allenges, this paper proposes the application of Pairwise Comparison (PC) in pHRC — an alternative survey technique that em
er-based network, showing a better prediction accuracy of the contact forces.
ns similar to humans. This allows us to generate synthetic training and testing data with 100x more objects than previous work.
ponding to their actions. At the top level, the controller consists of a set of finite-state machines corresponding to different levels
ctions between physical human-robot interaction and human modeling. On one hand, human neuromusculoskeletal and senso
that the human-robot pair will follow. An ablation test is also performed to verify the relative contribution to performance of each
C experiments. We explored how the task performance is affected by base compliance as well as human operator’s experie
porating viscoelastic properties at the attachment interface. The experimental method is implemented with the Maestro hand ex
rque sensor allows to detect users' intentions, which are translated into command velocities by means of a variable admittance
e agent to repeat a large amount of another agent's history trajectory so that a sequence of matched observation can be establ
n the self-driving car simulator CARLA at a pedestrian-filled intersection. The strength of our approach is compatibility with state
s is equivalent to a computationally efficient, convex quadratic program. Compared to the nominal TO, our method enhances n
PPI) algorithm that provides robustness against unpredicted disturbances and achieves real-time planning using a much smaller
ce problem by computing safe halfspaces based on obstacle sample trajectories using distributionally robust optimization (DRO
a hybrid belief. We illustrate how the upper confidence bound (UCB) exploration bonus can be leveraged to guide the growth o
To overcome this challenge, we introduce a pruning-based approach for planning with ambiguous data associations. Our key c
 e compute shields that restrict unsafe actions which would violate the almost-sure reach-avoid specifications. We then integrate
expected path cost in our graph and the expected path cost in an optimal graph to determine whether a given edge should be
d motion planner to plan low-risk motions without sacrificing path quality. Our method also explicitly bounds the risk of the path
re defined to satisfy assigned constraint budgets, then COBeTS will satisfy constraints anytime. Otherwise, COBeTS will guide
s 2D pixel-wise uncertainty, exhibiting high values for rays directly casting towards occluded or external scene content. To quan
he sampling distribution of a sampling-based kinodynamic motion planner when the outcome of the executed transitions is diffe
within AEP in DAEP has resulted in enhanced exploration within static environments. To thoroughly evaluate exploration planner
mance in challenging and large outdoor environments, including around buildings, across intersections, along trails, and off-road
ng, and searching over the graph to find belief space motion plans. IBBT is an anytime, incremental algorithm. With an increasi
l human trajectories and their predictions. We consider the risk of collision in the form of a chance constraint, providing an inter
environments on different robotic platforms. Simulations and physical experiments demonstrate that our framework is capable
ssimistically imagines more obstacles. We aim to be robust to uncertainty over potential worlds while still achieving the efficienc
he correlation with the environment by a set of general functions termed perceptual factors to construct a perceptual map, which
to provide more accurate time-varying transition estimates; and 2) an MPSE-integrated planning strategy that optimizes long-te
blem is similar to the Algorithm Selection Problem (ASP), but has several complicating factors that preclude application of exist
informative trajectory planning in continuous state and measurement spaces. An adaptive criterion is proposed in APFT to adju
em) is finite. The finite set allows us to enumerate the possible planning problems that can be encountered online and preproce
istic occupancy information. For avoiding conservatism during motion planning, we directly employ the probabilistic perception i
ed safety guarantees. Benchmarks compare computation time to conservatism of the collision checkers, in addition to character
rol. Further, uncertainty propagation extends the uncertainty metrics to explicitly consider the growth of uncertainty over time st
ating a non-convex optimization problem aimed at minimizing trajectory execution time while incorporating constraints on collisio
ally, through a new formulation of the team coordination on graphs with risky edges problem into Markov Decision Processes (M
s superiority over traditional heuristics such as dispatching rules in dynamic scenarios.
tion space. Meanwhile, distance-optimal path planners search for locally shortest paths in the augmented space, which become
y, as prioritized planning can be shown to be complete when start/goal configurations belong to an LWCS. In this work, we esta
ing time (before task initiation), agents need to carefully sequence tasks but also dynamically form/disband coalitions. While sim
blem with known task outcomes. The higher-level search intelligently generates more likely combinations of failures and calls th
ormation to predict the evolution of future robot density distribution. In sectorlevel planning, we generate sector-level paths that
n. We demonstrate the usefulness of this approach using a case study involving the disinfection of a building consisting of multi
hter and more compact design. A non-linear spring mechanism is proposed to generate the spring torque as needed. The imple
periment, where the ground truth locations of bones were calculated using bone pins. These data were used to train the propos
angles and enables us to position control a pneumatic cable-driven upper-limb exoskeleton, thereby assisting users in various
%, respectively.
, continuous model. Furthermore, this paper proposes a basis model to represent intertask relationships in the impedance para
developed to reveal the variable moment of the ankle joint, which is verified by human-machine simulation. An AFP prototype
reaching tasks while connected to an arm exoskeleton, experiencing one of five tested force fields. Their movements before an
combines decoupling of the electrical activations on motor neuron axons with extraction of motion patterns from the discharge t
based on accurate spring deflection feedback, and drives the hip joints via an efficient cable-roller mechanism. Taking advantag
actuators. The resulting prosthetic wrist, incorporating the artificial joint and artificial muscles, is lightweight at only 50g and can
ht- and curve-walking scenarios with four users. The study demonstrates that the exoskeleton frame adapts to the natural 3 Do
d exoskeleton through testing with a complete paraplegic user.
 ng time performance by up to 11 times compared to machine learning approaches. The efficacy of the developed finite-state m
ted Center of Mass, which has been established as a consistent dynamic for describing human movement. This real-time online
with the device turned off, on average, participants experienced a 47.5% (range 12%) metabolic reduction when walking with t
of task-space control for a powered ankle prosthesis, characterizing the emergent outcomes of this new coordination strategy.
r behaviour accordingly. To address this issue, we have developed a control strategy for a soft exosuit that employs computer v
r the number of cameras changes. Without any re-training or fine-tuning, our models outperform models trained with evenly dis
e provide the methodology and mathematical developments for three conversions considering three different types of cameras
s these challenges by using a radar configuration with 1/4th of the range resolution and employing a deep learning model with 2
ve matches allow us to recover the relative transformation between scans through weighted Singular Value Decomposition (SV
ailable at https://github.com/jisehua/Degenerate-Detection.git. Experiments with public datasets and self-build robots were cond
y usage, and local on-tree updates that significantly reduce computation time compared to existing static or dynamic tree struct
could be used as ground truth. In this paper, we introduce two novel point cloud map consistency losses, which facilitate self-s
he-art methods using the SemanticKITTI dataset and a novel evaluation method relying on airborne LiDAR scanning. The result
model is able to produce high-quality and complete meshes with finely detailed structures from a 2D image captured from arbitra
aluations on indoor and outdoor maps, verifying that the proposed system compares favorably to fixed-resolution RMP variants
e efficacy of tactile sensing for wall-following in intricate environments. When paired with an external method for pose estimatio
put sequence in the MPPI process. Owing to the interpolation, smooth control input sequences can be obtained even though th
f the robot’s embodiment with the reactive force-torque (wrench) field giving rise to flexible tool usage in non-stationary env
smooth, and efficient trajectories. The effectiveness of our proposed framework is validated through both numerical simulation a
man-robot collision detection, and robot programming interface. These features, integrated with an easy-to-use calibration metho
uman demonstration videos. The task representation is composed of keypoint-based geometric constraints on principal manifo
plex simulation environments and experimentally verified the approach in realistic scenarios on real soccer robots. Furthermore
s compellingly demonstrate that our proposed algorithm consistently navigates uneven terrains with high efficiency and surpass
Algorithm. Comparative analysis revealed that EC methods consistently yield more precise and optimal solutions than brute for
atic method for quantifying power transmission losses at each transmission stage from the battery to the wearer, using data ga
hed by high back-drivability, robust torque generation capability, and safety features. To attain mechanical transparency, we cre
ory are attainable for the patient, so reversing the trajectory is a safe operation. Since there is no clear way to optimize such a g
ferent walking styles defined by user-selected gait parameters such as step length and clearance. The primary aim is to improv
e model of the real system. We demonstrate our RL-based control on a 2-DoF planar setting where the subject's arm, placed o
ere the voluntary movement of the patient is prioritised and assistance is provided using FES and the robot in a hierarchical orde
ises. The robot's soft end effector serves as a target that users repetitively reach towards and press with a hand, foot, or knee.
s with various goals. To evaluate the method performance, we build a benchmark based on the robot-fetching environment with
Gym tasks, achieving significantly higher rewards and near-zero safety violations during inference. The method's real-world ap
precise, authentic sampling critical for robust learning, while also preserving computational tractability. Through extensive evalu
of policy and CBF coefficients, to enhance safety and long-term performance through reinforcement learning. By directly interac
ut the environment: (i) locomotion of a quadruped robot traversing challenging terrains and (ii) a quadcopter drone racing throug
ace. This allows SORL to find a natural tradeoff between safety and optimality without compromising the performance in terms o
reinterpreting stored image-based observations as a video, we leverage lossless differential video encoding schemes to compr
the MT10 and MT50 benchmarks of Meta-World consisting of 10 and 50 goal-conditioned tasks for a Sawyer arm.
poral forward refinement to achieve bi-directional temporal refinement (BDTR) while maintaining a robust temporal dependency
ecting the AR display’s visibility. Resolving the problem is quite challenging because the optical system operates on magni
Extremum-Seeking Control (ESC). On each action sampled from stochastic policies, we apply sinusoidal perturbations and que
gorithm, utilizing gradient inversion to reconstruct states, actions, and supervisory signals. The choice of using gradients for the
er than existing platforms. Furthermore, equipped with an attention-based Multi-Agent Target Generator and a Single-Agent Mo
mulation data. Specifically, we pre-train a generative adversarial network (GAN) model to fit the state distribution of the offline d
a supervised learning with on-policy datasets. By varying the disturbance strength in an adaptation module, i.e., adopting adapt
arning (HDC-RL). It renders robots to show precisely situated reactive behaviors in complex environments. Our empirical evalu
rizes the properties of different types of DLOs. The ground truth flexibility data is collected in simulation to train our flexibility es
learn an agent that can consider the diversity of LTL instructions due to uncertain event detection while avoiding task failure du
domain to encompass attributes like texture, mass distribution, or sounds, which significantly influence how we interact with ob
s the environment with the robot's in-hand camera. It then selects and relocates the non-target objects such that they do not blo
ke large-range or even 360-degree scenes is very challenging. In this paper, we explore the potential of MPI and show that MP
e speed of the deep learning-based signed distance computations with the strong accuracy guarantees of geometric collision c
ds have different scale factors. Key components of Reg-NF include a bidirectional registration loss, multi-view surface sampling
-OAE), to facilitate the detailed supervision inherited in local regions and global shapes. We propose to randomly occlude some
g "what" the entity is. Thus, our pre-trained object-centric representations for control are constructed by appropriately combining
objects. We also propose Shape-conditioned Trajectory Deformation Network (SCTDN), a model that learns to generate SKT
educe the concept, establish the task fingerprints as an intuitive image-based representation, experimentally collect task fingerp
of multi-object scenes. These slots robustly bind to distinct objects, even under heavy occlusion or absence. Our SlotGNN, a no
Rs, primarily data-augmentation and fine-tuning, also transfer to the real-world performance.
metric as a representation to keep the visual features consistent through a sequence with the robot’s relative pose transform
her, we find our proposed PCWM significantly outperforms prior works in terms of sample efficiency during training. Taken toge
el-based representations do not have this problem but they are not space-efficient especially in large environments with fine de
and describes them with PVT activations, producing an object-centric representation. Across standard choices of generic pre-tr

pled points according to their local geometric properties to facilitate multi- resolution segmentation. Subsequently, we arrange th
e inpainted images to 3D by predicting normals of the inpainted image and solving for the missing depth values. By predicting f
ution divergence, we design a simple but effective pseudo-siamese network to improve feature homogeneity. Experimental resu
specially in distant regions, we introduce a disparity-depth conversion module. Our method is both accurate and efficient. The e
nd 3DMatch, even outperforming some older supervised methods. We also integrate our components into existing methods, sh
ctive vital for addressing real-world scenarios with higher fidelity compared to datasets reliant on specific vantage points. Additio
unseen object categories in cluttered indoor environments. Our method, SupeRGB-D, groups pixels into small patches based on
central component is a hypersphere attention mechanism, which updates object queries on a hypersphere. To illustrate the effe
h prior. Volumetric semantics is computed by a 3D U-Net. We propagate the hidden 3D U-Net state using the sensor motion an
better domain adaption. Meanwhile, to prevent the model from forgetting previous curricula, we integrate contrastive learning in
which significantly reduce the package size to share among connected vehicles. Then we design a novel spatiotemporal graph
of estimating its own motion (from the static portion of the scene) while simultaneously recovering the motion of all dynamic obje
ucted DLO. By testing on a synthetic DLO dataset and further validating on a real-world dataset with seven different DLOs, we
e, which is not possible with proprietary distance estimates alone. This is accomplished via a differentiable rendering pipeline, w
l-text pre-trained models to a 4D point cloud network. Our approach involves aligning the 4D encoder's representation with a V
raphs is built by leveraging 2D foundation models and fusing their output to 3D by multi-view association. The resulting represe
an kernel that only penalizes lower-order spatial displacement derivatives to reflect DLO physics. Improved real-time DLO track
While previous work in the 2D domain has studied imposing constraints on the classifier weights to regularize the training, it is se
ability in environments with form-factor limitations and minimal human intervention. The robotic arm—comprising links, joints, a
trackers are too large to be installed on roofs and (2) even if traditional solar trackers were made in a more compact form-facto
d for open loop control only, we discuss how this method could be adapted for closed loop control and sim2real transfer to phy
sted in a table for reference. The transition pressures between the two stable states were experimentally characterized. Moreo
cs and little insight into their singularities can be gained by analyzing the determinant of their Jacobian. One such robot arm for
mpared against other robotic and biological fish. We operated the robot, untethered, in a variety of environments to test how it w
4 BL) and a figure eight, which it completed in 4.98 s (72.3 degree/s) and 10.74 s respectively. We operated the robot, untethere
ed hybrid parallel robot achieves a high workspace-to-footprint ratio comparable to that of serial robots. A CAD model of the rob
pe and a pulley-guided prototype, evaluating wire tension distribution and payload capacity.
odels to make it possible to explore this possibility in the context of robotic manipulation, alongside experimental results that pro
ained on internet-scale data, as well as (b) Point Cloud Transformer (PCT) robotic policies operating on large point clouds. We
mm basil seeds, thin paperclips, and items larger than 15mm such as bearings. This work demonstrates the potential of soft op
age large-scale audio-visual pretraining to obtain representations that boost the performance of robotic manipulation. To the be
ure hand and object configurations. We combine this with a `translation’ module that learns a plan-conditioned robot manip
metrization with existing motion planning algorithms, including sampling-based approaches, trajectory optimizers, and technique
ms under sensor faults and attacks. Our main contribution is the development and synthesis of a new class of CBFs that we term
ers as state estimator. We demonstrate BCBFs on a quadrotor that is exposed to external disturbances and varying sensing co
's efficacy through mathematical proof, affirming that the maneuver set obtained through the application of the algorithm is inde
ly evaluate the proposed CBF's behavior and show that the resulting optimal control problem is computationally efficient, which
ng an Extended Position-based Dynamics (XPBD) algorithm. Together with the desired control objective, physics-aware regular
omputational graph analysis tool to predict the possible future states of the system and verify that they satisfy safety constraints
testing equipment for legged robots. A pneumatic linear impactor is also adopted for the case study involving the humanoid ro
then used to train a classifier that is leveraged online to flag inputs that might cause system breakdowns. The anomaly detecto
y of the system up to some level of risk. For this work, we focus on learning the dynamics uncertainties of a quadrotor drone, w
d a control algorithm that utilizes the learned IPC, with the goal of landing the quadcopter safely on a landing pad. Experiments
distance functions. We also provide formal guarantees on the safety of the exploration targets and the completion of each explo
xact. This paper presents our recent efforts to address this issue. We present a novel Model Predictive Control scheme that ca
s optimal control problems to compute states that are guaranteed to be on the boundary of the set. This allows us to learn direc
he CBF-based safety constraints to adapt its navigation behaviour, which is particularly crucial when potential scene changes o
s. Dynamic models, validated with simplified 1 DOF experiments, show that the safety level of collaborative robots can be incre
ating a safety specification, such as avoiding a dangerous area, is bounded. We provide an open-source implementation as a R
tends results from conformal prediction to calibrate decisions directly, without requiring the construction of prediction sets. Expe
f human-driven vehicles in complex, e.g, off-road, conditions. While conservativeness can be reduced by using multiple backup
es aggregated in their local, cross-, and global context. In addition, we introduce a novel node pooling mechanism that leverage
efficiently computes epistemic and aleatoric uncertainty in a single neural network forward pass. Our method is enabled by thre
hile the never-before-seen efficiency is largely enabled by the novel SRF update method that is based on our new permuted-Q
egradation, our SeqTrack3D combines both historical point clouds and bounding box sequences. This novel approach ensures
eural implicit representation enables the rendering of high-fidelity RGB-D images with the extraction of explicit, dense, and inter
UVD discovers subgoals by detecting phase shifts in the embedding space of the pre-trained representation. Operating purely o
rformance improves when the gripper yaw w.r.t. the apple has two suction cups on the bottom of the apple and one suction cup
 a. To address these issues, we propose a dataset for estimating strawberry weight, which includes strawberries with different h
r automatically pre-processing grain kernels and efficiently acquiring high-quality images for these kernels. The framework emp
nt activities. To address this, we propose a novel inter-day registration approach that only relies once on GCPs, at the beginnin
t by farmers today or on future viticulture robotics platforms. We achieved an R2=.80 for VSP vines (better than state-of-the-art
in current breeding practices, we select a subset of leaves to be used for phenotypic trait evaluation as this allows us to allevia
ulate tree crops. We conduct a comprehensive set of experiments in a simulated environment and demonstrate generalizability
utdoor environments, was shortened considerably (3.38 m, 3.6 m, and 5.6 m, respectively). Especially in the face scenario, the
nstance segmentation on 3D point clouds, which is a first step toward phenotyping in 3D. Separating all leaves allows us to cou
ed planning, to determine where to collect leaves to get the most informative readings; 2) system design and integration for aut
ith the image feedback from both cameras and the proof of asymptotically convergence is provided. Furthermore, the cumulativ
agricultural domains that lack publicly available labeled data. Motivated by similar success in urban scene parsing, we propose
osed skeletonization method is shown to be performing better than state-of-the-art, especially in handling very thin and delicate

modes, i.e., free-flight and object-extracting, and the rule of state jumps. Also, we design control strategies which enhance the
ctories onto the constraint sets before evaluating the cost functions. We further learn a deep generative model in the form of a C
s to produce a filtered virtual velocity control input, which is tracked by an adaptive controller to overcome modeling uncertaintie
o these novel scenes. Motivated by this challenge , this paper presents emph{Osiris} which is a 3D scene graph builder for agric
s the overall time needed until the data can be served via the spatio-temporal-semantic query interface of the semantic environ
ense the sweet pepper peduncle position and the end-effector state (harvesting failure) to complete the precise cutting. An end-
gular position of the transplanting part is also proposed. For path following control, sliding-mode control and inverse optimal con
olutional Neural Network (CNN) on a novel dataset we built for this purpose. Predictions are performed while the robot is moving
nomously detect and cut pruning targets with minimal control of the environment, laying the groundwork for a fully autonomous
mark correspondences approach, integrated into an underlying Kalman filter system to recognize the surrounding trees and join
and maintain the high occupancy resolution required to informatively plan around the fruit. When sizing, a robust estimation and
obal view planning and used random sampling of candidate viewpoints for exploration, which could suffer from high computation
e motion constraint arises during insertion of the sapling within the growing tube, whereas for harvesting, it arises during extrac
e map as a pretext task. Our approach enables the network to learn unequivocal representations of objects seen in one datase
train machine vision systems. We use StyleGAN2 to generate images that appear like they are from desired new field or robot
g and control of the log loading maneuver. Our main contribution is in the full integration of the necessary elements to achieve a
dopted for obstacle searching and comparing. Using the detection's searching results as feedback reference, the Extended Kal
er at breast height together with the individual tree locations in a plot. We validate our results against manual measurements co
ity relationships and evaluate these hypotheses in real-time, allowing this new knowledge to be immediately exploited for future
works on herding using flocking control, local observations of the nearest animals and communication with other robots within
strate that a collective modular robot achieves desired motion tasks while keeping all normal contact forces of the connectors be
eline algorithms. PLASTR performs equal to or better than the best heuristic in terms of material coverage for both simulated an
bes the design, development, and evaluation of this assembly system, including the robotic hardware, lattice material, and plann
criptors. We demonstrate our approach using datasets collected in a simulated environment and a real-world dataset captured
y simple, and we implement it on AUV without any offline computation requirements. Combined hardware experiments and simu
equiring 20 bucket fillings. Additionally, we showcased the adaptability of a controller pre-trained on woodchips for an unknown
empirically show that uniformly sampling the domain parameters is insufficient when the task difficulty varies according to the t
hree modules: 1) an Omni-directional robotic data collection platform equipped with a Realsense D435i camera for localization,
ocess, which takes action language and environmental sensing as observations and outputs the action. The results of action ex
We present Grasp-Anything, a new large-scale grasp dataset synthesized from foundation models to implement this solution.
m to compute a sequence of finer-granularity actions that jointly achieve these goals. We ground and evaluate our framework's a
he behavior of high-level plans using end-to-end demonstrations. Using Language-World, we show that PCBC is able to achieve
the environment using a robot, and infer the state of the underlying problem from collected observations while guiding the robo
cene graph and the task automaton, and provide semantic guidance via an LLM heuristic function. To guarantee optimality, we
uman-annotated plan correctness metric from 28.89% to 49.63% over SayCan, whilst achieving competitive executability. Our
eriments on Language Augmented TaskGrasp (LA-TaskGrasp) dataset and demonstrate that GraspGPT outperforms existing
bidirectional communication feasible; (iii) An analytical profiler that collects any set of user-defined performance metrics for syst
n of the control gains, and chattering attenuation. Furthermore, the stability of the system in closed loop is guaranteed through L
nd rotational levels into a single vector term. The weights of the NN are determined online by an adaptive law based on the Lya

em states, predicting forthcoming real states as initial conditions for the imminent OCP solution. Anticipatory computation store

es in tire models that can occur during actual races, particularly when pushing the vehicle to its handling limits. we present a un
es of uncertainty as compared to a standard Extended Kalman Filter (EKF). The filter has the same computational complexity a
minated, such that the chattering effect of the controlled system on the surface is restrained consequently. Finally, simulations o
aints. The proposed method is formally proved to ensure closed-loop robust stability, and is intended to be applied to any kinem
xperiments using a humanoid service mobile robot alongside simulated human counterparts. Our results demonstrate the capa
pproach geared toward ensuring the safe execution of learned orientation skills within constrained regions surrounding a refere
al-time for online robot control 3. as precisely as possible for optimal control. In this paper, we propose a batched, fast and prec
Network (DNN) regression model was implemented. First, a robust human-pose keypoints detection to calculate shoulder-elbow
anner naturally leads to prosocial interactions. Specifically, we introduce the notion of a markup factor to incentivize legible and
human-driven vehicle by formulating a human planning model as an optimization problem. Lastly, we present a Stochastic Mod
gorithm development and tuning, we develop an attentive processing strategy that focuses exclusively on key input features. Th
ontradicting, constraints. Joint space constraints are formulated from efficient data-driven self- and external C^2 collision bounda
-based robust fast (RAST) learning method that selects training data near to the boundary by considering adversarial samples.
siders non-standard human body pose. In this work, we introduce a stereo vision system for enhancing UHRI that utilizes three-
refine their thermal model and optimize the line capacity during heat waves. All three applications were recently field validated
es. An H-infinity-based disturbance-rejection controller was designed to suppress the vibration of the head nozzle unit posture a
oblem, and proposes a search-and-rescue framework for cellular-enabled collaborative robots use cases that, taking as input th
ge only on overlapping regions of sequential sub-graphs. Different from the state-of-the-art graph based exploration methods, t
-object search.

ttern with 3-connectivity is investigated in Open space and a linear 4-connectivity pattern in Corridor space, respectively. The p
d lower-cost hardware while still retaining good autonomous performance. The approach further consists of a methodology to in
or sensor data (e.g., occlusion, transparency, poor-depth, in-hand). This does not only make a substantial step forward compar
the global frame, the proposed method generates a semantic map as a distribution over possible region labels at each location
roposed objects to make a final grounding decision. Our method does not require any labeled training data and can generalize
e self-supervised labels, we then train a neural network that identifies terrains that are safe to traverse from an image using a o
blem of the receptive field enlarging along the radial distance, which is not desirable in LiDAR point clouds since the size of an o
ds for semantic segmentation and boundary estimation. We show that our approach, albeit being trained with only ten annotate
nt minimization of both losses via the Hungarian algorithm. We thus achieve end-to-end training for a semantic segmentation n
former introduces the deformable mechanism in multimodal fusion, avoiding over-attention and improving efficiency by adaptiv
ud robotics platform, to offload resource-intensive tasks. Lifelong LERF obtains poses from a monocular RGBD SLAM backend
uerying and integration with additional knowledge sources.
ing, we propose a novel geometric quasi-planar over-segmentation method that groups 3D map elements likely to belong to the
ading to frontiers. The algorithm for generating semantic topometric mapping utilized by the proposed method is lightweight, res
ted with the 3D knowledge from TSDF using an enhanced Hungarian-based feature matching mechanism. In particular, Open-F
tead, Mask4Former introduces spatio-temporal instance queries which encode the semantic and geometric properties of each
approach that can be trained end-to-end and directly predicts a set of non-overlapping masks along with their semantic classe
Priors (HSP) and Deep Reinforcement Learning (DRL). The HSP contains two components, i.e., the egocentric semantic map-b
twork (GCN) model is developed, to learn CECI from a repository of 3D scene graphs. As no database of 3D scene graphs exi
o semi-supervised machine learning in the robotic VTM domain. We show through simulation and real-world underwater data th
o-end collision-free space detection network, referred to as SG-RoadSeg, built upon our previous work SNE-RoadSeg. A key c
sed to conduct transductive semi-supervised learning to propagate labels. To optimize the graph structure to learn discriminativ
work.We propose a full-resolution backbone to prevent information loss in sparse point cloud processing. Our instance transfor
calibration, GLFS, which achieves simultaneously higher accuracy and 3D map calibration while retaining real-time capability a
research. To enable better integration of satellite maps with existing methods, we propose a hierarchical fusion module, which
ity. It also improves the accuracy and robustness of the neural network by forcing the network to segment and classify objects
ures. To semantically decompose the obtained 3D map, we build a lane graph from the paths of ego agents and their panoptic
mera branch and induces accurate location of dense camera features in BEV space. For an effective BEV fusion between the s
e predictions by capturing contextual information and occlusion area features. The framework subsequently employs a query-b
c objects. Moreover, spurious radar points from the curse of RF multipath make robots misunderstand the scene. mmEMP first
erated high-quality framework for data acquisition, model training, and performance evaluation that functions with the CARLA s
a lightweight spiking neural network (SNN) to efficiently leverage the asynchronous property of events. Buttressed by it, we firs
nhances the 3D feature via 2D-to-3D fusion and 3D-to-2D fusion, respectively, which surpasses either one of the single fusion
of deep learning models for missing modality scenarios for semantic segmentation tasks. Gated Spectral Unit (GSU) is also pro
which uses an image as input to infer a dense LiDAR shape. Boosted by our UYI block, our experiments show a significant perfo
ety of autonomous driving. To this end, we propose a novel network named SuperFusion, exploiting the fusion of LiDAR and ca
t and experiments. The proposed sensor fusion system is real-time capable and delivers superior performance over state-of-th
we open source the GNSS/INS/Camera Integration Library (GICI-LIB), together with detailed documentation and a comprehen
or event-based moving object detection. Specifically, we first compensate the motion of background events using inertial measu
des a deeper understanding of large-scale patterns; the Sparse Fused Feature Attention (SFFA), which fine-tunes the fusion of
with sparse Radar points, quasi-dense scale estimation through learning the association between Radar points and image patch
n. This paper introduces HabitatDyn 2.0, a dataset specifically designed for enhancing object localization capabilities with sema
s that leverage the strengths of both modalities. Through comparative experiments, we demonstrate that our approach surpass
ed using simulations. Our experimental results showed that the proposed method accurately performed calibrations using a ca
y in the modality fusion process, thus degrading the depth estimation performance.In this paper, we propose a novel Spatial Re
all performance when compared to the state-of-the-art homography nonlinear deterministic observer in both simulations and ex
if the robot is physically capable of continuing the mission, its operation will be compromised due to a corrupted representation
to overcome dynamic environments, which captures more target-aware information through the hierarchical attention mechani
 dality, a new data set with synchronized event- and depth data was recorded. The results show the viability of temporal encodi
pact reduction feature can be added to an already-implemented impedance controller because the parameter design is divided

closed-form matrices, we are able to provide conditions for stability with respect to errors in extrinsic parameters and surface n
the optimal exposures for the next bracket are obtained directly without attempts to ensure a quick response. Experiments sho
ii) significant affordability, and iii) fabrication-friendly construction that requires minimal assembly skills. We evaluated the prop
del incorporates a Mel scale learnable filter bank (MLFB) and a hierarchical frequency-to-time attention network (HiFTA-net). Th
peech signal is computed from the 4-mic circular array microphone. The speech module is integrated with the Robot Operating
control of various writing utensils. Specifically, an unsupervised image-to-sequence model decomposes a given calligraphy gly
the benefit of having a template representing our object of interest with the advantages of using a neural network to learn how
as also been successfully executed on the Perseverance Mars Rover, demonstrating the practicality of our approach.
e three-dimensional spatial awareness for navigating under overhanging obstacles. Our study includes the development of a pr
h explores the necessary information and modifies its prompts to recover from failure. We experimentally confirm that the syste
ns. To improve the sensitivity of the network in distinguishing state variation of the handled quilt, we introduce an enhanced los
This design utilizes both advantages: generalizable intermediate representation from keypoint correspondence and strong mod
an access the ground truth in simulation. With this adaption mechanism, NUVS outperforms the classical IBUVS algorithm whe
ree-wheeled omnidirectional mobile platform, a torso and a novel movable neck mechanism with a face. The prototype robot ca
ffers many inherent advantages for nursing robots. The prototype is tested and analyzed, and the kinematics and statics are ve
applicability of fiducial markers, contributing to the creation of robot-friendly space for future service robots.
hs of the Unstructured Information Management (UIM) principle and the flexibility of Behavior Trees to model task-specific perce
ue by considering an environment of interest and design constraints. The area coverage of a candidate morphology design is ev
, particularly its impact on executability, correctness, and time complexity via empirical evaluation in the context of a simulation
plications. Our method is made available in the open-source robotics toolkit Drake.
ation in crowded environments. SocialGAIL utilizes an attention-based graph neural network to encode observations and emplo
specifically designed for DRL applications. Due to its high level of abstraction, complete implementation in C++, and rigorous t
annot robustly prevent penetrations. In this paper, we propose a framework for strongly coupled simulation of rigid and deforma
eal fish kinematics. Our results highlight the advantages of using real fish body kinematics for a more accurate representation o
dings to control not only the robots, but also the environments. This work also performs sim-to-real rock instance segmentation
ments show that SceneControl achieves greater realism and controllability than the existing state-of-the-art. We also illustrate ho
frictional contacts. We conduct extensive experiments to demonstrate the effectiveness of our differentiable physics simulation
priate colliders for the rollers of a Mecanum wheel. Additionally, the impact of significant physics parameters is presented. To br
our TCKDD achieves textit{state-of-the-art} performance in point cloud registration. Furthermore, we design a keypoint detector
truth and are not flexible enough to adapt to new research trends. To address this gap, we introduce Campus Map, a novel larg
m for reliability and accuracy. The effectiveness of DISO is evaluated using both simulated and real-world sonar datasets, show
based LiDAR odometry techniques, CURL-MAP formulates LiDAR pose estimation as a unique optimisation problem tailored fo
wed smoother with sensor information-specific factor formulations which enable, in the case of degeneracy, partial information to
computational load in check while enabling an accurate representation of the scene. We evaluate our method in terms of accu
nisms to detect sensor anomalies and degradation, handling them adeptly to maintain system accuracy. Our experimental resu
changes or sparse data collection. Our comprehensive experimental results on benchmarking datasets validate the effectivene
tion invariance offered by feature points. Consequently, ONeK-SLAM achieves high-precision localization and detailed object-le
eover, since landmarks are conditionally independent, we can do this in parallel, disconnecting a pose from all the landmarks.
 escriptors. In contrast, we use a density image bird's eye view representation, which is robust to viewpoint changes. The utiliza
hen a new scan arrives. To efficiently estimate the distribution of points, we extend the ikd-tree to manage the map in voxels an
a filtering step to improve the accuracy of the points in the resulting map. We put our work to the test using the View of Delft an
m of finding the maximally consistent set of measurements in an incremental fashion. It evaluates the consistency of each loop c
e nearest neighbor search strategy, thereby preventing the exclusion of repetitive yet valid matches during the early stages. Ou
andmarks in a coarse-to-fine manner instead of directly using object residuals in bundle adjustment. Firstly, we introduce an imp
t efficient multi-feature sharing architecture. Lite-SVO is designed to bolster self-supervised SVO, facilitating its adoption on edg
it yields far more stable predictions when compared to state-of-the-art rational and polynomial spline methods.
s is also improved by combining the temporal stereo results and the static stereo results. To circumvent the degeneracy issue o
loor level. The above graph is optimized in real-time to obtain a robust and accurate estimate of the robot’s pose and its m
for researchers acquainted with the VPR problem, it examines the intricacies of different VPR problem types regarding input, d
-inertial odometry based on Doppler velocity information, challenges remain for the estimation of 3D motion because of the disc
nature, our algorithm also amplifies the efficacy of modern keypoint detectors, such as SuperPoint. The implementation of our
well. Extensive experiments validate that our proposed method significantly outperforms the state-of-the-art methods on several
ation combined with IMU measurements to obtain a highly accurate and robust LiDAR inertial odometry estimation. Furthermore
ient constraints on the LiDAR odometry. To solve this problem, we propose a marginalization method to retain prior scans' cons
vely and quantitatively on two open-source datasets. This evaluation illustrates that the proposed loss function converges faste
that considers the relative pose uncertainty, the map discretization errors, and the LiDAR noise. In this way, we can substantia
ment the accuracy of scan-to-map registration, particularly in large-scale and complex scenarios. We validate our system's perfo
s not necessarily lead to the most accurate and robust solutions. This paper conducts and presents a thorough analysis of vario
odule can distinguish the kid planes in different voxels and merge these kid planes to estimate the father plane. After merging, t
d formalises the concept of point cluster, which encodes all raw points associated to the same feature by a compact set of param
ng offline methods. DORF first conservatively preserves the definite static points leveraging the Receding Horizon Sampling (RH
ontribution of this work is the meshing module, which represents a scene by an efficient voxel structure, performs fast finding o
o formalize the sensor arrangement problem as a form of subset selection under the E-optimality performance criterion. While
cessing of hybrid point features. We present an efficient matching approach to exclusively perform the data association within t
a translation with reduced computational cost. For the rotation, the parallel optimization is performed with multiple initial values,
ation gives virtues twofold. First, it is scalable. For any dataset, when a weight for trade-off is given, this algorithm determines th
eal-world tasks and demonstrate that our technique converges to model estimates that lead to better solutions as evidenced by
ing the existing infrastructure is not always feasible. UWB has also been used in peer-to-peer ranging collaborative SLAM syste
lso reveal that using the physical properties results in fewer radar points for odometry estimation, but the performance is still gu
vironments, we overcome the challenge of SLAM with sparse sensing owing to the four-point LiDARs. We expand the applicatio
ue of the conventional metric, enhancing the identification of true loops. Employing a state-of-the-art (SOTA) keypoint and featu
use this representation for flexible room classification and segmentation, serving as a basis for room-centric place recognition.
ng dynamically, and camera-object relative transformations at each time step. This shift not only offers a solution to the challen
ame-to-map factors for the live estimates or as map-to-map factors aligning the submaps with respect to one another. Experime
existing approaches in the literature with additional proposed baselines and propose a novel data-driven approach. The result d
needed for view selection. Further with these two tasks, we propose an adaptive strategy for switching modes in view path pla
provide the largest rewards, we can maximally reduce pose graph uncertainty while avoiding long travel times. The effectivenes

mplete stereo VO pipeline that estimates directly with individual event-measurement times without requiring any grouping or app
 ieve improved behavior, particularly when linearization points largely deviate from the truth (i.e., on transients upon state disturb
ap. Finally, we use the built local lane line map to correct the position of the vehicle. Based on the corrected position and prior p
ulating the behavior of non-cooperators, which can effectively identify vulnerabilities of a navigation policy. We evaluate the app
the car to know whether it is an advanced model car made out of Legos or a toy car built by a toddler. We propose an approach

dal than unimodal explanations. Additionally, our findings reveal low user satisfaction with explanations of a robot with extreme p
robot navigation. We employ spatial graphs to capture the current spatial interactions, and through the integration with RNN, th

hese observations suggest significant differences in cognitive and communicative strategies in human-human versus human-ro
ducted a pilot study using the system with groups of high school children in Japan and Australia and our results show the poten
n-robot interaction which requires less user feedback to achieve a good adaptation result. In the proposed ChatAdp, we use Ch
The interdisciplinary approach, combining electromagnetic theory and robotics engineering, shows promising potential for adva
posture states, facilitating autonomous navigation and obstacle avoidance in rapidly changing environments. Simulation results
heoretical model is introduced to construct the energy landscape of the structures and the force response of the gripper when p
fect during the deformation of the actuator, we used compressed air to drive the actuator and injected the beads into the actuat
ned with a rigid, bistable mechanism there is great potential for improving the accessibility of sampling and research in this cha
8 kg (limited by the tensile strength of the vine robot body material and winch actuator torque capacity), as well as novel materia
ry for sample efficiency by augmenting the training data and constructing auxiliary losses to force the agent to adhere to the sy
a qualitative measure for the contact estimation. This paper first presents the probabilistic force estimation model and then sho
rised raspberry physical twin. Then, followed grower-guided protocols to pick blackberries in farm polytunnels, and to evaluate
es are proposed based on a theoretical model. The experimental results demonstrate that the robot can move at an average sp
engulf objects then subsequently transport them with the crawling gait. The rolling gait discretises Tetraflex’s locomotion int
h a novel flow-controllable open-circuit hydraulic actuation, we could keep the manipulator smoothly operated and depth-compe
The wire and its array facilitated the perception function of the soft gripper and acted as the Electronic skin (E-skin) to acquire in
ents helical spring structure features scalability, hollowness, lightweight, and a big length-diameter ratio. To perform the critical u
scopy robot is developed base the actuator and tested in a phantom. The results of the task simulations substantiate that the pu
luding gecko-like crawling and inchworm-like crawling. A single gait or a combination of multiple gaits, as well as shape-morphi
h specialized TPU finger pads for gentle picks, a near-infrared (NIR) reflectance-based probe for detecting full ripeness and a st
eral fins fanning to propel itself forward. And the flexible printed circuit board (FPC) constitutes the body of the fish and is also a

4 degrees with zero-power shape- persistence. The design is modular and rapidly manufacturable via a layered laser-cutting tec
ment soft fingers that are able to house the new illumination system, and we integrate them into these various palm configuration

” to retain its deformed shape. The interlocking force between the jammed sponge and the asperities, and the increased are
h the required mechanical properties. We utilized Twisted String Actuators (TSAs) to fold the target 3D structures from flat plate
OF control, the liquid robot is able to perform various actions and implement versatile manipulation tasks, such as transporting o
ess this challenge. The proposed approach involves developing a computational modeling framework that evaluates the electro
bing soft robots and exploration soft manipulators are demonstrated. The soft actuator could exhibit a combined twisting-bendin
tendon-pulling and air-pushing, the robot can achieve independent control over stiffness and position in three-dimensional spa
t any assembly process and therefore our hand can be manufactured for less than 300 dollars. We test the hand in conjunction
nd 0.67 Nm at the PIP and DIP joints. Performance evaluation, including comparison with existing methods, demonstrated the
ators can twist 9.6 degree per millimeter in length, which is significantly larger than previously reported soft twisting actuators. A
neumatic force, eliminating the need for additional rigid hardware, power sources, or complex control procedures. This mechani
ne side of the vine robot to enable high curvature and large force application. Our key development is the introduction of an ext
monstrated. A specific architecture has been pursued to reduce the overall number of fluidic tubes required when adding a new
, with reference to the flexible state. Likewise, an internal routing scheme vastly improves the stability of the proximal segment w
under continuous contact conditions to address the problems of non-linearity, hysteresis, and non-injective mapping between sig
measurements with minimal disruption to the grasping surface. With only 12 sensors, 6 per lattice, this gripper can estimate an o
nd for the control of accuracy, while the CNN model reduces the computational cost, offering a sufficient accuracy during the sy
nerally lack the ability to sense the force between the human hand and the glove. Therefore, this paper designed an array flexib
tial to replace current systems based on the results of an application case study presented in this paper. Simulation driven optim
e were validated through both bench tests and human subject tests. The gradual sliding of the structure’s layer resulted in
d the application of forces over 0-5 N showing a 0.38 N average prediction error and resolution of 0.25 N. The sensing capability
on experiments, including a real-time shape prediction through a graphical user interface.
e, which led to the optical waveguide sensor being sensitive to bending deformation. Compared with EMGs, the accuracy of con
M) based model to then estimate the shape of the soft body and the magnitude of an external force acting at a known arbitrary lo
rformance on increasingly complex testing data generated by a robot, and also demonstrate the full pipeline from human demo
scoelasticity. Then, an Enhanced Exponential Reaching Law based Sliding-Mode Controller (EERLSMC) is proposed to minimi
sing framework for soft continuum robots. The core idea behind this work is that point loads lead to unique deformation profiles
mming, the proposed actuator has a special function of shape sensing that estimates the tip location of the actuator based on im
oncept, design details, and experimental results showing its behavior in response to infrared (IR) and visible light. Our simple d
eform boundary surfaces, and use nonlinear mechanical modelling and shape derivative based optimization to navigate the hig
biochemical reaction of sugar and yeast. The biochemical reaction is utilized to generate contained pressure, which in turn pow
on of single and multiple press events, detecting single indents of sizes greater than 1 cm, forces larger than 2 N, and various l
se can be interconnected for forming user-defined multi-rate control structures which can be executed as synchronized real-tim
e a tooling suite for automating the development of microservices for robotic applications based on the Robot Operating System

ntly share, reference, and compare tasks, robot models, and solutions.
ation, featuring NVIDIA's Isaac Sim and OpenFOAM computational fluid dynamics software (CFD). Realism is further increased
e found on the internet to a set of LTL formulae, grounding high-level cooking actions to a set of primitive actions that are execu
bly by decomposing complex tasks into simpler subtasks, generating robot control code, executing the code in a simulated wor
th synchronous motion planning.
ted into the software pipeline. This paper presents a new, lightweight approach to address this challenge and improve outdoor
n by a factor of 2, outperforming other approaches and benchmarks. The enhanced images improve the reliability and detail of
proach learns from MPC predictions, adapting to unmodeled dynamics and disturbances, resulting in a resilient control framew
ms, with 4 and 8 maneuverable units (MUs), respectively. Reinforcement learning (policy gradient) is used to design the centra
ound training of the NN using only synthetic data. The experiments on real hardware-in-the-loop trajectory images show that O
nction that offers the agent clear and insightful feedback. Our case study examines a pre-grasping task where a Robotiq 3F gri
narios. Event sensors have made significant advancements in hardware and software in recent years. Moreover, the characteri
n gradients at the individual cell level, thereby avoiding the computational overhead typically associated with such analyses. It le
using real nadir-pointing imagery and find that 85% of simulated satellites are localized to under 5 km within 6 hours (4 orbits). T
ntroller design and performance is demonstrated by simulating a 3D SMS controlled by model-based and robust controllers. In
 along the Z-direction. Furthermore, for four-armed space robots, a hierarchical planning method is proposed, which innovates

control schemes. This paper outlines the design and manufacture of a soft, inflatable, spherical shell designed for a robot drive
eriments presented include undocking at various inclinations with different loads expected for the application scenario.The max
control method of a robotic arm on a free-flying robot with a gecko adhesive gripper. The gripper utilizes a single-motor contro
el differential driven robot and spherical robot driven by equivalent pendulum.VWDER is equipped with six retractable shell-sha
herefore, the aim of this work is to study and compare robust controllers and examine their applicability to SMS. An LPV plus H

lacement of surface points surrounding the manipulators. Then we develop a 6-degree-of-freedom velocity controller to manipu
e clinically deployed robotic hardware, show improvements in control, force estimation and reflection. The significance is that w
89.5% accuracy for gesture prediction through effective fusion of kinematic features with spatial and contextual video features.
ethod is unable to achieve success without the help of expert demonstrations. Comparison results show that our method achiev
s reasoning about randomized dynamics parameters. We successfully transfer our agents from simulation to real and show rob
6D pose datasets. However, this strategy does not translate well to surgical domains as commercial graphics software have lim
rior knowledge in successfully accomplishing tasks. Due to the intensive nature of simulated data collection, previous works hav
oal-conditioned decision transformer to achieve sequential representations with goal-aware future indicators in order to enhance
t enables the bronchoscope to move in a follow-the-leader motion mode or fixed-angle motion mode, navigating through narrow
red environment as well as an unstructured environment and measure to demonstrate the efficacy of our methods.
drift, we then propose an adaptive-template matching method, which updates the tracked template based on the reliability of th
an online, adaptive parameter tuning approach for simulation optimization that (1) bridges the real-to-sim gap between a physi
nerated to balance between exploration of unsensed regions and refining the existing probabilistic understanding. We demonstr

evels and for separate robotic arms using only kinematic data. The models are evaluated using the LOUO (Leave-One-User-Ou
We also show that applying domain randomization techniques make the policy more robust against changes in geometrical and
cal skills. ORBIT-Surgical also facilitates realistic synthetic data generation for active perception tasks. We demonstrate ORBIT-
uced to overcome the ``big-to-small'' problem in typical robotic scenarios. Finally, an iterative reweighted least squares is applie
remained a challenge due to differences between the simulation and hardware domains. With simulation-only data, in this work
er) and embed the constraints of concentricity. We use an energy method to determine robot shape as a function of actuation a
n process includes a reproducible calibration process and a novel tip curvature compensation method. We validate our approac
netic steering. In addition, a cylindrical array of magnetic sensors is used to measure the field from the catheter tip to enable re
enchtop experiments and an artificial phantom quantify the performance of the developed robot controller and navigator. Valves
ombining initial vessel structure-based 3D US – 3D CT registration with intraoperative intensity-based 2D US - 3D US registra
e independent key-value pairs (delta s, delta d), where delta d is the partial derivative of the distance measurements along a sp
od by introducing an obstacle into the environment. The strategies rely on local interactions among UAVs, and the sensing of t
ed on the synchronisation over a network connection of multiple ARK systems, whether real or simulated, serving a twofold pur
ramming local interactions, through physical contact or communication, among Morphobots. We characterized the mobility of M
d for manufacturing the continuous docks in a variety of geometries.
e non-reciprocal coupling to elicit collective locomotion toward or away from target sites, and we use the control barrier function
p. In our study, the group of shepherds is the swarm that is automatically designed, and the sheep are pre-programmed robots
ebo Sim. The plugin transmits simulated sensor information to the firmware along with a socket link interface to run external sc
n policies for moving towards semantic targets at unknown locations. Experiments on both our proposed Habitat transport task
 ategy. We validate COAST on three challenging TAMP domains and demonstrate that our method outperforms baselines in te
ual-arm tasks. We also present how this CGA-CDTS can be seamlessly integrated with an optimal control framework in geome
on to state-of-the-art techniques. We validate our approach through simulation and demonstrate its reactive capability to cope w
ultaneous trolley collection, we propose an efficient exploration-based object search method. First, a traveling salesman problem
resent an infinite choice of random initializations of computational sub-problems. Decision making in such trees means to decid
s of continuity in sampling to enhance system efficiency. Firstly, we decouple and comprehensively evaluate the impact of both
ient geometric constraints. This reachability-aware heuristic efficiently prunes infeasible sequences of actions in the continuous
ach on various manually designed and automatically generated scenes, demonstrating the benefits of auxiliary subproblems in

ct while interacting with the environment. Specifically, HAPFI integrates multiple modalities, including historical RGB observatio
e items. Case study also reveals that our approach performs more effectively on the overall task than the Belief-Desire-Intentio
elying on prevalent simplifications that exist in the literature. The evaluation results demonstrate the planning efficiency of the p
bustness. We demonstrate our framework on several case studies, including an aircraft turnaround task involving three Jackal r

t configuration towards the next sub-goal of the task plan. When the RRT is extended, the probabilistic model is updated alongs
ation through communication with other robots and considering its remaining budget. We evaluate our proposed method for var
nd physical experiments. The results have confirmed that our algorithm has a lower energy consumption compared to a two-sta
ch enable researchers to focus on individual sub-tasks of the assembly challenge if desired. We provide a full digital twin as we
so that robot integrators can perform well-reasoned, time-efficient risk assessments. In this paper, we provide the initial datase
ion, and use it to guide the planning and strategy grounding processes. Together, our observations yield an affordance-directed
in a hierarchical planning pipeline to generate executable task plans. We demonstrate our entire approach on long-horizon task
h-cost nodes to improve planning correctness and improve execution efficiency. Our evaluation results show its superior perfor
s are iteratively optimized using a neighborhood search. While aiming to find a near-optimal solution rather than an optimal one
d growth step, which leads to the difficulty of growing branches in narrow environments, making the efficiency and correctness o
rable haptic interface to cope with the limitations given by the previous work. Leveraging the additional haptic feedback modalit
rve as teleoperation transparency measures and demonstration quality indicators under varying communication delays. We val
ved, exploiting the feedback from the FDA actuators to provide the kinesthetic feedback to the LDA, using the pulse width mod
ctions and resource limitations are the main bottlenecks for teleoperation. In this study, we propose an intelligent mode-switchin
igital Twin-driven Mixed Reality (DTMR) manipulation with haptic rendering. This system comprises a commercial remote contr
hen the geometry and impedance parameters of the remote environment are known. In addition, a point cloud-based force rend
plemented robot avatar compared to stationary and teleoperated camera pose conditions. Our results demonstrate that the pro
put compared to requirements. This paper proposes a novel haptic command interface (POstick) specified to lump iron remova
kills, during both design and execution. We demonstrate that the SCTs so acquired can be successfully used to pick up an obje
acy and user-friendliness of TELESIM, we conducted a user study with 37 participants. The study involved a simple pick-and-pla

roposed method outputs a probability distribution over the possible placement candidates, considering the robustness and reas
The experimental evaluation confirmed the effectiveness of our control algorithm, demonstrating an average reduction of 4.67%
s frustration in a subjective load index questionnaire. Our findings strongly suggest that iRoCo is a promising new approach for
an subjects study (N=19), we find that operators using this framework complete a range of tasks significantly more quickly and r
ction windows of input data. An analysis of the predictive power with various inputs demonstrates the predominance of the deep
eedback for wheeled humanoid control. Finally, the system's effectiveness and mappings for DMM are validated through locom
k, where we introduced a deep learning model and an algorithm to predict as early as possible the user's desired motion. Movin
pose a novel dynamic aggregation (DA) method to replace previous static aggregation like mean or max pooling for pre-trained
n (VGI) to rectify the imbalance supervision signals by inserting prior rare objects collected from the wild while avoiding introdu
tion spaces across domains with unpaired data. We propose effect cycle-consistency, which aligns the effects of transitions acr
D point cloud understanding. Comprehensive experiments are conducted to demonstrate the superior parameter and data effici
challenging. In this paper, we propose a Multi-space Alignment Teacher-Student (MATS) framework to ease the domain shift a
to 6-DoF grasps and novel objects both in simulation and in real-world settings, without the need for additional data. Suppleme
gressively learned via a set of simulations in a coarse-to-fine manner. It is expected that most RL trials are performed in coarse
ysical environment (target domain) and a simulated environment (source domain) as a Gaussian perturbation and utilize Kalma
ch as attachable IMUs and motion capture systems, to practice dance in a virtual reality (VR) environment and to control variou
experiments were conducted by mounting the DARFT-1 on the humanoid robot DRC-HUBO+. Through the experiments, the D
capabilities. Compared to robots with fixed configuration, there are superior advantages of this robot in terms of adaptability, rob
propose a pose control method for the suction chamber based on an adsorption performance metric which depicts the stretch o
ector of the manipulator and the grasped plate is known, we can also obtain the motion of the manipulator for inserting the plat
robot to efficiently use customized tools for a wide range of cooking tasks. Additionally, we designed a modular cooking cell co
Scene, which requires only one Standalone Free-moving Device, is Scene-invariant, and can sense up to a 180-degree field of
ontrol when needed. Through Co-RIFT, a global volunteer force aids experts in addressing environmental issues by guiding and
s users to manipulate objects via hand gestures. Its overarching goal is to demonstrate that improved human-machine commun
mages while having measurable physical sizes in meters. The proposed model enables fine-grained human motion analysis and
tested on a collaborative assembly task, which included fetching and storing parts. The Neural Networks were trained on a dat

performs non-monotonic logical reasoning with prior commonsense domain knowledge and with models that are learned and re
k, and the LiDAR is not portable for human localization. In this paper, we present an inertial-only IPS that uses an external port
ulator, we can efficiently improve the control system.
algorithm with the detection, we improve detection accuracy and estimate the target's state, velocity, and pose. The AUV can tra

al robot, Pepper, guiding persons with dementia through a music intervention session. This includes three activities: a Keep-with
Moreover, both the patient and operators will be exposed under the radiative fluoroscopy during the procedure. The ARTERY p
some of its solutions for the navigation problem (Move Base, TEB, RTAB-Map, Robot Localization) and manipulation (MoveIt).
in conjunction with Reinforcement Learning (RL) in robot training. Through our approach, these extracted features enable robo
exist in ground-foot static friction, requiring smaller steps to mitigate foot slippage. The second approach involves adjusting leg
ngaging in negotiation, and executing transactions. The FiReSpARX project is dedicated to designing market mechanisms, ince
e grip. A full shared control framework was developed by integrating the automatic grasping control with an EMG-based artificia
sical robot checks the workspace for a cube with a wrist camera. The Unity environment functions under physics that are mana
unconventional mechanical design, with a highly extendable tape spring limb that allows it to place its magnetic gripper at distan
ough a variety of multi-modal input devices using the Avatar concept via GARMI. The attached video illustrates the proposed co
ondestructive sampling tasks, including fragile porcelain retrieving and precision instrument manipulation at a depth of more tha
e, sand, grass, and rocky surfaces, and climbs inclines of up to 28 degrees. The robot's innate compliance and shape adaptabi
ing, presented at ICRA2011. This operation was made possible by the structure of multiple flat tubes and a mechanism that au
ascent to descent in 20s. At 2V voltage, within 220 seconds, four buoyancy regulators together generated 5g of buoyancy, with
 ways with low friction, good back-drivability, which makes the mechanical device more lightweight and compact and have better
reduced performance. In this study, we investigate how the stiffness parameter in impedance control changes muscle activation
y mixed reality to wearable robotic limbs, demonstrating initial potential for improving wearer efficiency.
R (ENS-LSR) on simulated box stacking and grape harvesting tasks as well as on a real-world robotic T-shirt folding experiment
eaching 2 m/s.
ly available new source of surface normals: the specularities. Our first technical contribution is a method which detects specula
or maybe when in motion (e.g., a drone following the robot). The scenario itself presents a complex challenge: single-image re
i-object tracking benchmarks.
pipeline for precision agriculture based only on motion information that allows for accurate and fast propagation of tracks betwe
upled estimator. Our approach uses visual object detection algorithm to provide UWB anchor initial position and refine it in the f
ement, we proceed to update the rotation estimation by performing IMU integration with gyroscope bias removed from gyroscop
mprovement over the state of the art, achieving up to a 177× speed-up in surface reconstruction, and up to a 31× improvem
module and continues driving an overridden policy optimization. XGX, with effective decomposition and novel guidance, improv
lly stable and robust while considering the uncertainties of the closed-loop system. When deployed on a Clearpath Jackal robo
all on different benchmark settings for the retrieval task. For object navigation, we show the benefit of our approach in simulation
detection and classification in camera images. The synthetic data generation is employed to augment the training set and boos
dynamic horizon planning through the introduction of variable speeds. Further enhancing our methodology, we incorporate a n
l that maps the visual input from various pixel spaces to the local occupancy grid space. To tackle the error and misalignment o
nstrained by the number of IDs or recognition speed when used with standard RGB cameras. In addition, recognizing object po
ease computational costs without significantly impacting estimation accuracy. The adaptable recurrent encoder produces multi-
To address this insight gap, we propose an Object-Attention VLN (OAVLN) model that helps the agent focus on relevant objects
aging large-scale datasets from non-robotics sources, increasing training data availability. Bayesian filtering, enabled by place r
ur approach introduces continuous modeling of the hierarchical scene architecture and leverages visual-language pre-training to
ving the first phase of the Open-Vocabulary Mobile Manipulation (OVMM) 2023 challenge, demonstrate that an agent equipped
ixels as the goal specification, our method becomes a versatile navigation policy towards all different kinds of objects. Besides,
seudo-labels against domain shift, improving the performance and reliability of our Test-Time Adaptation approach. Comprehen
em is trained using the Imperative Learning paradigm, for which the network weights are optimized end-to-end based on the pla
cy on IR instead of raw data, the second phase is domain-invariant --- the navigation policy does not need to be retrained if the
m established terrain preferences within the inertial-proprioceptive-tactile domain. Leveraging this insight, we introduce Preferen
ation between different size patches and ambiguity, and identifies non-uniform deblurring spatial features, facilitating the recons
date keypoints as vertices are divided into sub-graphs according to the feature distributions on the graph edges. Finally, each s
enabling direct 3D supervision from 2D semantics and depth labels. Additionally, we introduce an Auxiliary Ray method to tack
his way, our network is a gradient-free zeroth-order optimizer designed for V2D and can be applied to both supervised and self
rrespondence robust to the dynamic changes of the scenes. Trained on an in-the-wild video dataset, Doduo illustrates superior
ning to downstream robotic vision tasks and yields superior performance. We introduce the RoboLLM framework, equipped wit
oth modalities. Through extensive experiments, we demonstrate that our method significantly surpasses previous state-of-the-a
ment the training process, and (4) an optimization-based post-processing algorithm in test time, fusing the results from visual o
pectives, considering the specific characteristics of DVS. A series of carefully designed losses helps enhance the quality of gen
that accumulates information across multiple frames of a video sequence. Our MOTPose method takes a sequence of images
onal Adapters, which not only significantly improves accuracy and robustness but also enables impressive zero-shot generaliza
dditionally, we address the issue of over-confidence in the closed-set by refining model calibration, avoiding misclassification of
method leverages a multi-stage training pipeline, designed to efficiently transfer synthetic performance to the real-world domain.
al-resolution motion information from event streams, namely FE-DeTr. The network leverages a temporal response consistency
ry differential equation. The resulting framework offers control over the generated videos' dynamics, content, and frame rate, a
dalities. Our network comprises a generator and two discriminators. The generator aims to produce fusion images with salient o
e image appearance of panoramic cameras. With only a few 2D boxes annotation as guidance during the training phase, our n
synthetic and real images. For all datasets, our method outperforms state-of-the-art methods, in particular for depth prediction ta
s to fine-tune these advanced models, like Stable Diffusion, for specific conditional generation tasks emerges as a promising av
tational overheads. Through evaluations on real-world datasets, such as KITTI-360, Argoverse2, and our private dataset, we de
es but also conduces to expressions of local geometries. More importantly, we establish the first large dataset for event-based 3
ble, its limited depth sensing capability hampers comprehensive environmental perception. Addressing these limitations, this stu
-specific image outpainting model. To assess the performance of NEO, we conduct comprehensive evaluations on three photor
hness of the pose. We then design a lightweight neural network to regress the pose by developing a differential pose layer. Fina
jectives, we introduce Crossway Diffusion, a simple yet effective way to enhance diffusion-based visuomotor policy learning via
d sub-symbolic task representations. Our bimanual task representation is object-centric, embodiment-independent, and viewpo
cloud to aid the model in better learning the geometric structure of the lunar environment. Through experimentation, we have d
ws (BoV), a fully appearance-based model used to assign utility to the captured views for both offline dataset refinement and on
stly, extracting all lines from the image; secondly, identifying corners; and lastly, searching for quadrilateral candidate regions u
pose a robust collaborative perception system that operates independently of external localization and clock devices. The key m
s with waterdrops removed. Extensive experimental results on both synthetic and real datasets show that our method is able to

view, thus allowing for real-time prediction. Moreover, to improve the accuracy of 2D-3D correspondences, we propose a 3D co
mation of the skeleton. Then, we select an optimal group constructed considering the occluded body parts. Each group comprise
ts with the temporal order of input frames to eliminate errors induced by object motion. Meanwhile, we enhance the original dis
nce visibility in a single real-world hazy RGB image. First, both Low-level and deep features are extracted, and then a pre-traine
ed data is then proposed to constrain the model on not only semantic prediction but also the BEV feature. Furthermore, we prop
e Language Models (LLMs) and the cross-modality capabilities of vision-language models (VLMs). To the best of our knowledg
ios characterized by complex lighting conditions and high-speed motion. Motivated by these observations, a natural recourse is
es. This work opens avenues for robotic applications where low-light conditions often pose difficulties such as disaster recover
o target spike data. To this end, we propose a cross-modality cross-domain (BiCross) framework to realize unsupervised spike
generated for each active pillar. PPG generates the synthetic points for each active pillar based on the information provided by
grammetry system requires professional skills and these would cost tremendous labour for rapid configuration each time. In this
ormation. In the MMA-Net, we first feed spine X-ray images into the segmentation network to produce multiple morphological inf
e challenges, using path tracing and physically-based, data-driven models, is evaluated on an existing large real-world dataset.
he target pose using monocular RGB images. Since contours play an important role in both keypoints prediction and pose estim
grees of freedom and can be deformed in 3D space by inflating or deflating each chamber. Experiments showed that the soft ac
a robot arm UR10, 2-finger gripper, depth camera L515, and 10 objects arranged randomly in the bin to evaluate the efficiency
dditional adaptation. To enable this, we introduce a method for generating a comprehensive synthetic dataset using a physics s
scale synthetic datasets showing warehouse traffic based on real vehicle models and geometry.
ption. Following recent advancements in group-equivariant learning, RIDE models rotation equivariance implicitly within its arch
s from diverse surgical operations collected from multiple departments or clinical centers, and ii) severe data imbalance arising
Segmentation (MSPR-DKS),' designed to address these challenges and provide input for the precise control of robots. Moreove
 grasps a static target in 3D space, grasps a pitcher handle, and picks and places a box by determining the approach angle usin
od, to tackle these problems, a camera is used as an external sensor. First, contact between the hand and the object is detecte
monstration for a specific task. This enables us to compute the shortest path through the graph. Ultimately, we show that recom
this work, we propose an acceleration-based controller for the position-based visual servoing framework, which models the erro
surveillance and autonomous navigation.
known and can drive the object toward desired shapes through large deformations. The inputs to our approach are the point clo
d undermining projections for darkness clues. It then injects these learned visual prompts into a daytime tracker with fixed para
r visual tracking. Our system operates efficiently on limited hardware, relying solely on an onboard camera and an inertial meas
roneMOT, which firstly proposes a Dual-Domain integrated Attention (DIA) module that considers the fast movements of drone
le measurements. Through several interactive scenarios, we demonstrate that the proposed OPF approach provides robust tra
D videos. Given a monocular video input, our method predicts object and part mask tracks with associated language description
xt information. Then, cross-attention is used to inject the template information into the search area. This strategy allows the fea
s supervised weights, most existing self-supervision techniques actually make performance worse instead of better, which sugg
pproach is formulated to solve the multidimensional assignment problem with lifted graph edges introduced to account for misse
uncertainty around the boxes, no work in 2D MOT for autonomous driving has studied whether these estimates are meaningful
a heavy and wide bobbin using bimanual grasping and bipedal locomotion at a time. This experimental use-case is found in som
etween the template and search region for better feature fushion and cutting redundant computation, and 2) pruning encoder la
rrection method to maintain the constructed parallel matching relationship and implement holistic pose correction for keyframes
dynamics factor into VIO and to optimize the model parameters online together with the VIO state variables. We validate our m
This approach enables the tracking phase to utilize the full image and inertial information. During intense motion, the inverse dis
the proposed SRI-GVINS gains significant numerical stability and computational efficiency over the start-of-the-art methods. A
nce as output. The final dense depth is the linear combination of the multiple depth bases. Thanks to the multi-basis depth repr
ning fast at 300 FPS with predictions from 400 Hz IMU measurements to provide accurate and smooth trajectories.
ual from the standard camera, and the residual from IMU pre-integration using a keyframe-based graph optimization framework
eedup of up to 4.23x in the linearization stage and an overall enhancement of up to 2.08x in the overall optimization process.
multi-camera systems, and introduce implementation details that contribute to the performance of our system in challenging sce
s, we diligently collected and annotated a new dataset (MAT) containing 1000 images. In planning, we integrated the methods
body weight with one motor. Owing to these unique characteristics, the VGFL can realize the strategy without multiple actuator
nce is significantly enhanced with a supernumerary tail and that there is positive impact on muscle fatigue.
leton transparency in the meantime. In active mode, an air pump inflates the pneumatic artificial muscle (PAM), which stores th

ve learning. Specifically, the robot first learns from a group of healthy subjects to generate a reference motion trajectory via prob
MU sensors to detect lifting posture and correct it in real time. The exoskeleton’s effectiveness was tested through lifting exp
rce vector, enabling continuous and user-independent estimation of GRF. The sensor module is integrated with machine learni
-exoskeleton locomotion interaction experience transfer (LIET) framework, which could speed up the exploration of human-pre
d Coronal planes. Previous knee exoskeletons primarily focus on aligning the joint in the Sagittal plane, neglecting alignment in
with five healthy subjects and compared to state-of-the-art controllers.
ed in this type of approach. The effectiveness of the proposed control strategy is demonstrated through experiments involving f

her than itself and then smoothly transition back to the wheeled mode after leaping. The dynamic driving-to-leaping maneuver e
he contact status converges toward the correct result. This paper tests the algorithm through simulations and validates the meth
ntrol signals to stabilize it on parabolic terrain. The stabilization problem was solved as a quadratic program in < 3 ms which sho
rametrization, the trajectories are rendered inherently contact constraint-consistent. We apply our method to the specific case o
and their performance is compared over a range of inputs. The CPG-based controller is used to drive the developed robot prot
identifiable among the twelve-dimensional WFs. This indeterminacy turns out not to affect the rollover prediction, yet does so fo
Fs), so that the over constraint is avoided. A transition mode based on oblique vehicle motions is used to bridge the transformat
requirement that the vehicle's velocity can cross zero and at the same time have a global attraction region. The design of the c
ously recomputing the feet position while the robot is falling. In this way, the quadruped is ready to attain a successful landing in
se optimization aligns the robot's trajectory with kinematic constraints. The resultant robot reference trajectory is executed via a
very in the wild is difficult. We then present a set of benchmark policies trained in simulation using deep reinforcement learning
ficients with the environment by conducting evaluations using both a simulator and an actual humanoid.
egrated into the proposed whole-body standing-up control to create the trajectories of the hip and the upper body joints online.
ptimization problems. Additionally, the 0-level set of the robot corresponds to the robot surface, which can be seamlessly integra

cilitates a characterization of the force-tracking performance of learned policies in simulation and the real world, indicating their
erent shared-control methods for obstacle avoidance in diverse environments. A time-derivative Sigmoid function (TDSF) is imp
 design an efficient super-resolution based end-to-end transformer baseline called DA-TextSpotter which achieves comparable
nsfer knowledge from source to target, we propose prototype-calibrated self-training. By exploiting class-wise prototypes in the s
ir appearance and disappear from the camera's field of view to test the module's ability to handle challenging cases of occlusio
sistency. This score takes into account the subtle geometric differences between structurally similar point clouds and therefore c
GFA can leverage existing large VL models, enhancing their zero-shot capabilities of new modalities due to their natural pairings
atures. In PS-VLADs, we replace half the parameters used in the original NetVLAD with trainable scalars, which further reduce
tes the challenges in estimating bounding boxes under non-uniform aspect ratios. We have developed two tracker versions, na
re-trained Vision Transformers (ViT), are used for matching rendered templates against query images of objects and for establ
uplex attention function, flow-guided information, RGB information, and spatial support tokens. The proposed duplex attention m
urate torque controller with inertial compensation is developed for the VTSEA via the singular perturbation theory, and its stabili
to 43% and 44% for joint 2, and by 15% and 21% for joint 3 compared to the configuration without springs.

anscendental equation of the relationship between the actuator stiffness and the rotation angle of the stiffness motor. SVSA-III h
a hybrid system model, as well as the mechatronic implementation of the actuator. We corroborate the feasibility of the concep
lexity and cost alternative to classical pinwheel designs, while offering similar performance.
esign. The verifications for static characteristics of output torque, as well as the dynamic characteristics of power density, mech

force-deflection behavior, but most typical variable stiffness springs require energy to amplify force similar to motors. In this pa
nt and the tether. Then, the letter presents a trajectory planner based on non-linear least squares. The optimizer takes into acco
odeling method is employed. In addition, a new LOS judgment model that constructs communication relay LOS links between co
online planner to estimate energy consumption and perception quality, reducing UAV battery usage and localization errors. Sim
y of trajectories derived from the Clohessy–Wiltshire equations. This advancement opens up avenues for replicating close-pro
(more data increases the regression performance but requires more memory and longer computation); textit{ii}) methodologies
e for online motion planning, is obtained. The algorithm is demonstrated in extensive flight experiments encompassing six aero
mality in real applications. The lightweight nature of EVOLVER enables online computation. The framework is thoroughly evalua

odel-free control policy. In-field tests are based on the Crazyflie 2.1, extended by a custom multi-zone ToF deck, featuring a tot
m can perceive its environment and adapt painting policies to natural painting media. Our results highlight the effectiveness of

s can be used to actuate an imprecise robotic arm to play sequences of notes with varying loudness on a Chinese Guzheng.
lenges of programming embodied agents in real-world environments. However, teaching with real robots remains challenging d
del online improves 3D trajectory tracking by nearly 36% on average (and vertical tracking by 56%).
ticularly when latent disturbances continuously vary. This paper presents the Hierarchical Meta-learning-based Adaptive Contro
, we introduce a wide-angle camera-based method to generate a panoramic probability map (PPM) for estimating high-probabi
nd, we proposed the distillation strategy to distill the knowledge to segment the free space of the walkway for robot navigation w
on: How accurately can ViLaIn and the symbolic planner generate valid robot plans? To evaluate ViLaIn, we introduce a novel
omplexity with respect to the MHE horizon. As a case study, we evaluate the proposed trust region NeuroMHE on real quadrot
network methods since interpretability is particularly important in industry to ensure stability. Experiments on texture surface sa
ting framework that prompts LLMs with kinematic knowledge of objects to generate low-level motion trajectory waypoints, supp
t. During preprocessing, an LLM translator is employed to convert natural language input into a Planning Domain Definition Lan
 r sensor drift compensation for robustness evaluation. Additionally, for efficiency evaluation, we compare the energy consumpti
model interpretability and performance in OOD scenarios by disentangling the latent space of edge features, thereby strengthen
k reveal that DeFlow achieves state-of-the-art results on large-scale point cloud data, demonstrating that our network has better
diGraph notably enhances generalization and sample efficiency.
space and offer an ethical framework for virtue ethics inspired by Aristotle's Doctrine of the Mean. Finally, we demonstrate that
bot to perform the specified actions in a sequence. To improve our model, we incorporate codebook-based vector quantization,
eaningful insights into the expert's locomotion strategies. Furthermore, we empirically demonstrate that training a bipedal locom
ork while maintaining a low false negative rate in practice (whenever there is a distribution shift in our experiments, our method
ny loss in accuracy. Moreover, using our proposed framework, we contribute an adaptive method to find a maximal feasible retu
active perception with a physical quadruped. Empirical results demonstrate that our approach outperforms five baselines. It is a
he effectiveness of our approach is underscored through a high-dimensional synthetic state estimation task and a 6D pose estim
the attributes of model application scenarios, we can design more effective test suites. Additionally, we establish a comprehen
udged the system efficient, useful, and intuitive, and showed a positive inclination to re-use the robot in the future. Such outcom
ous, and result in a lot of food waste. Therefore, we proposed SliceIt!, a framework for safely and efficiently learning robot food-
and depicted within a scene graph, guided by either commonsense or user-defined criteria. Then, this scene graph subsequentl
we compare the task success rate and token efficiency of four multi-agent communication frameworks (centralized, decentralize
nally, we present a feature reuse mechanism to integrate the vision-language features from off-the-shelf pre-trained MLLM into
keeper rod as an initial setup with the intention to be extended to the full game as soon as possible. Our experiments reveal th
anguage model aligned with the policy networks, which allow the robot to recognize user's intention or perform reasoning tasks
context-length limits in LLMs. However, it is not clear how to document robotics tasks and providing full robot APIs presents a c
from discrepancies between simulation and reality. In an experimental case study we apply the framework to train and deploy
at effective strategies are highly dependent on the environment as well as the global and local planners. Based on this insight, w
milar scenes better. In settings such as office buildings, hotels, and warehouses, this approach helps to improve the robustness
ods on the task of upright pose prediction for chairs. The developed system can be easily transferred to a wide variety of applic
iffusion models and querying LLMs to determine plausible physics parameters. Given URDF files of generated and human-dev
ithfulness of LLMs and make the planner more reliable. We conducted experiments in VirtualHome household tasks. Results sh
patch-aligned feature descriptors to expand foundational model capabilities to create an end-to-end multimodal driving model, d
ng LLMs to directly plan task sub-goals, we instead perform few-shot translation from natural language task descriptions to an i
mics Spot robot in eight novel real-world environments (one apartment, one lab, two microkitchens, two lounges, one office spa
getting mechanisms that are periodically evaluated for offline data deletion inside the LTM. We compare different mechanisms,
es the collaboration combining a robot planning algorithm for obstacle avoidance with a variable admittance control. This enable
dhere to regulatory constraints amidst complex flight conditions, while significantly reducing the burden of UAV operators.
x planning process is decoupled into two simpler sub-problems. As a result, the proposed planner can be solved as two small q
oning uncertainty in the presence of noise.

t attitude similar to that of a kite to keep position stability and increase damping. Meanwhile, the winch only needs to recover th
utonomous drone navigation. Also, a generic framework for creating synthetic fisheye images is provided. Furthermore, we prop
he regressing approach to distance estimation is more robust to varying relative pose of the targets and that it is suitable to be u
mpling toward the unexplored region next to the free space, defined as the heuristic frontier regions. The heuristic frontier regio

ding neural networks with Morse Graphs. MORALS shows promising predictive capabilities in estimating attractors and their Ro
 ctory tracking and dynamic obstacle avoidance.
movable microfluidic chip. The method has significant advantages in biomedicine engineering and micro-nano-manipulation.
ation of regressor channels at different instants is exploited more effectively and efficiently to achieve fast and accurate parame
autonomous driving in challenging real-world scenes with high fidelity and low risk.
challenge and explores a wide range of safety-critical scenarios. Our evaluation demonstrates that the proposed method gene
Convolutional Neural Network. The network takes a single RGB image along with cars' velocities and estimates the target vehi
nd vehicles within high-risk locations.
ency. We also introduce the first dataset for realism alignment in traffic modeling to support such research. Our framework, nam
rial factory environments worldwide. Aiming at deploying the LLM agents in a collaborative environment, we propose a queryab
interaction score for each CCV. These scores prioritize vehicles that are most relevant to the current driving scenario, offering v
ms the baseline method in closed-loop simulations, including safety metrics. Qualitative results illustrate the superiority of POP
he interacting entity pairs through future affinity learning and top-k filtering strategy. Experiments show that our future interactio
s, and ego vehicle behaviors. Thus, it strongly supports scenario-based testing by forcing pedestrians to produce certain situati
g a condition-action-control sub-flow. We embed a novel grammar structure called stringBT that encodes knowledge, enabling
hen applied to identify the order of task execution for an agent. The proposed DMCHBA is proven to produce conflict-free assig
ch exhibits competitive performance in path following tasks and is 3 times faster on average for the same prediction length.
robot navigation and present an analysis of various prompting strategies affecting the model output. Finally, we showcase the
autoencoder, a Wasserstein autoencoder, and a spherical autoencoder are evaluated against each other. The latter two varian
complex robotic tasks, our method outputs variable BTs that can add and execute new actions according to environmental cha
effectiveness of prediction, we evaluate our proposed method on two well-established trajectory prediction datasets: the ETH/U
driving scenarios, answering questions, and decision-making. Our findings highlight the potential of LLM-based driving action g
c participants. Moreover, we propose the type-aware two-level hypergraph message passing module (TTHMS) with learnable h
al outlier detection. Based on the filtered data, radar motion state estimation is performed by use of curve fitting methods. To fus
entities. The system utilizes our architecture, TrajNet to interpret thermal visual inputs to produce desired trajectories and emplo
results demonstrate that a well-designed, purely imitation-based planner can achieve highly competitive performance compare
s framework is highly automatic and scalable, which can greatly speed up HD map production and decrease the cost. The prop
clusions that might intersect with the AV's planned path. The framework estimates occupancy probabilities and likely trajectories
meras or LiDAR sensors. Then, online visual localization is performed through data association of semantic features and map o
er-representing unlikely predictions. DiPA addresses these limitations by using a Gaussian-Mixture-Model to encode the full dis
Moreover, to generate diverse driving maneuvers, we introduce risk-aware contingency planning, a bi-level optimization algorith
T-GCN) is proposed to match the recovered cost functions with the traffic scenarios, making the lane-changing trajectory planni
show that conventional wheeled robots have previously unrealized potential of moving through vertically challenging terrain. W
on is unnecessarily extreme: using a large-scale real world social navigation dataset, SCAND, we find that geometric systems c
ex environments. Simulations and experiments are performed and verify the performances of our method better than the state-
grounding these queries to the visual context embedded into the language-enhanced BEV map. To enable further research in
on from multiple camera views and LiDAR sensors, enabling comprehensive scene understanding for multitask learning. Furthe
ur novel localization recovery method generates the potential future trajectories by learning from the navigation affordance, and
en outperforms prior state of the art on the Waymo Open Motion Dataset.
query ray, we generate the 3D target point embeddings on the query ray using on-the-ray positional embedding and local featu
tperforms baselines in terms of robustness and precision. We also implement our method on a real UGV and show that it perfo
tracking dataset captured from a mobile robot adapted for the prediction task. Furthermore, we identify new agents with limited
e training loop of the RL algorithms, providing diverse driving guidance whenever the intervention is required. Additionally, to be
in focus from per-step actions to temporally extended skills, thereby enabling long-term reasoning into the future. Extensive res
onal traversability network, termed MultiTCNN, which is capable of directly generating MTraMap. We evaluated both networks o
deling, employing LiDAR for object modeling and cost-effective sensor cameras for object tracking through image-point-cloud re
, resolves the inherent conflicts between gas exploration and collision avoidance. Our approach aggregates the local observatio
etter stability and generalizability. Experimental results demonstrate that our method can achieve robust navigation performanc
oy large continually learning robot fleets in a cost-effective and scalable manner. All materials can be found at https://www.open
(WOMD) with large-scale, high-quality, diverse LiDAR data for the motion forecasting task. The new augmented dataset WOM
data points. As a result, it can be shown that the absolute position accuracy of the industrial robot can be improved by 96 % wi
t policy. Through extensive simulated and real experiments, we show significantly better categorical generalization compared to
enables our agent to exhibit a diverse repertoire of skills in novel situations specified using task commands. Using merely 7500
ding complex multi-object relations, and readily applicable to both tabletop and 6-DoF rearrangement tasks.
n. Our insights are to utilize calibrated offline reinforcement learning techniques to ensure efficient online fine-tuning of a pre-tra
sk of motion forecasting using real world data from autonomous driving systems. We develop ensemble models that are very c
rrors detected by the software analysis technique, Phys. We identify three types of physical unit mismatches and present a tax
ot can achieve self-takeoff with no potential energy cost and maintain high maneuverability (The video shows a rapid climb imm
with the help of origami flippers. The origami flipper also shows good adaptability when in contact with the external environment
ate. To validate this simplified model, a simple control is presented for continuous circular trajectory tracking inside an indoor flig
built a multi-joint robotic fish, and developed its dynamic model and control method to perform all the maneuvers under the gui
significantly reduced step length. At low water content, the increased mud suction force caused appendage entrapment, result
nt the RoboTwin platform designed to aid such data collection: it allows us to replay the dynamic movements and body posture
gh updating rate of 100Hz. Finally, we applied the method to swimming direction feedback control for evaluation and compariso
dding shape memory alloy wire into a silicone matrix. TSCP can adjust the stiffness in real time through current and increase th
arm robots in dynamic environments. Similar to fish escape maneuvers, simulations and real-robot experiments show that the s
with respect to robustness, energy-efficiency, and tracking performance in sim-to-real navigation tasks? 3) How do animals ma

the robot on each face of the shell before releasing the robot – and an intentional fall – walking the robot off a ledge at var
t level, and added to Cartesian PD impedance control and Virtual Model Control to stabilize the jumping motions. After optimiza
toppling. While the yaw rate of the robot was actually higher without the tail, the tail substantially reduced unwanted roll, confirm
.0 m/s, respectively. The results also indicate that reinforcement learning is a feasible way to investigate how the spine should
ybrid motion, offering optimized solutions for intricate robot locomotion. Furthermore, we employ parallel underactuated MPC co
ntegrated with the power source, wireless communication module, and control module, the untethered operation is also present
owing depths. We hope these results will serve as a proof of concept for how optimization can be used to unlock the secrets of
total mass of 2.1 g, which is an eight-fold reduction compared to the lightest battery-powered tailless flapping-wing MAV curren
d a distance seven times its body length. Remarkably, the jumping height and distance relative to the body length of our jumper
pable of crawling, when actuated at frequencies lower or higher than the jump frequency range, demonstrating autonomous mu
g robot can reach maximum velocities of 0.9 m/s (corresponding to 15 body lengths per second) and can initiate both left and r
which is driven by two propulsors and fully controllable in the two-dimensional (2D) space. The VLEIBot+ can swim at speeds
as well as successful homing control of a simulated quadrotor. The average errors of the inferred home vectors in general stay
WR are optimized considering the total lift and energy consumption through multiobjective optimization based on NSGA-II metho
lymer flexibility, light weight, and low cost, enabling further weight reduction, easy assembly, robustness, and cost-effectivenes
smission) and local chemical stimulation (for inducing the evoked signal transmission). The neural signals of the motor area of t
n model to predict an external force applied to the robot's tip using inertial measurement units. This enabled us to determine th
dels for both the foldable arm and variable wheel, and carry out extensive experiments to test the performance of the foldable a
ation system that resembles organismal muscle actuation mechanisms. We develop a feedforward controller that incorporates p
patial reorientation (regulating robot body's 3D orientation). Inspired by this, a hybrid tail-wheel robot is built, i.e., the tail that cre
a wide variety of different types of structures by combining different modules, this research studies a method of automatically se
nally, we derived and simulate the kinematics and an effective inverse kinematics algorithm. Finally, we validate the stiffness an
mbining the designed observer and the learning-based controller, the closed-loop output feedback control is implemented on th
inates the need for the laborious manual design of robot joint poses. Our evaluation demonstrates superior performance in syn
nts on different real-world dataset demonstrate the accuracy and robustness of our proposed method. The code is available at
and three components in the translation vectors for performance quantification. Instead, we introduce four new evaluation metric
ction of informative data that limits their divergence. The statistical accuracy of both calibration monitoring methods is over 98%
novel convergence voting algorithm, achieving targetless, real-time, and outlier-robust performance compared to existing metho
This innovative technique sidesteps the common challenges associated with lidar feature extraction. Our method's distinct adv
ly obtained. Specifically, This method utilizes the planar features in the environment, making it possible to quickly carry out cali
et robust initial extrinsic parameters. Then, in the second stage, we design a new optimization formulation with semantic inform
ti-camera extrinsic calibration, with cameras possessing highly different points of view and a wide baseline. The 2D pattern can
we apply them to correlate RGB images and depth maps derived from the projection of point clouds. Through weight sharing be
system responses, and estimating dynamic parameters. The proposed method is applied to predict and identify collaborative ro
needed to find good parameter values, we also propose partial evaluations of the reward function, wherein the selected param
The former is used to regress large extrinsic offsets. The latter is employed to iteratively fine-tune extrinsic parameters, while th
moroplasty. We also discuss the generalizability of GBEC for camera-in-hand and marker-in-hand sensor mounting methods. In
sed on the derivation, the mass and COM of the object are identified by weighted least squares. In order to improve the accura
bFormer, an end-to-end network for automatic LiDAR-camera calibration. We aggregate multiple layers of camera and LiDAR i
when an event camera is in motion, changes in reflectivity and geometric edges in the environment trigger numerous events, w
alib (Long Baseline Radar to Radar extrinsic Calibration) is proposed to address these problems. The novelties are: (1) A new ta
er with no explicit pose estimation and no fine-tuning using real-world data. To the best of our knowledge, this is the first examp
eroception into our policy inputs, thereby enabling the policy to sense collisions on different body parts and pits, prompting corre
Gibson, Habitat-Matterport 3D (HM3D), and Matterport 3D (MP3D) datasets within the Habitat simulator. Remarkably, VLFM a
ovel environments, as compared to approaches that use subgoal proposals from generative models, or prior methods based on
mework in online and offline learning methods to demonstrate its ease of use. Through experiments conducted on various challe
onal collision detection libraries. Furthermore, we obtain better accuracy, TPR and TNR results than state-of-the-art learning-ba
control, the planner provides probabilistic collision avoidance guarantees in the real-world. Relying on this guarantee to remain
-deep-learning methods in complex scenarios. The approach enables safe interaction between the mobile robot and stochastic
eness and practicality.
long-short-term memory (LSTM) to propagate historical environmental information from the sequential LiDAR observations. Ad
f the trajectory using DRL and plans a QP trajectory based on them. By utilizing the terminal state and confidence of the trajecto

on the layered structure. We conduct both simulation and real-world experiments to evaluate the developed method and its key

ams than previous suggested ones. We show that this formulation can express a variety of scenarios in a unified manner.
 potential paths often involves higher dimensions comparing to low-level commands, which increases the difficulties of training;
y often find unsafe motion plans. This paper proposes ZAPP (Zonotope Agreement of Prediction and Planning) to resolve the re

at 0.15 bpp, surpassing DSIN (0.925), NDIC (0.956), and ATN (0.955) on the KITTI Stereo dataset.
goal-oriented.
erated based on algorithmic calculation of a signed distance function. Comparative experiments showed that the proposed appr
our method with the baseline [1], which the proposed approach builds on. We demonstrate that the approach results in confide
approach to adaptively inflate or enlarge the obstacles detected on MIMs based on their solidity, and the robot's preferred veloc
s combination of SGP-based terrain interpretation and RRT* planning enables ground robots to safely navigate environments w
ctive collision path planning method, which effectively reduces localization uncertainty.
c rewiring enables fast and robust embedding with respect to unforeseen categories of geometries. Through 41,412 random be
s method significantly reduces the number of grids examined, resulting in higher efficiency compared to traditional LOS checks.
which make no assumptions about the agents’ distribution. The framework is modular and outperforms the existing contro
eir reliability to augment the weights of the feature graph representation used in our GNN. During runtime, our model utilizes m
e that the proposed E-RRT* method effectively solves the path planning problem for HRMs. Especially in narrow environments
ctify the safe control input. The effectiveness of the Lidar-based CBF synthesis for navigation in unknown environments was va
mproves its understanding of physical object concepts, including generalization to held-out concepts, by capturing human priors
r the data acquisition based on the object complexity and plan a globally shortest path. To train our PRVNet, we generate supe
amed OSVP, to directly predict a set of views given a dense point cloud refined from an initial sparse observation. To train our
ries can access all sensor modalities throughout the decoder. In experiments, we outperform the state of the art in transformer-
ble parts and joint parameters during the reconstruction process, obviating the need for annotations. Furthermore, we introduce
branch. Our straightforward yet distinctive solution can make the most use of both range images and residual maps, thus great
ask scheme to adaptively influence the selection of supervision constraints, thereby improving performance in a self-supervised
useful geometric information for grasping. Our neural volume deploys a voxel- grid based representation, motivated by acceler
sizes, patterns, and physical properties. Videos, datasets, pre-trained models, and code can be downloaded from our project w
s task. AmodalFlowNet consists of a transformer-based cost-volume encoder paired with a recurrent transformer decoder which

e and dense samples. Experiment results show that our method outperforms state-of-the-art baseline methods, especially in sce
comprehensive clues. Furthermore, a novel 2D projection constraint is established by observing objects from different perspect
lation volumes. We further propose two iterative update schemes: "ID" which iterates over the same batch of events, and "TID"
ctFormer, a Transformer that learns bird's eye view (BEV) representations by using predefined BEV queries to interact with mu
ain DDPMs in the LiDAR domain, we first conduct an in-depth analysis of data representation, loss functions, and spatial induc
sks and arrange the task-sharing pattern, all while leveraging the efficiency of the hard-parameter sharing approach. To the be
ust the learned features by leveraging the categorical feature representations. Furthermore, we combine the segmentation and
periments are conducted. Based on contact sensing, the developed adaptive gripper can perceive the boundary geometry and
rs using continuous marker patterns. Experiments show that the rotation measurement system could achieve an average static
s. These parameters then guide a spline-based patch generation, which allows us to recreate larger surfaces without an increas
arized or efficiently used in gradient-based optimization. These models are then applied for: offline gradient-based parameter fi
m the distribution of black pins and white markers, creating a 2.5D visualization. Based on the imaging effect, we improve the S
iled deformations. We've made our code and dataset public to advance research in robotic simulation and grasping.
based planner outperforms a reactive pushing strategy in terms of efficiency, reducing the robot’s travelled distance by 23.
 dback from a high-speed vision system.
or on Fn or v. For a given distribution, the likelihood of slipping is a continuous function with respect to the tangential-to-normal
de RL- based skill learning and creates abstract state space to enable skill reuse. More importantly, LEAGUE learns manipulat
domain) and conduct the TTA based on the compound domain knowledge. In addition, to prevent overfitting of the TTA model, w
getting. Our approach also leverages an exponential-moving-average-based semantic memory to further enhance performance
ss (MDP) data with high compression ratio and information content. To ensure consistent data distributions before and after dis
improving robustness when multiple environmental changes occur at the same time.
tractor that distills recent experiences into new library skills, and 4) a lifelong learning algorithm for enabling human users to bo

ailable datasets to evaluate our method. Experimental results demonstrated that, compared to other CL methods, PSNN-VCL n
rning process. Our comprehensive experiments show that LOTUS outperforms state-of-the-art baselines by over 11% in avera
nfigurations, a dynamic model considering servo response and non-minimum phase is established and simulated, and flying po
es in the simulator but also have the ability to transfer to the physical robot without any sim-to-real fine-tuning. Leveraging the d
e this problem and navigate search spaces exceeding those in prior work by magnitudes in the number of possible composition
es locomotion and adhesion. Window frames are detected using image processing with an onboard camera, and design optimiz
ypes and multi-agent systems. Results indicate the feasibility of real-time collision avoidance through deformable object simulat
sion algorithms are navigation-oriented in that they are designed to approximately preserve shortest paths between locations of
d map and predict possible failures caused by the uncertainty of its wheel odometry and unknown risky areas. The UAV dynam
d algorithm is evaluated via the comparison with two state-of-the-art flocking control methods in an environment with obstacles.
g optical coherence tomography (OCT) imaging together with the aforementioned force data. We introduce a unique dataset, in
anoramic images and 2D LiDAR readings. We validate the capabilities of PanNote by collecting a real-world dataset. On these
segmentation approach which jointly classifies the human action, robot action and overall outcome of the action. The results s
mall and simple objects such as bolts to sizable and complex ones such as large car parts. We provide baseline results and the
nd local feature extraction and a Completion and Segmentation Cooperative Module for joint completion and segmentation. Poin
facilitate the application of deep learning models in industrial settings. Project details are available at https://baikaixin-github.io/
sh a classification taxonomy for spatio- temporal interactions. We use our taxonomy to annotate OOIs in video clips of dynamic
e, we compare the performance of the RTS-based system to a Global Navigation Satellite System (GNSS)-based setup. The da

nding. Then, the performance of mainstream indoor scene understanding tasks, e.g. 3D object detection, semantic segmentatio
nditional intent prediction model on large human-human datasets and fine-tunes on a small human-robot dataset. We evaluate
uation to get fine-grained insights into the performance and characteristics of a planner. Beyond the dataset, we provide a simul
an behavior in the presence of a socially appropriate "robot", and human-verified labels grounded in the metric space. To the be
verse (29,520 unique instructions) dataset dubbed RoboVQA containing 829,502 (video, text) pairs for robotics-focused visual
ention and triple embeddings. The input and perturbed graphs are then imported back into simulation to generate corner case s
multi-modal perception. To achieve this, we have collected a dataset comprising over 110,000 contact-rich robot manipulation se
e robots to behave in ways that do not alter the natural behavior of humans in the shared space. Instantiating this principle req
design a scenario taxonomy and augmentation pipeline to enable a systematic collection of ground truth risks under different sc

ories. The dataset also incorporates rearview information to provide a more complete representation of the driving environmen
m, this paper presents LiDAR Dataset with Cross Sensors (LiDAR-CS Dataset), which contains large-scale annotated LiDAR p
sm. The testing subsets cover different real-world scenarios, including the Pool subset and Maoming subset, focusing on challe
 dataset standards and guidelines. This initiative aims to establish a common framework, enabling the seamless inclusion of ex
dataset generated from the pipeline and conducted bin-picking experiments. Our large-scale collision-free annotation enables th
ng with their estimated relative locations. The accuracy of this matching is refined by human annotators, who utilize our annota
mantic labels for tree stems and field-measured tree diameters. We share benchmark scripts for these tasks that researchers m
ric similarity between real and virtual points with respect to the estimated glass plane. We provide a dataset of LS3DPC with sy
lly, the transformation consistency is first utilized to vote pseudo-labels and task confidences. Furthermore, to filter out high-con
on and selection module. Our method achieves a substantial increase in recall rate compared to the current state-of-the-art met
erimental results on several category datasets demonstrate that our methods outperform existing alternatives for articulation es
weather effects than previous state-of-the-art methods. Furthermore, we show how our method can be used to perform simultan
peed. By applying descriptor quantization, the requirements for storage and communication can be reduced by up to 97% witho
ns toward meaningful points and to structurize point clouds following tree structures. In each layer, the top S key points with the
ably, our novel algorithm can now achieve superior prediction outcomes using solely RGB images, eliminating the previous nee
r, we propose a distribution-aware tuning (DAT) method to make the semantic segmentation CTTA efficient and practical in rea
ction. The entire 3D shape dataset containing 4 motion patterns, including stationary, pressing, rolling and slipping. Self-attentio
ct classification, object pose estimation, and so on), as in previous studies, and thus is robust to perception errors. Our model a
density field architecture for dynamic object-centric scenes conditioned on high numbers of degrees of freedom (DOFs). In a 7-D
t generates 6-DoF poses based on the affordance text. We also introduce a new high-quality dataset for the task of language-d
s over other recent RANSAC applications. We conducted an experiment on robot suction-based grasping, comparing our meth
ruct the parts of the robot occluded by the ArUco markers. As such, the markers are artificially removed from the training image
model is optimized primarily on hybrid sets of real and synthetic data instead of the real or synthetic data separately.
ach on the task of agile, vision-based quadrotor flight. Extensive simulation and real-world experiments demonstrate that our ap
age captured by a moving camera against an unfamiliar background. We first record and label a set of optical flows in an indoo
ically integrated system based on a novel vision-based fully convolutional neural network (FCNN), which runs at 39Hz within 10
ve higher precision and recall. Regarding cluster-level pseudo labels, it introduces a hybrid grouping strategy to generate more
aseline approach the peak signal to noise ratio is improved from 21.34dB to 23.37dB while the panoptic quality improves from 5
twork and proxy semantic labels derived from a pretrained segmentation model. Experimental studies on three indoor datasets,
sed techniques. A vital component of our method is the lifting of 2D projections to 3D keypoints. Common deterministic regress
e the geometric traits of lane lines and the intrinsic spatial attributes of LiDAR data to design a simple yet effective automatic an
d, which means it does not require any manual labeling and has a solid generalization ability toward different environments. Th
ntrol to the torso joint and the norm of the end-effector motion velocities, uses a more general reward function and, thereby, en
lso design a navigation system that uses these predicted SOGMs within planning, after they have been transformed into Spatio
We employ ourmethod to learn challenging multi-object robot manipulation tasks from wrist camera observations and demonst
stribution over reference paths in the HD map using the path features and predicts trajectories in the path-relative Frenet frame.
e also introduce modifications to the well-known Residual Learning architecture, retaining high accuracy while reducing inferenc
city, and utilizing Fisher information for weighting parameter importance. In general, we believe research in this direction could h
el that works with a goal estimator and dynamical system to learn features from pedestrian motion history. The results show tha
the cooperative nature of MARL frameworks. Experimental results show that POAQL outperforms existing reinforcement learn
veral robot planning tasks. Additionally, it has the potential advantage of scaling up to more challenging long-horizon planning ta

em, we establish that the choice of the exploration strategy has only a minor influence on the recognition accuracy, whereas da
nected networks. We conduct experiments with autonomous lane-keeping through a photorealistic autonomous driving simulat
age the efficient model as the teacher to distill knowledge into the lightweight model. Finally, we systematically prune the lightw
ated by a voxel grid, and the view-dependent color information of the points is inferred by a shallow MLP. The expansion of the
81.9%, a significant improvement compared to the 34.7% achieved by the classifier trained solely on simulated images. This d
when only one human is in the scene. We evaluate our model in multi-person and human-robot motion forecasting tasks and ob
oint cloud compression and preserves more semantic information, enhancing the performance of machine vision tasks. Our sol
e evaluate our system against other state of the art methods on a set of established baselines as well as new ones introduced
F) on top of the RL policy is introduced to ensure safety during real-world deployment. We provide theoretical safety guarantees
cly available reference model. The median position deviation from the reference model lies below 1mm (0.5% of the simulated
nerates plans satisfying an arbitrary set of constraints and computes them in a short constant time, namely the inference time o
simulated and real-world environments demonstrate the effectiveness and robustness of the proposed framework. Furthermor
esearch and development.
and emph{distributional ensembles}. We carry out evaluations on real-world bin picking scenes of varying difficulty. The objects
orld evaluation data and demonstrate promise for improving online grasp learning on bin picking tasks using a physical 7-DoF F
P) algorithm and leveraging 3D robot joint keypoints, we establish correspondence of 3D coordinate points between the robotâ€
ems. We detail the design of GANDER and provide results of a proof-of-concept system, establishing its efficacy in an autonom
multi-object scenes. Embedded one-shot learning models leverage references within a database for OS-TOG to generalize to n
consisting of physically valid sequences for industrial objects. Experimental results show that our approach generalizes better th
ps://object-adaptation.github.io for additional visualizations.
r this task. However, a visual encoder trained from scratch consistently outperforms the frozen pretrained models. Moreover, w
n and stochasticity of soft-rigid body systems. We demonstrate the effectiveness of DeRi-Bot in moving rigid objects to a target
cy learns robust visual features that allow it to locate the object of interest accurately and consistently, regardless of the camera
n both general manipulation tasks and scenarios with low visual acuity.
rthermore, we show our approaches outperform an implicit memory baseline.
entangled objects into a buffer bin to reduce the degree of entanglement; Pulling to separate the entangled objects in the buffe
without any robot data collection at all. We demonstrate our approach on a fabric smoothing and folding task, showing that our p
ting to avoid end-effector collisions. The design of hybrid-action reinforcement learning module is trained with unified latent acti
ard function for the pushing action, thereby enhancing its effectiveness through increased intentionality. Simulations show that
Notably, in the surgical retraction task, even when trained with as few as 10 demonstrations, our method achieves a median su
f variation to facilitate more controlled evaluations of generalization. From our study, we determine an ordering of factors based
imulation and the real world. Extensive experiments in simulation and the real world show that our method significantly surpass
h-following tasks such as applying glue and sticking objects together, all from demonstrations that can be collected in minutes.
nditions, as an analytic approach requires accurate physical modeling of the interaction and contact dynamics. In contrast, we
environment interactions, and tested this on 200 unseen target deformations. In three attempts, 82% of the trials achieved a g
ticulation modeling and grasp pose affordance from diverse articulated objects with different categories. In addition, GAMMA ad
ssible ground-truth samples. Thereby, we also consider the grasp uncertainty enhancing the model’s robustness to noisy in
pper of the robot and object displacement. Our prioritization scheme favors sampling of contact-rich experiences, which are arg
parent object geometry reconstruction. We create an extensive synthetic dataset through domain randomization, which is based
arning algorithms. We conduct experiments by combining the two-stage agent with conventional offline reinforcement learning a
lled on the robot's parallel gripper. By moving with the robot gripper, this camera gains the ability to actively perceive the object
d on multiple articulated object manipulation tasks simultaneously in simulation and shows great generalizability to novel catego
ation tasks.
 does not rely on additional human domain knowledge such as a simulator or demonstrations.
efficiency in simulation, and it allows us to learn grasping affordances in a small number of interactions, on a real-world setup w
gh grasp pose prediction from a parallel gripper model, then predicts an adaptive action using a convolutional neural network, a
n-uniform anchor sampling mechanism is designed to improve grasp accuracy and diversity. Benefiting from the high-efficiency
vo real world objects. With self-supervised end-to-end training, the performance of the integrated model can be further improve
alidated in the real world on 6-DoF grasping tasks, achieving state-of-the-art results of an 86% mean zero-shot success rate on
int clouds. We conduct extensive experiments on publicly available datasets and report impressive performance outperforming
eature spaces via pseudo labeling for the unlabeled data. First, as the inputs of the ViT student are image patches, the feature
ation-invariant representation. Furthermore, we utilize few-shot segmentation to solve the inefficient training data problem. In Cr
ot adhere to simple probability distributions and can vary with input data. Furthermore, in our network, the binary attention modu
s to dense LiDAR-like point clouds. We evaluate our approach on two different datasets, and the experimental results show tha
nts that demonstrate the practical applicability of our approach to robotics applications. Our results indicate that our method can
on over a new image to predict the original pixel in the original image, while scaling error terms based on the estimated confide
ach shows 40.8%(except Completion Ratio) improvement in accuracy compared to NICE-SLAM that does not use prior informa
any types of 2D and 3D supervision, including depth maps, segmentation masks, and meshes. We show that COV-NeRF match
on our robot to optimize tactile-based policies that maximize the visual reward. In six challenging tasks, such as peg pick-and-p
tive shape control framework which avoids, detects and corrects anomalies automatically. Real-world experiments validate the
obots to confirm the robustness of our approach in various untrained and highly cluttered environments. The results from exper
the proposed approach and shows its great potential for other tactile servoing scenarios with measurement noise, which is inev
ng how this action significantly extends the robot's operational reach. Ensuring safety, we developed a simulation environment i
tion movements. This method requires no a priori knowledge of the specific hand-object system and learns the system's invers
ne key module to tokenize 20 sparse binary tactile signals sensing touch states for the learning of tactile context and the other
the object upon release of the grasp. The proposed method is demonstrated in various pairs of objects that are used in a very p
or the screwdriver's pose via optimization under visual and kinematic constraints, and for its applied wrench on the screw via so

ed robustness, and superior adaptability when confronted with unknown tasks. Moreover, in the context of transfer learning, ou
ectively learn five different DLO manipulation tasks with the same framework parameters. We further provide a thorough analy

ation by two cooperated manipulators. The proposed method was verified by simulation and experiments.
eometric and physical models of the object are treated as low-dimensional modal parameters which can be used to linearly par
mulation of contact dynamics, which is amenable to both smoothing schemes. Our final contribution resolves (iii), where we sho
th intrinsic (using the fingers) and extrinsic (also using the environment), mostly in seconds. Second, our framework allows the
classified small objects in cluttered scenarios.
The former deals with the action conflict problem caused by the disappearance of the UAV agent when multiple UAV modules a
safely and efficiently navigate dynamic, human-dense environments. We evaluate FogROS2-LS with two latency sensitive cas
ions with 5 agents and an example transition with 35 agents. The convergence time grows with the number of agents, however
mulation results that scale up to thousands of cars and hundreds of drones demonstrate the effectiveness of the algorithms. Su
ution is ensured. Furthermore, to accelerate adaptation to dynamic targets and resource distribution that are only perceived onl
only local sensing information through spatial decomposition and spherical mirroring techniques. Then, deadlock-aware guided
deas from LC-ADMM and the Bayes tree. The iDFGO algorithm incrementally recomputes estimates when new factors are adde
communication overhead, and efficiency. For the dynamic scenario, CBBA performs best optimality, while DHBA exhibits bette
 the explicit fourth-order Runge-Kutta method. To achieve the autonomous maneuver, model predictive control is employed alo
adrotor are incorporated into the MPC formulation to simulate actual flight conditions, and an orthonormal function is employed
l network capable of extracting features concurrently from both the workspace and the B̈uchi automaton. This neural network
oblems for unknown systems. Specifically, we approximate both control objectives in a way that enables us to apply the standar
mplementations. Moreover, our friction and rotor inertia identification is validated, and the performance of our model is analyzed
erical examples from the Cosserat rod literature and from recent robotic applications.
n lemma. We generalize the elimination ordering and support MuJoCo-type soft constraint models to obtain O(n+m) complexity
eneral co-simulation scheme for systems consisting of both rigid body components and elements with structural flexibility. A rob
vations and states evolve within this latent space in response to navigational behaviors. Furthermore, our approach leverages e
ally adjusts the compliance control gains to accurately track the desired contact forces during task execution. The key insight of
sis considering the contact state combined with the wrist compliance space. The effectiveness of the proposed method is expe
entail an improved initial pose estimation network and a novel pose optimization technique. The improved network is dedicated
ly realistic assembly sequence plans on a large dataset of hundreds of complex product assemblies. We further demonstrate th
ated through various examples and benchmarked against existing methods in the literature.
he initial transformation matrix. Thus, this paper proposes a novel probabilistic magnetic-aware Map Matching framework for Mu
h a reward model showing how effective patrolling strategies can be obtained from repeated attack-defense interactions betwee
at includes a shared portion of the environment patrolled collectively by all robots, along with disjoint areas allocated exclusively
ptimizes the rate of material flow using the global state estimate in real-time. We show that our proposed framework outperform
nstant wind disturbances and uncertainties in the rotational damping parameters.
h a measured impact response, which is affected by difficult-to-model post-impact oscillations. The classifier can be used for sin
e several of the contained anomalies are not task-specific but general, evaluations on our dataset are transferable to other robo
robot executions should be continuously monitored, and failures should be detected and classified appropriately. In this work, w
hen the complete motor failure occurs. It is shown by both extensive simulation and hardware experiments that FT-Net learns to

explicit knowledge of the occurrence and nature of the failure. The proposed metric is also sensitive enough to detect LOC in m
cation violations of 3 top AVs from the CARLA Autonomous Driving Leaderboard, and found that the AVs violated 71% of prop

nerate each SER, enabling rapid exploration through a switch between local path planning and global path planning to each fron
ese measurements are sent to a local position controller that guides the UAV to and from the power line. Once perched, rollers
creased computation, redundancy, and complexity. GIRA bridges this gap by providing a uniï¬ ed perceptual modeling frame
h a predefined traversal sequence. A multi-objective cost function is formulated to further improve the traversal performance. Ad
ated into the scan-to-map point cloud registration to generate drift-free pose updates by enforcing controlled updates or leaving
novel stance control, Graphic User Interface (GUI), and tendon-driven gripper have been developed based on the lessons learn
As a result, our method yields a computationally efficient source localization and field belief module providing substantially mor
ore, it has been observed that during long-term sailing covering a distance of 1200km, apart from wind speed and robot velocity
task for the robot and generates the target point. Then, the mid-level path planner finds the waypoint path and optimizes the co
solar panels in the direction of the sun. The performance of the proposed solar tracking platform was experimentally validated i
he robotic arms. We propose metrics to quantify these faults and evaluate the damage of the inspected components. Our expe
e identification of the electronic components consumption, mechanical losses, electrical losses, and required mechanical energ
ntrols for both the drone and the robotic arm. The arm compensates for UAV positioning errors and aligns the platform's attitude
ative search of high concentration regions, resulting in overall shorter mission times. The proposed framework is designed to it
but also determine the worth of deploying the ASVs that day at all. Therefore, we propose a persistent monitoring strategy that
georeferenced ground truth point cloud results in 0.8 cm after parameter optimization. Furthermore, our results show the impo
orted away from their source. We demonstrate this system applied to measuring particulate matter in smoke plumes and discus
xcels at avoiding obstacles, and correctly detects that traversable terrains, such as tall grass, can be navigable. By using a seq
d approach for modality switching between different perception modes, effectively alternating the emphasis between vision and
ogy. The sensing performance of the sensor is validated by touch location, sensitivity analysis, real-time touch discrimination, a
omorphic tactile sensor, a spiking neural network and a novel decision making algorithm. Our real-time system achieves classif
g, sensor-specific mechanical, or optical modeling, and instead takes inspiration from electromechanics to analyze the vector fie
ecise and reliable policies that can generalize to a variety of unseen objects and pushing scenarios without domain randomizati
estimation, we show that the proposed method can decrease operational forces significantly over transparency-optimized bilat
via calibration. We further revamp the sensor’s optical system by fixing the ARTag on the base and reflecting it to the same
lLink only requires one motor to actuate the three phalanges, and one camera to capture tactile signals along the entire finger.
nally, we show the advantages of our new hardware design and improvements in calibration methods for accurate depth map g
We collect tactile data from human demonstrations across various tasks and leverage this data to construct a shared latent spac
mark between various configurations of illumination and contact elastomers are provided. Furthermore, the robust design of All
es of the translucent gel so that 9DTact can perform 6D force estimation without the participation of auxiliary markers or pattern
Possible image stitching algorithm to form a global view of the contact. We demonstrate the effectiveness of the GelFinger sen
mer coated with a thin layer of paint with a level of transparency that can be adjusted by varying the sensor’s internal lightin

To this aim, we used a soft optical biomimetic sensor that transduces surface deformation via movements of marked pins, simi
g a novel goal-update mechanism with deep reinforcement learning. We also apply these policies to real-world settings with a ze
ge on unseen YCB objects while just being trained on primitive shapes.
c data generated from the interaction between the robot's legs and terrain to automatically extract terrain attributes. The results
luate our method and to facilitate future research, we construct a real-world dataset for visual-tactile in-hand object pose trackin

detecting the contact patch, and with very little calibration data.
g epoxy. From evaluation tests, this lightweight sensor demonstrated a wide measurement range of up to 9.87 N, with a high s
and demonstrate how SATac could benefit robot perception capabilities.
ocesses without sacrificing performance. Experimental outcomes substantiate the effectiveness of our method, showing 26.9%

m-to-real processes over difference images. It is shown to generate real-like synthetic images while maintaining accurate conta
ception of large-scale object surfaces through rolling. To tackle the challenges posed by laborious calibration processes, we pr
taset obtained from deforming the sensor against various objects. This allows the model to translate simulated depth image se
posed approach holds potential for tasks requiring precise manipulation and limited intrinsic in-hand dexterity under visual occlu
tiveness and feasibility.
y improve completion times as compared to the primitives alone or a naive combination of them.
and physiological information monitoring.
epresentations, a latent garment dynamics model, and future predictions of garment states. Two cost functions are explored: th
ntations, resulting in enhanced material property classification and more adept grasping prediction. Through a series of experim

perty corresponds to the robot-actuating trajectories. A depth-free adaptive controller is proposed based on image feedback, w
 tion identification including 5 already identified in literature and 4 new motions with both synchronous and asynchronous senso
align to the contact angle between the tool and the environment, a Koopman operator-based optimal control scheme is propose
e actual prototype sensor based on the proposed method show that the maximum nonlinearity error and decoupling error in the
s orientation for robotic ultrasound systems (RUSS). Experimental results show that the contact sensing method is robust to fric
robot BRAVER validate the capabilities of the proposed estimator and controller under various disturbance conditions.
ecoupling the contact force from residual dynamics and system noises, thereby reducing the dependence on accurate dynamic
pedance control and formalized holonomic constraints to enable flexible task execution. Later, we propose the quantified analys

hm is probabilistically complete, i.e., if a winning strategy exists, the algorithm will almost surely find it. The case studies and be

ed predictions so that robustness values can be efficiently computed online. We evaluate our approach for the use case of auto
eferred to as the "single tangent space fallacy". This approach involves merely projecting the data of interest onto a single tang
n explicitly accounts for the progress towards satisfaction, and the MILP-based optimization approach avoids the state-space e
ient Mixed-Integer Linear Programming (MILP) encoding for the latter. Our method provides a framework for agents to make in
un-time monitor, that provides a list of safe actions for any delayed sensory observation, given the expected and maximum net
epending on the time of day. Then, we apply this planner to optimize for MITL temporal robustness in Markov Decision Proces
a semantics-preserving rewriting operation, LTL progression is exploited to decompose the complex task into learnable sub-go
n-robot interaction, which subsumes the expressivity of all previous representations. In addition, it allows us to make fewer mod
e encode the environment and the task as the Boolean logic and linear temporal logic (LTL), respectively. Based on the establis
Inter- task Relationship Tree Search (IIRTS), is developed for the fast task allocation of heterogeneous multi-robot systems. It i
50 tasks, LTL-Transfer can solve over 90% of 100 challenging unseen tasks and 100% of 300 commonly used novel tasks with
cted values were compared. The L-shaped surface is a typical shape in which the film thickness distribution changes with the a
lation on an exploration task and on hardware with a robot scooping granular media.
mational overlap between multiple simultaneous queries and availability of multiple robots working in concert to answer those qu

esigned, which improve the grasping success rate by 36.7% compared to direct grasping and the classification accuracy by 39.
roduce a variable stiffness achieved by modifying the kinematic structure of the fingers using snap-fit joints, a very simple alter
ng the most similar-looking surfaces to the one on which the expert grasp is demonstrated. Our model is trained entirely in simu
coder enables information-gathering heuristics to minimize adaptation interactions. The amortized distributions are efficient to c
ments show that our model generates physical plausible grasps for the anthropomorphic hand effectively with over 70% success
howcase the effectiveness through extensive experiments in both simulators and the real world. Remarkably, our framework sig
hat selected grasps are successful. Our experiments show that both kinds of attack lead to drastic drops in algorithm performan
ge and matching challenge. We propose a training strategy that tackles the challenges of learning domain- invariant image emb
scenes and determining the appropriate grasping order. Firstly, the Positional Encoding Module in GSAGED enhances object
onfigurations yield more than 93% of the force estimation within ±1.5 N tolerances for a range of 0-30+1 N and contact locatio
re shows correlations between several Domain Randomization-based quality criteria and sim-to-real transferability. Key challen
ke grasps under direct control of high-level grasp parameters. It is based on via-point movement primitives trained on synergy t
of our sampling region. We demonstrate that our algorithm can run in real time (20 Hz) and is capable of improving grasp perfo
nt Fractal Fingers. The companion paper delves into the kinematics and grasping properties of a specific class of Fractal Finger
rchitecture for object-centric grasping. The method uses pointcloud data from a single arbitrary viewing direction as an input an
1,020 grasps. The grasps were created with a Kinova Gen3 robot arm and Robotiq 2F-85 Adaptive Gripper to enable training o
del-based runtime monitoring algorithm that learns from deployment data to detect system anomalies and anticipate failures. Un
ures during the grasping process. Based on the novel representation, we introduce a deep reinforcement learning method to lea
ons to develop a system, called MoDem-V2, capable of learning contact-rich manipulation directly in the uninstrumented real w
oses are targeted by novel task or joint space reactive controllers formulated using Gaussian Mixture Models and Gaussian Pro
man-robot handover scenario using a multi-finger hand.
t solely relies on the point cloud as input, eliminating the necessity for complete mesh information of the object. By leveraging t
– and conduct experiments on both stationary and conveyor belt setups with realistic vision-based grasp planners. Results s
h has significantly improved the grasping success rates of baselines with only partial reconstruction by nearly 30% and achieve
ctical comparison of four grasp planning algorithms. In particular, we show that the proposed approach allows us to carry out se

sis of the mechanism and control methodology used to achieve the desired behavior are presented. Furthermore, the validity of
sition and force sensing can be achieved at any location of each finger. When gripping an object, the size and stiffness of the o
The results were compared with those of the fingers with single infill densities and a commercially available Festo MultiChoice

s used to transport a part against gravity. Experimentally obtained motion tracking data confirmed the theoretical model. A serie
inger and more than twice the stiffness of a normal condition. Additionally, our design enables multimode grasping and incorpo
en angles and the model-predicted values is 2.31â—¦. Subsequent experiments demonstrated the necessity of the magnetic gri
ved repeatability and the integrated functions.
ometric model, circumventing the need for their explicit definition. We showcase the performance of this system both virtually o
methodology for this new class of robot hands, and multiple prototypes.
xploits the possible environmental constraints regulating the opening of the two fingers and the approaching direction of the grip
nvironmental constraints for iterative design updates in the co-design space. Using these two conceptual tools, we co-design m

ic model for HASHI, as well as demonstrate grip force and torque readings from the sensorization of each chopstick. We also e
d with the fully integrated system in a tactile-aware task - bin-picking and sorting. Code, documentation, design files, detailed as
nstrated that the integration of bilateral control-based imitation learning and the cross-structure hand is effective for grasping va
nylon fabric. A small fan blows gently to help the fabric conform to irregular surfaces prior to lifting.
into a net, and sweeping.
cted with 10 fabrics from everyday textiles, aiming to extend the data collection action modes of current datasets by simulating h
highly under-actuated hand exoskeleton. Our key contribution is the arc-fitting algorithm, which is able to calibrate the misalignm
ce enhanced performance compared to its absence without causing discomfort for a prosthetic user and able-bodied participan
ator and construct its vibration pattern model for any position on the feedback board surface. Optimal phase and amplitude mo
perimental study shows that the introduction of enhanced robotic assistance into a surgical robotics training curriculum improve
e difference in stiffness that users can perceive using Hapstick. Lastly, we perform a study in which participants use Hapstick t

hat entropy conditioned on the sensor state variables. Using a coverage-based controller ergodic controller, we train perceptual
0.75 mm or 0.47% of the display’s maximum vertical range. We also show that our model is consistent with our display, wit
l-in-the-blank pattern for guidance efficiency. Besides, we introduce several optimizations on data transmission and human-com
nd computes a corresponding transparent velocity trajectory. The velocity commands are complemented by an Adaptive Frequ
ation and accurately recognising human activities. This work underscores the potential of robotics in addressing global challeng
evaluations and comparisons with traditional unsupervised approaches, and an online session for assessing realtime control of
cators of human comfort and discomfort as well as calculating the robot energy consumption in performing the desired trajector
on backbone deep learning networks to predict object positions and orientations for robotic grasping. Predictions with Intersect
nformation, we show that our method makes effective use of a limited budget of expert calls at run-time, despite having no acc
{kunde2022recognizing} to assess user proficiency in flying UAVs. We presented a UAV pilot training schedule for novice users
an online planning algorithm to compute the optimal robot policy. We demonstrate our approach on a simulated experiment in
termined by reward estimations. Additionally, by introducing variation without deviating from the original human's intents, more
dressing these challenges, we train a lightweight interaction policy that allows robots to decide when to proceed autonomously
g-tail data samples in commonly used real world autonomous driving datasets. We demonstrate that models trained with CARL
usage history and current status (e.g., energy level) and incorporate the vehicle selection model into a robust adversarial reinfo
ntention. Then, a trajectory planner is proposed to efficiently generate intention-aligned, smooth, feasible, and safe trajectories
le. Two sets of experiments are designed to test the proposed load-releasing method on a bench-top setup with 2 kg, and 3.9 k
orms avoidance by following the surface of obstacles. Finally, we design and deploy several planar compliant contact experimen
s the overall energy flow, regulating and separating non-passive behavior, caused by workspace restrictions. The approaches a
ramework is evaluated by a user study with inexperienced participants that verifies our approach. Further analyses show, that e
solely on the current state. This problem formulation eliminates the need to store previous states. We conducted experiments u
y by (1) modifying the training process to systematically introduce deviations into a pedestrian model, (2) updating the value net
model users' decision making, we evaluate three popular risk models and found that CPT captures people's decisions more ac
ls. Trained using curriculum learning, the model incorporates both time-based and event-based signals. Experimental results d
th a superior performance compared to existing pedestrian motion models. For evaluating socially-aware navigation methods u
fine-grained high-quality motion plans, it is difficult for them to leverage ego-conditioned prediction due to their iterative nature
plying two vectors starting from the obstacle point closest to the waypoint and the boundary point farthest from the waypoint, an
ns where computation speed is not a concern (e.g. post-hoc debugging). However, the optimization-based method is 4000 time

nship among electrodes. To validate this approach, a public EEG-based human trust dataset called EEGTrust is constructed. E
ecognition due to the industrial relevance, difficulty, and competitive baselines of this application; ablation studies confirm the ut
nces. Identification of the mode average most-preferred meta-behavior and minimum covariance bound allows us to analyze gr
esults from a pilot study with 31 participants, showing that such a self-confidence-aware policy increases capability-based huma
ained large models have great generalization ability and do not require additional annotated data for training, allowing fast deplo
modules as the perception unit for neurorobotic systems. This paper proposes a novel deep-learning model providing resiliency f
he target for handover with 90% target prediction accuracy. The target prediction accuracy reaches 99.9% when less than 20%
ot theory of mind model to infer the human's trust beliefs, including true belief and false belief (an essential element of ToM). W
ce than self-learning from pre-defined reward functions via traditional deep reinforcement learning method.
eedback to the user. We conduct an experiment to explore how accurately people can recognize the directions conveyed by the

Our robot-assisted peeling system uses visual, haptic, and vibration sensing modalities to peel a diverse range of food items wi
network along with alternative implementations of standard multilayered perceptron, convolutional, and long-short term memory
oring implicit maps in LLMs, and the other for generating explicit maps using LLMs. The implicit map is in the LLM’s memor
sed into a geospatial distribution representing an operator's preferences over an area. An online, POMDP-based planner, optim
h enhances the speed of the entire inspection process. Limited controlled experiments demonstrate its effectiveness across var
he groundwork for further new robotics applications at large-scale scientific user facilities that depend on Bluesky.
amework at the two test sites we found that the collection of questionnaires were easy to adapt to each work cell and the practi
 identified through high-frequency perturbations in four directions, and two instructed co-contraction levels. The results uncover
es the environment with language instructions, determines the best camera viewpoints for perception, and then iteratively select
ontroller with a greater emphasis on kinematic state-dependent control produced a more natural gait in terms of spatiotempora
ywords: Robot surfaces, User studies

he social interaction dynamics characteristics between the robot and the target person are described by human-robot interactio

ssiveness while simultaneously introducing added complexity. This also presents a new challenge as the belief predicates mus
n in data-sparse environments. We then show how to utilize a conditional model for proactive goal selection and safe trajectory
s a predefined risk threshold or in the case of potential execution deadlocks. Our experiments and simulated demonstration on

atically, triggered by a human-dictated initial guess. The created merged map proves advantageous in scenarios where the rob
one of possible labels for the observed motion, while the recurrent generator networks produce fake RGB videos that are lever
e conduct a pilot study with non-expert users to provide feedback regarding the usability of CoBT. More videos and generated
problem is solved online to ensure that the robot stays within these tubes and the velocities are compliant with the ISO/TS 1506
egies are transferred from an expert human to a da Vinci Research Kit (dVRK) using an epsilon-greedy on a maximum entropy
on from their system in a pool at various distances. This study's results demonstrate that sound is a viable method of underwate
ere impedance and wrench are adapted based on the human-robot autonomy factor derived from interaction forces. System pa
ough trials and errors. In addition, to robustify the care-giver’s policy, we propose a strategy for sampling a care-receiverâ€
using this method in a supervision framework boosts the performance of the human operator when completing the peg transfer
sed on textual and visual similarity for improving performance in novel settings. DROC is able to respond to a sequence of onlin
d without interfering with the main task by exploring the null space with constraint optimization.
ulti-subject experiment involving a human-robot co-carrying task is conducted. The performance of the ergo-interactive control i
oposed method is executed using the da Vinci Research Kit (dVRK) robot in a simulation environment, and its effectiveness is a
on architecture that offers higher transparency and dexterity, addressing the challenges of signal latency and low-bandwidth co
ed in the planning problem, including specific subgoals derived from an NL abstract goal, as well as recommendations for subgo
s by giving the landmarks to an LLM. Through a user study, we find that when presented with explanations from our proposed fr
oblem is addressed using the Model Predictive Control (MPC) approach. The proposed strategy has been experimentally valida
time-varying zeroing control barrier function is proposed and defined to guarantee the bounded approaching velocities by a sa
g completion times with other state-of-the-art planning algorithms, and then experimentally, demonstrating the performance of t
step. These discussions cover critical navigation subtasks like instruction understanding, environment perception, and completi
a task- and platform-independent framework, evaluated in a multiparty collaborative scenario. Obtained results show that a rob
on methods within a preference learning framework. To this extent, we present our method which is able to greatly reduce the n
edits to the canvas and overwriting content. We propose a self-supervised fine-tuning procedure that can tackle both issues, all
th participants, our wearable system reaches a score of $28$ in the NASA-Task Load Index, indicating low cognitive demands
ble for guiding the individual back to the intended path if they veer off course. This is achieved by establishing a target arm end
tion VMPs within a library. Moreover, our approach stores a fixed number of parameters, thus complying with the essential prin
e human-in-the-loop scheme. In order to validate this control scheme, a full-scale wearable analog space suit was constructed
orld. We further evaluate our approach in a human study with 12 participants where we demonstrate that a mobile manipulator
vate appropriate levels of human trust. We conducted a human-subjects study (n=53) testing the influence of the serial-position
sitivity to anomalies in long-term prediction, we propose the Laminar Component Enhanced LHMP approach (LaCE-LHMP). Ou
el interactive continual learning architecture for continual learning of semantic knowledge in a home environment through huma
lebLine, which composes of high-quality portrait line-drawings, with dense labels of both 2D semantic parsing masks and 3D de
environment on a 2D display (VR-2DD) and in a virtual environment using a virtual reality headset (VR-HMD). We trained 15 p

s termed as Noutan (shade) and Kasure (scratchiness) were applied to the strokes by robotic control of a brush's translation, d
em to new locations, we propose a drag-and-drop interface. An operator can grab any robot hologram in a 3D mini map and dra
ulation and hardware, and generalizability is validated by using same learning framework for different walking speeds and terra

raint to stabilize the training process and prevent model collapse. Our system is evaluated on a full-sized humanoid JAXON in t
seconds for different behaviors including walking, running, and jumping, even with intuitive initial guesses. To make the templa
d evoke the NWR. The experiments on a full-size humanoid robot show that the method enables robots to avoid potential dama
me, demonstrate the efficacy of our approach. Particularly, our model achieves impressive lead times and response times acro
patches, forming a human-like surface of the robotic mannequin. To imitate a human body, the linear actuators drive the patche
mances of the proposed approach are shown through extensive tests on the new humanoid robot ergoCub, currently being dev
relief the Deployment Gap and an agent-based feature adaptation module with inter-agent and ego-agent discriminators to red
e before fusion. Moreover, we design a Dissolved Query Recovering (DQR) mechanism to remedy inherent conflicts by preserv
ntion (MCA) and Camera-aware Channel Masking (CCM) modules to enhance infrastructure and vehicle features at scale, spa
hitectures where our proposed method achieves state-of-the-art results. Additionally, we conduct an in-depth performance ana
rehensive experiments on synthetic and real-world datasets affirm the superior strength and adaptability of our method.
ts and colour and depth information of images can be learned. To demonstrate the effectiveness of our framework, we conduct
DAR Simulation) to construct a novel heterogeneous LiDAR dataset named OPV2V-HPL. Additionally, we present HPL-ViT, a p
ancy prediction head, it is found that the occupancy prediction head holds considerable potential for accelerating the model whi
predict human keypoints based on these features. Our method demonstrates that 3D HPE can be seamlessly integrated into a
mework by proposing an auxiliary semantic loss as supervision to the shallow layers of the framework and an image-conditione
d detections in consecutive frames. The affinity is a probabilistic matching that leads to robust data association, track lifecycle m
tribution Gap in multi-agent perception. FDA comprises two key components: Learnable Feature Compensation Module and D
althy for individual agents within the system, significantly reducing object detection accuracy. To enhance the success rates of t
onventional forecasting methods are usually not trained nor tested in real-world pipelines (e.g., with upstream detection, trackin
et, DAIR-V2X-Seq, demonstrate the effectiveness of QUEST and further reveal the advantage of the query cooperation paradig
defogging. This paper is one of the early works focusing on point cloud defogging. It's core insight lies in the estimation of the
Bird's Eye View (BEV) context into compact center representations, enabling feature aggregation at sequence-level to significa
ors from different connected autonomous vehicles (CAVs) may not be properly estimated to utilize the theoretical optimality prop
ith those of offline training data. 2) Closed-loop testing under distribution shifts and out-of-distribution. Contrary to recently repo
vironments and embodiments, H2RIL leverages task-agnostic play data for low-level policy training and paired human-robot dat
ory in the end. We need to allow sub-optimal or negative actions occasionally. But existing methods focus on generating the opt

d for learning behaviors from videos of tasks being performed without access to any low-level states and actions. We do so by
ons of what makes two rewards similar. We introduce a tractable framework that can capture such definitions of similarity. Our
rameter weighting is considered during action generalization to adapt to a wide variety of unfolding scenarios. As revealed by th
simulator CARLA and an assessment tool called Behavior Metrics, which provides holistic quantitative fine-grain data about ro
 y produces a sequence of upper-bound estimates that, however, appear to converge rapidly to actual escape energy. We pres
veloped, and the effectiveness of the proposed system is experimentally verified.
evaluate the workspace with the same adduction and abduction angle ranges. The effects of the parameters of the spherical m
ture space as the critic and then get rid of the hard-to-extract features step-by-step for higher training performance and better a
fabrics capture a much richer and desirable set of behaviors via artificial, second order dynamics grounded in nonlinear geome
framework in the presence of kinematics parameter errors.

how the integration of vision and touch can improve reinforcement learning and Sim2Real performance.
g fingers. To adapt to changes in forces due to model uncertainties and unexpected disturbances, we employ an adaptive torqu
sful climbing. The control scheme of the robot integrates the servo control, the suction control, and the application control in a m
eal-world settings. Given our model's implicit collision sensitivity, we achieved commendable success rates of 92.3% for known
two trials. When the laser tracker is removed, the learned feedforward controller can sustain the accuracy achieved even witho

ehicle information, which addresses the issue of incompatible vehicle resources across various smart city scenarios. And the oth
l level, semantic level, and texture level, inspired by the principle of spatial consistency. These levels of consistency enable our
ectory. The proposed method allowed the camera to stay 94.1% of the time within the real FoV, compared to 54.4% without the

with the robot platform by commanding motion in the mobile base in order to increase the workspace and maneuverability of th
in simulation and on hardware using a 1/10th scale rally car and a 24-inch wingspan fixed-wing UAV.
the planner query occupancy at relevant spatio-temporal points, restricting the computation to those regions of interest. Exploit
nts. The result shows that the safety filter is capable of generating safe inputs in real-time, while addressing hundreds of constr
hor-based future data synthesis, the NVFormer can infer optimal control command for reasonable fuel consumption. The physi
eline, the efficacy of which is validated through simulation experiments.
nitial guess. We validate our approach with Monte Carlo simulations of traffic scenarios.
monstrates capabilities for high-dimensional control and avoiding suboptimal policies via policy alignment. The efficacy of our ap

us approaches.

properties affect performance in three planning scenarios: icy gridworld, simulated plant watering, and real-world plant waterin
ories. We then present an algorithm which leverages a way to efficiently query the model along candidate paths and speed prof
s dependent on reward shaping. We also propose a modification to the uncertainty estimate for the Random Forest surrogate m
nning problems with over 30,000 problem instances and show that planning speed can be significantly reduced by up to 89% w
ilizing iterative route fusion is devised. By incorporating auxiliary supervision and employing a dedicated decoder, we demonstr
plicitly takes into account the multimodality of the training data and simultaneously achieves time efficiency and path optimality.
ore dense subgoals when the MPC method needs more guidance. The density of subgoals is determined dynamically based o
l as a 7-degree-of-freedom manipulation task with dynamics uncertainty, suggesting that the approach finds better solutions fas
e user. Perhaps more importantly, it provides guaranteed bounds on the pose. The computation of such bounds is made tractab
t, that is only available for a restricted set of run-to-failure experiments, typically gathered in an R&D setting. Although being art
ashboard (WD) that provides a living performance benchmark. We share the results of a proof-of-concept electronic task board
t analysis— and allows the implementation of sophisticated PDE control schemes considering fully nonlinear PDE models with
ntrolled by an external computer, exhibits a spatial resolution of 10 um, a translation speed of 60 mm/s, and an adjustable stitch
 nning coverage paths is presented, in which the contact area is modeled to realize an adaptive spacing between path lines. Th
el parameters, as well as those that only address maintenance scheduling and quality-related parameters.
gh extensive testing in three different relevant scenarios in a simulated warehouse environment. Our experiments demonstrate
efining the concept of following vehicles. Experiments show that performance is improved by utilizing new conflict prioritizing rule
e experiments on three datasets demonstrate that our proposed contrastive divergence loss significantly improves the performa
e study results show that the proposed framework has the potential to fully utilize all the database scenarios at query time while
y by orders of magnitude and counterintuitively requires fewer supervisors (per AV) as more AVs are adopted. These analytica
uman decision-making during driving. We ensure that our model exhibits robust generalization capabilities by conducting simul
ctions at a lower frame rate serve as intermediate goals; and 3) uses an attention module designed to temporally align predicted
nted to solve the reformulated optimization problem, with an embodied iterative nearest neighbor search strategy to determine
d driving data. Specifically, we define a novel and differentiable measure to quantify the level of courtesy of driving behavior, lev
en created and updated by minimizing the discrepancy between the predicted and actual behaviors of the driver. With the cogn
propose a novel, safety-adapted loss variation that leverages the estimated per-pedestrian criticality during training. Therefore
n a more balanced distribution and improve the segmentation performance under different spatial distributions. In addition, a sa
the standard tracking metrics like MOTA and MOTP do not capture the combined performance of the two tasks in the wider sp
while considering communication constraints. We then train a learnable encoder to dynamically adjust the compression ratio bas

esembles the dynamics of a typical quadrotor system. The proposed controller ensures uniformly ultimate boundedness of clos
notification from the vehicles themselves. Essentially, it builds on the cooperation among vehicles, which share their data acqu
tracking precision of moving objects, largely surpassing the performance of the state of the art. We release our code and mode
e with vast sensor environments and communication to potentially unwilling human drivers. We consider image observations, a
sensory input into learnable multi-modality tokens, thereby intrinsically alleviating description bias by separated pre-trained per
aset and its associated planning test platform. Our frame-work not only matches state-of-the-art planning methods but outperfor
vant agents for motion prediction. Specifically, we achieve this by implementing a semantic-aware selection of relevant agents
dels exhibit large shifts under non-causal perturbation: we observe a surprising $25$-$38%$ relative change in minADE as com

MRTL for fleet-level emission reduction in mixed traffic using autonomous vehicles as a means of system control. By analyzing t
ottleneck of multiagent rollout by proposing an approximate multiagent rollout-based two phase algorithm that reduces comput
mization framework featuring Closed-Loop Individualized Curricula (CLIC), which we factorize into a set of standardized sub-m

egy (MTOS) is presented to enhance learning of topology and rule constraints. We also devise bilateral query scheme in contex
otion-primitive-based and an optimization-based planner and evaluate the performance on 2000 challenging CommonRoad traf
ory prototypes of the historical vehicle behavior features in training data on a neural decision tree. Each leaf node of eMem-ND
s work, we study the problem of optimally generating driver warnings by considering the interactions among the generated warn
e for real-time implementation through the utilization of GPUs. We illustrate the effectiveness of our proposed methodology thro
odel-based agents.
ns of dense and dynamic traffic. It is noteworthy that, given a narrow chance, the proposed approach generates high-quality lan
encies among agents. Benchmark tests on the NGSIM, HighD, and MoCAD datasets reveal that our model (GAVA) outperforms
ompliant and socially-aware trajectories. We perform extensive testing on the nuScenes prediction challenge, Lyft Level 5 datas
is proposed, which is applied to a home-based bilateral training system. The sEMG-driven CNN-LSTM model builds the relatio
mbines RT's predictive capacity to a representation of the locations of potential reaching targets. Our experiments demonstrate
 8% is achieved, surpassing prior works. In robot-in-the-loop settings, our system successfully detects intended motion solely fro
and let it learn to magnify the optical flow. Finally, the magnified optical flow guided the microexpression images to generate ma
ointly utilizing in-cabin and external features improves performance compared to using in-cabin features alone. Furthermore, ou
n Execution for Robots) can parse descriptive language queries up to 42% more reliably than existing LLM-enabled methods b

multimodal sources of data fused in an effective and efficient step-wise hierarchical fashion and two auxiliary tasks designed to l
n saccadic eye motion due to factors like fatigue or inattentiveness, we incorporate stationary gaze entropy to gauge visual con
parameter to adjust the mechanics of the gripper for smooth recycling of the impact energy. Our gripper closes within 45 millis
ceptive feedback control of trajectories in the task space while eliminating the necessity for any additional sensors. We assess t
e analytical methodology used for validation is also presented, which connects the subsystems of the actuator and allows for th
p between joint stiffness in the robot and the disturbance rejection for force-based control during scrubbing. There is growing de
r variable cross section. This paper fills a gap in the current literature and offers valuable insights for researchers in the field of
nimum tendon deformation, and the coupling of tension and displacement are obtained; the numerical optimization method verif
gn parameters. Finally, the buckling phenomena are used to actively soften the robot for a better compromise between complia

xhibits excellent conformity across a broad range of shapes and provides initial insights into hysteresis modeling attributable to
matics challenges, our method also unveils previously unidentified joint parameters. Real experimental data from knee and hand
rage modality, rowing trajectory and supporting leg shapes through the theoretical analysis and experiments. The fabricated rob
propose an iterative scheme to adjust the temporal course of the selected reference path length. More precisely, the temporal r
also addressed. These outcomes stem from a specific PCC parametrization, with implications extending to the general PCC m

m behavior. We demonstrate the effectiveness both as a constraint and objective function for inverse kinematics optimization. M
task. The approach could reduce the time-to-market of robotic arms and enable the standardization of modular robotic parts. N
sed segmentation approach, both in simulation and using real recordings of human-demonstrated pouring motions. The results
st common behaviors observed in soft robots can be replicated. Case studies include isotropic objects subjected to different bou
e robot is presented. Furthermore, a Quadratic Program (QP) formulation for achieving a desired inertia matrix at a desired pos
oncept implementation of a data centre storage media robot, and show through a series of experiments that its design, speed,
es are readily investigated. The Lagrangian method is employed to formulate the closed-form dynamic model. The quintic B-sp
wrench system are not singularities of the geometric model of the sagging CDPRs, contrary to what happens in rigid-link paral

space condition and a new metric are introduced based on the available twist set. The metric shows the maximum allowable M
comparative analysis with its predecessor, AcroMonk - a two-link brachiation robot, to demonstrate that the presence of a pass

an algorithm using reinforcement learning (RL) for synthesis to unify the learning process of safe controllers and energy function
raining and single-step execution paradigm to make CBF farsighted while the execution remains solving a single-step convex q
g a skill library to address different safety risks. Furthermore, a skill selection policy that considers estimated risks is used for th
anguage Model (LLM) to infer a human’s next actions from a range of modalities without fine-tuning. A novel aspect of our
hat it refers to vision during ``approach" and force during ``insertion," providing interpretability regarding modality use. Furtherm
ally, we employ Parametric Bias to capture changes in this mutual relationship caused by variations in the weight and length of
We evaluate the parameter prior accuracy in five different transfer scenarios, including sim-2-real, transferring parameters to mo
he variants outperform ES-based baselines in benchmark robotic locomotion tasks, while being comparable with or exceeding
 e exemplary motion of skilled operators, collecting data based on these constraints, and conducting imitation learning based on
y from previous work, our method does not restrict the structure of its function approximator, enabling its use with arbitrary DNN
sed Editable Movement Primitive (EditMP), which can learn movements from high-variance demonstrations. These demonstra
arning the association between them with respect to their type for the Pick-and-Place task. The experimental results show a hig
orces and torques for motion segmentation.
obot to another during a physically collaborative task. We chose the collaborative sawing task as a study case since it involves
ges with different stiffness and surface friction. Our results demonstrate that pre-training significantly improves the ability of the
in the potential subspace. We construct a neural cost estimator to select the best informed state from those samples. Then, w
nd uses a causal transformer that learns to predict safe robot actions autoregressively using a critic that requires minimal safety
at govern action primitives in robotic worlds through human demonstrations in simple domains. We demonstrate that our appro
Therefore, this paper demonstrates how the torque output controlled with series clutch actuators can be used to compensate fo
ngle demonstration. Furthermore, we demonstrate its success in tackling complex environments where standard approaches fa
his enables a large number of trials to be performed by a robotic system and the analysis of surgical techniques and their down
m parameters through optimization. Our experimental evaluation demonstrates that this optimization regularizes the forcing term
hat is comparable to or even better than one learned from teleoperated demonstrations lasting over 20 minutes. Experiments d
violation. Our empirical results demonstrate that our proposed Bayesian approach infers constraints of varying severity, more ac
ay-tracing Probabilistic Trajectory Learning (RPTL) framework for Diagrammatic Teaching. RPTL extracts time-varying probabil
monstrations with no human annotations across a diverse set of objects within a category. Supplementary information can be fo
DS's ability to address instability, accuracy, and computational intensity challenges highlighted in the literature, positioning it as
he literature concerning obstacle avoidance within the DS learning framework.
roposed framework, PromptAdapt, leverages the Transformer architecture's capacity to model sequential data to learn demons
a robot arm to produce stable motions for wiping and stirring tasks that do not have a single attractor, while being robust to per
accuracy while simultaneously preserving global stability. To comprehensively evaluate the effectiveness of the proposed algor
ition constraint is further formulated. We also quantitatively analyze the generalization ability of the learned ADS by utilizing the
utperforms state-of-the-art IRL and ICL algorithms, achieving up to a 72% higher task-success rates with accurate decompositi
ures of human motion are saved as a path experience and modified through the path adaptation process of an existing ERTC p
L algorithm as a Worker. More specifically, we pre-train a temporally abstract IQL-TD-MPC Manager to predict "intent embeddin
ing multiple critics in an actor-critic framework to gracefully combine multiple reward functions leads to a significant improveme
as validated on a real robot setup based on four case study materials with varying structural and mechanical properties. We de
simulator and on a real robot kitchen manipulation task show that SPRINT leads to substantially faster learning of new long-ho
es to the Teacher Encoder, thereby ensuring effective information exchange between representation learning and RL. We refer
an interaction using a stick with makers attached to it. Based on the motion pattern of the stick, the robot exhibits walking, runni
er. Specifically, a teacher world model is trained efficiently on state information. At the same time, a matching dataset is collect
interactions in the wild. Despite all the challenges, the sensing constraints bring about creativity: the robot is pushed to gather
y (WODTEM), to improve sample efficiency. We first attempt to introduce an episodic memory (EM) mechanism into the seque
g (MARL) problem. Each segment of the snake robot is an agent, using its local observation to independently determine its acti
ergence and as low as 50-60% actuation effort compared to the vanilla reward. In a sim-to-real experiment with the Go1 robot,
rior sample efficiency and overall performance in multi-robot tasks. Extensive experiments in both simulated and real-world mu
g resources produce faster convergence to higher-performing architectures. We also demonstrate that the neural policies estim
and adapts to changes in dynamics. Videos and code to reproduce our results are available at: https://sites.google.com/berkele
nsfer of a torque-based deep reinforcement learning policy on a real human-sized biped robot.
ecentralized motor control policy. Our method improves the robustness and generalization of the policy without sacrificing perfo
m, we encourage collaborative data aggregation. These resources lower the barrier to entry to utilizing large-scale datasets, the
n integrated real-time framework designed for simultaneous UIE and UOD with domain-adaptation capabilities. Specifically, bo
utilized to train a PSO-kriging neural network (PKNN), enabling accurate predictions of the global mobility map. Through rigoro
e robot-agnostic. Additionally, we can leverage UNRealNet’s predictive uncertainty to both produce risk-aware traversabilit
d to GPS, GLN is less susceptible to external interference, such as building occlusion. In urban scenarios, our proposed D-GLS
s robot properties in addition to environmental information to predict the traversability for various robots in a single model. We va
mantic information for the image and proactively determine a time allocation. Our planner is integrated into a complete mobile r
human behaviors in realistic diverse contexts. We achieve this goal by modeling the explicit and interpretable objectives so tha

. Both affect the efficiency and performance of planning and policy learning. In the paper, we propose a goal-conditioned RL alg
asoning in the keypoint space, our model automatically generalizes to scenarios with a different number of objects, novel backg
the direct use of LLMs in decision-making, (ii) can be combined and complement existing learning methods highlighting its mo
action increase. This paper presents a time-contrastive learning approach tailored for off-policy RL. Our method incorporates a
eralize to unseen instructions and tasks. Compared with the state-of-the-art method, our method achieves higher success rates
ects from cluttered scenarios in simulation and the real world based on human-understandable prompts. Furthermore, we show
uencing two manipulation skills.
we blend the approach of an explicit, yet learned, target state predictor with RL. We further show how a tightly coupled predictor
he learned model representation can directly associate with the policy or task performance. We apply the proposed method to
towards better outputs and smoothly enhances the performance in multi-task benchmarks. We evaluate this approach within th
ories that capture skill stages, robot states, visual indicators, and crucially, tactile signals. Leveraging these trajectories, our offl
elines. Our project page is available at: https://tactilerl.github.io
rmore, a cost predictor is also introduced to directly evaluate actions during both training and planning, significantly expediting
ow that the learned policy directly transfers to objects that substantially differ in scale, position, and orientation from the source
o achieve this, we reformulate the adversarial risk to yield a trade-off between rewards and robustness (TBRR). We show that
m a DLO into the desired shape. Unlike previous DRL-based studies, MultiAC6 is able to solve the sim-to-real gap, achieving lar
nowledge. When the agent is encountered with different appearances under new domains, the complete processing pipeline ca
of interest based on the current robot observation. The PNSS values for a few candidate positions around the robot provide a c
h the target within a finite horizon. Finally, our trained policies transfer well to the real 12-DoF quadruped robot, Solo12, and ex
navigation performance of the snake robot in complex large-scale cave exploration with challenging terrains to exhibit improvem
a shared embedding space between the two modalities via cross-modal metric learning. Additionally, we perform metric localiza
l local bundle adjustment module that utilizes a sliding window mechanism to manage adjacent keyframes requiring optimizatio
-efficient learning and {zero-shot} generalization, this work proposes a novel physics-informed RL (PIRL) where a distance-to-ta
ntal results show that the proposed method achieves better navigation performance than baselines, especially in challenging sc
modalities. Our work motivates the use of foundational vision models for generalized structure prediction-driven applications, e
the-art RL methods, (b) by leveraging expert knowledge in form of an action imitation reward, the RL agent can outperform the
ent's robustness against unseen crowds, we propose diverse scenarios inspired by pedestrian crowd behaviors. Our behavior-
g simulator. Specifically, the proposed pipeline is compared against a modular setup that combines the well-established object
s a first- order method to avoid expensive computations and high dimensional issues, enabling fast and safe policy optimization
he motion, such as the gait patterns. In this paper, we revisit this topic for goal-conditioned tasks in robotics, where symmetry li
ment decision method based on RL, and the computation time of our method only increases linearly with the number of objects.
 ulk policy networks cannot meet the computation speed requirements at inference time on real-world tasks such as autonomou
ach learns to successfully solve a popular real-world labyrinth game in record time, with only 5 hours of real-world training data.
ning (DRL) has shown great potential for learning (near) optimal policies through fast end-to-end inference. In this paper, we pr
imating their corresponding depth values. To train the segmentation task, 464 distinct images are manually labeled into rigid an
guide optimization in later learning. Simulation experiments show that our policy increases the success rate by up to 61 percen
ciency of the learning process. Importantly, our framework proves particularly effective for articulated soft quadrupeds, whose i
ntroller parameters and also reducing the number of evaluations required for the tuning. Moreover, it is shown that the method is
of 99.99%, which is over 10% higher than the baseline random approach, and outperforms other state-of-the-art methods. Using
upedal robot, and showcase interesting human-robot interaction modes including mimicking human videos, following natural lan
n an RL phase trains a control policy in simulation to mimic the generated human motion. Finally, with domain randomization, ou
verify that our framework enables the Unitree’s Go1 robot to walk under various impaired joint conditions in real-world indo
king scenarios in parallel, introducing a symmetric reflection initialization technique to ensure rapid and stable gait skill transfor
on for learning policies that solve a new task, while preserving the style of the motion enforced by the skill library. We show tha
nature of the resulting control policies, as well as the benefits that compliance brings to legged locomotion. In our simulation an
IFM framework can significantly improve the performance of the given MPC controller on rough, slippery, and conveyor terrains
the synthesis of stance foot control with swing foot reflection, improving agility and robustness in locomotion with adaptive beha
e neural net policy operating directly from a camera image, trained in simulation with large-scale RL, can overcome imprecise s
O). The risk preference, ranging from risk-averse to risk-seeking, can be controlled by a single parameter, which enables to adju
arios via a risk distortion measure. Extensive experiments in both simulation environments and a real Aliengo robot demonstrat
gy consumption objective within the optimal policy space for task performance. This guarantees that the learned policy excels i
avoidance and robust legged locomotion. Instead, we propose to learn locomotion skills end-to-end from goal-oriented navigat
framework outperforms a state-of-the-art locomotion controller in a high-fidelity dynamic simulation, especially in scenarios inv

ocomotion for both flat-foot walking and multi-contact walking on the Atalante lower-body exoskeleton.
ble forward walking at 0.75 m/s along with backwards walking, side-stepping, turning in place, and push recovery. We also pre
ied formulation of the CoM and DCM waypoints at the start and end of each motion phase, which makes the framework extens
aches, our framework exhibits superior performance and robustness in simulation. The trained policy effectively tracks a wide ra
the proposed method. The dataset is collected during robot locomotion, and our approach reached the best quantitative result
sumption and cost of transport, which converge to a value of 0.7 at 1.4m/s, regardless of the payload mass.

ctuated spine. Based on the lateral flexion of the spine, we establish a kinematic model to quantify the impact of spinal flexion o
relies on easy-to-obtain videos, readily available in online repositories such as YouTube. It converts videos into motion primitiv
very high obstacle and stairs-like terrains using an Aliengo robot. We demonstrate how Aliengo can safely descend a step heig
within 9.7 degree inside a rotational gimbal and jump in a counterweight setup in Earth's gravity. Overall, we consider SpaceHo
s. The robot differences encompass: a variable number of DoFs, (i.e. 12 or 16 joints), three distinct morphologies, a broad mas
for the design and stability control of quadruped robots, setting the stage for future work on more complex terrains.
rious real-world tasks such as door opening, sample collection, and pushing obstacles. We demonstrate load carrying of more
ork to track diverse commands and trajectories. We initially train the policy within a simulation environment and subsequently fin
o the system, enabling it to achieve desired states even in the presence of inaccuracies in the model parameters or disturbance
s are developed to impose the full-order dynamics to track the prescribed reduced-order trajectories generated by MPCs. The e
work through a series of experiments. In these experiments, our method showcases its capability to generalize reference motio
m, numerical simulations show that there is a big gap between this and the linearized model. Through analysis of the typical no
ed using the same leg-wheel transformable robot, and the results confirm the effectiveness of the strategy.
continuous jumping can be learned in a matter of minutes, with high robustness against various types of terrain.

er reliability and speed in bipedal footstep planning, than was possible earlier. The complete algorithm is also demonstrated usin
2-norm, which represents the worst- case growth of perturbations. In this paper, the worst-case perturbation analysis is used to
ontrol with whole body control framework is used to track the generated trajectories and stablize the system in real-time. To val
uccess rate in the presence of environmental obstacles. Our results indicate that the torso-path guided experience accumulatio
nments and is suitable for use in dynamically changing situations. Our proposed approach is suited to the hardware augmentat
mulated environments during training while eliminating the reliance on exteroceptive sensors during real-world deployment. Ano
of the model proposed. By numerical simulation, we observed change trend of locomotion performance with the configurations
with the original nuScenes static map element, models trained with annotations from CAMA achieve lower reprojection errors (e
vehicles). The dataset includes an accurate 6-dof groundtruth based on a novel methodology that refines the RTK-GPS estima
r that generates high-quality depth maps while maintaining a lightweight design. Experimental results demonstrate that our prop
ation is based on tracking neuron activation values over the list of images to represent and is not limited to a particular neural n

ve re-ranking, offering data-efficient training and strong generalization across several benchmark datasets
e empirical analysis, and presents crucial insights. The results of this study can provide useful guidance for the design of future
es node information across time and space to deal with perceptual aliasing. Extensive experiments on the Oxford Robotcar and
hance the feature representation by the cross-attention mechanism. Then, we propose a Dual Dynamic Fusion (DDF) module to
e with these quantized measurements (termed zQVIO). To improve this ad-hoc quantizer (although it works well in practice), w
lobal semantic map. Evaluations on the Occ3D-nuScenes not only showcase a 20.6% improvement in Success Ratio and a 29
obabilistic mapping. Globally, the map is built volume by volume independently, avoiding catastrophic forgetting when mapping
ned to filter out such pixels. Lastly, addressing weaknesses in the evaluation of current depth estimation algorithms, we present
lidar sensor suite onboard a legged robot, hand-held while scanning building scenes for 600 metres, and onboard an aerial rob
date and refinement during re-localization, while maintaining centimeter-level accuracy. We validate the performance of the prop
pear and suddenly disappear. For these two kinds of dynamic objects, we propose downward retrieval and upward retrieval me
our method can generate semantic object map with shape reconstruction, and be competitive with offline methods while achiev
PVM using the updated voxels. Our experiments using various large-scale real-world benchmarking environments with dense a
is influenced by tilted perspectives. Lastly, we collected a dataset pertaining to staircases and introduced new evaluation criter
arios. We compare our method with recent baselines on both a common benchmark as well as our own datasets, covering a div
them to extract features optimized for traversability estimation. The second strategy involves developing an FSM, which is a se
nfinite data. We demonstrate how to infer global navigational patterns by fitting a maximum likelihood graph to the DSLP field. E
ng deployment. We conduct comprehensive evaluations on large-scale datasets, namely Pitts- burgh30k and Pittsburgh250k. E
d state-of-the-art global registration methods in terms of accuracy and processing speed.
r method is available as open source at: https://github.com/Zhaohuanfeng/DynaInsRemover.git.
he agent learns a spatially-correlated representation of visual and audio inputs that can be applied to work on environments wit
tes corresponding masks across images to 3D labels. These almost view-consistent pseudolabel masks are then used in the s
quently selecting the most pertinent salient ground features through moment computations. For thorough evaluation, we valida
 graph-based optimization method is adopted with a dynamic sliding window that determines what factors are being marginalize
ructed. Finally, we perform robot localization by feature matching under a particle swarm optimization framework. We have imp
el cost-effective reconstruction pipeline that utilizes a pre-established LiDAR map as a fixed constraint to effectively address th
atches is then fused with inertial measurements and point-to-plane registration in an iterated Extended Kalman Filter. The propo
pose initialization and re-localization without an initial pose estimate. In experiments, the proposed method showed an extreme
method outperforms existing state-of-the-art methods in extremely severe situations where the point cloud data becomes dege
ns. The proposed algorithm demonstrates remarkable improvement over the state-of-the-art, achieving up to 91.7% recall at 10
, pose estimation and visual localization especially in challenging scenes.
ettings. Our code is available at https://github.com/alejandrofontan/ORB_SLAM2_Distribution
method. This approach combines semantic information with a reward and penalty mechanism, enabling us to obtain a global pro
y for guiding the scale recovery. The proposed method is evaluated on KITTI odometry, and the experimental results demonstra
s, we also explore its performance on the widely used RobotCar dataset, for broader applicability. Code and models are availab
nd has outperformed existing state-of-the-art LiDAR-Inertial odometry and LiDAR-Visual-Inertial odometry methods.
serve spatial relationships among the backbone features. To extract abstract features with spatial structure information, the mu
tion of the LiDAR-IMU extrinsic and time-delay parameters. The experiments on the public and private datasets demonstrate th
dataset of a large office scene. Notably, our proposed point completion module improves the pose registration performance by
maps. Providing explicit guidance to the two imagination modules significantly improves the precision of the localization system
essor. Benefiting from these delicate designs, our model is able to generate reliable image representations for absolute pose re
accuracy with significantly improved efficiency. Our approach’s core component is SuperGF, a network that aggregates lo
es, which reveal data reliability at the pixel level. (2) SCR is formulated as deep evidential learning with epistemic uncertainty, w
t and compact embedding from a sequence of single-frame embeddings. We show that our model is able to outperform previou
views the contrastive memory bank and applies a distribution-based knowledge distillation to maintain the retrieval ability of the
te estimates, including uncertainty evaluations, but can minimize the nonlinear errors that arise from sensor measurements and
n scenes lacking distinct features and when training and testing environments have different distributions (domain adaptation).
ecognition, enabling relocalization in a previously built map. Our experiments show that ColonMapper is able to autonomously b
tion is demonstrated with a magnetic catheter prototype with a Hall-effect sensor embedded at its tip. We demonstrate position
ntal results on the KITTI and KITTI-360 datasets demonstrate that our approach outperforms state-of-the-art methods. The cod
o achieve real-time monocular visual localization in a pre-built map. In addition, a series of experiments have been conducted to
thin the neural distance field. We illustrate the capabilities to globally localize within an indoor environment utilizing a particle fil
nst several recent state-of-the-art baseline methods applied to GPR image tasks. Experimental results on large public and self-
racy, we provide an iterative trajectory fusion pipeline. The proposed framework is evaluated on two well-known autonomous d
and MapNet. It thrives on large-scale outdoor car-driving scenario, ensuring simplicity, end-to-end trainability, and eliminating ha
ors. Experimental evaluations on KITTI and KITTI-360 show that our method outperforms the baseline methods, and improves
pose refinement to enhance the performance of APR. The extensive experiments demonstrate the effectiveness of our framew
geometric manner. We present a learning method that uses these pose labels and rigid alignment to learn two 3D geometric re
he top-scored database images, discerning the spatial repetitiveness based on detected key text bounding boxes. Our pose ver
e (GRM), which is composed of online Visual Odometry (VO) and a point cloud optimization module, such that the local scene
rate coding that determined the number of spikes; improving spike efficiency by over 100%. VPRTempo is trained using Spike-

ate upon which to build such a universal VPR solution. Combining these derived features with unsupervised feature aggregation
system on a low-cost single board computer (Raspberry Pi 4) to produce global localization estimates at greater than 4Hz on a
on tracking with bounded errors. We extensively validate the proposed NeRF-VINS against the state-of-the-art methods that use
ng failures. Experimental results indicate that our proposed system achieves accurate and robust localization with less than 0.2
ed algorithms.
object detection model on a high-performance multi-core microcontroller onboard the drone, consuming only 2.5mJ per frame a
different sensors can be used separately or jointly to perform localization tasks, and data from the motion capture (mocap) sys
mplemented to achieve effective relative localization of an HMD with respect to a vehicle moving in the real world.

The system was validated using a small ASV in a 2.4x3.6x2.4 m water tank, with a motion capture system as ground truth, and
sts involving dynamic trajectories in multiple real-world cluttered indoor environments. Our method achieves decimeter-to-sub-d
filter (EqF) formulation for the INS solution that exploits a Semi-Direct-Bias symmetry group for multi-sensor fusion with self-ca
ch is optimized for robust performance for our ranging-IMU fusion problem. We evaluate our solution in a densely occupied rea
ramework to address the challenges of radio map inaccuracies and sparse coverage. Our proposed system integrates an unce
iciently extracts correspondences compared to methods using only object categories. Through experiments, we demonstrate th
pdate PF, leverages prior knowledge regarding the spread of distribution for each likelihood function and conducts weight upda
(e.g. moved furniture), unlike our approach, AGLoc, which matches against lifelong architectural features such as walls and do
priors within the factor graph. This refinement provides more accurate spin priors at the beginning of the factor graph, leading to
itialization to more complex localization algorithms or mapping pipelines. Not only does it reduce the error of the mapper, but it
s with data from two recent 4D radars and two IMUs. Surprisingly, our results show that the odometry from Doppler and IMU da
er pruning. Such a hierarchical process couples the feature retrieval and the correspondence extraction to resolve the substruc
cy and accuracy. Capitalizing on this unique aspect of the FEJ strategy, in this work we carefully design the covariance recove
to seamless and pleasurable interactions.
uct a series of experiments to validate the data-driven ability of our approach and to compare it with existing navigation method
arrier functions (HOCBFs), synergized with the previously trained learning model. By leveraging the compound kernel from the f
ralization capabilities of DeformNet for various deformable object manipulation tasks, even in the presence of previously unsee
benefits of both model-based control and end-to-end learning. We validate our method in both simulation and the real world w
actably model the causal interactions between agents at each step. We validate our approach on a suite of more dynamic merg
object parameters with only a single real-world demonstration by minimizing the disparity between the grid density of point clou
lts, we demonstrate successful sim-to-real transfer on the humanoid robot Digit. To our knowledge this is the first demonstratio
istration to train a Hamiltonian neural ordinary differential equation. The learned Hamiltonian model enables the design of an en
nd, we propose Deep Model Predictive Optimization (DMPO), which learns the inner-loop of an MPC optimization algorithm dire
T uses attention mechanisms to distill target system state-transition observations into a context vector, which provides an abstra
e, we developed a carefully implemented library containing a sample efficient modern deep RL method, together with framewor
train a Gaussian process (GP) representation of the constraint manifold. We evaluate our approach on a physical differential-d
ning on video datasets learns representations that are more conducive to downstream robotic offline RL than other approaches
s to explore the rover's turning capabilities. The rover is capable of lifting, sweeping, and spinning its wheels, allowing it to exec
ders improving the task performance, and (2) a recovery policy that maximizes safety. The recovery policy and a corresponding
experiments demonstrate that both methods effectively tackle REMP, with HBFS excelling in planning speed and PMMR produ
representation feature space, which is subsequently leveraged to condition the diffusion model. The diffusion model samples in
ect handovers with an 81% success rate in unseen environments with object instances that differed significantly from the traini
epresentative shapes can be generalized to a wide range of everyday objects. Additionally, a curriculum on terminal goals is de
propose a grasp loop signature which categorizes the topology of these grasp loops and show how it can be used to guide pla
 pipeline to mitigate the effect of point cloud noise in two stages. In the first stage, we learn the affordance on the noisy far poin
mulate actions in a physics-based simulator when necessary; finally we relax the assumption that we require perfect knowledge

Transformer-based lane-topology method. Code is available at https://github.com/NVlabs/SMERF.

ceptron. We demonstrate that this approach facilitates smoother reconstruction with a higher completion ratio with fewer holes.
calable global map is further developed to archive learned features for reuse and maintain constant video memory consumption
uncated dynamic low-pass filter (TDLF) and a numerical gradient averaging technique to smoothen the process. Experimental
demonstrate the effectiveness of our concise representation and the robustness of QuadricsNet. Our code is available at url{http
tations of prior approaches. Through extensive evaluation on open-source datasets, our approach demonstrates superior perfo
y in marginal areas. To the best of our knowledge, our method is the first to achieve high-quality NeRF-based mapping on edge
object shapes and poses during our optimization, resulting in a high-quality object-level mapping system. Moreover, the estimate
r pixel centroids. This unveils a `continuous sense of a place', defined by inter-image persistence of segments along with their i
alth of pre-existing, abstract and inaccurate maps, ranging from floor-plans to sketches. We evaluate the performance of SAMs
he map space can then be estimated with the particles' weights. To reduce the noise in modeling static and dynamic obstacles
e environment can be occupied by moving objects. Our experiments show that our approach outperforms existing moving objec
and label them using the angles of the lines connecting them. When estimating normals at these points, we only consider point
ur pipeline works in an off-the-shelf manner that does not require any feature extraction or matching process. Instead, we find o
lock techniques. The proposed method is evaluated on LiDAR point clouds of city buildings collected by a mobile mapping syst
s in both public and private datasets. Our framework demonstrates superior efficiency in comparison with other state-of-the-art
p Removal (S2M-R) is proposed to discriminate and remove dynamic objects. Moreover, the lightweight back-end removal mod
show that our system performs on par or better compared to other state-of-the-art ad-hoc SLAM/BA strategies, free from data

through extensive experiments. The study effectively demonstrates the robotic fish’s capability to swiftly and efficiently nav
nuous bending. The system can be adapted to various sizes of host systems, and the links and joints are printed in place, simp
ion-class UVMS. WAVE includes an underwater perception sensor and has 2 operating modes: rigid or passive-mode. In passi
ymmetric dead zones of the actuator are removed by adding compensation to the algorithm. During a 10-day field test, the veh
mprovements in speed and maneuverability, achieving swimming speeds of 1.5 m/s (1.7 body lengths per second) and perform
omnidirectional robot has a significant improvement compared with the conventional one with single passive joint in the leg. Mo
spots along the Tyrrhenian coast, in Italy. Sediments collected throughout the tests have been analyzed, extracting information
r a walking cycle's discrete ZMP biases and then parameterize them as polynomials. Then, we use several previous walking cy
he constructed objective function allows deciding the local goal during optimization to both maximize safety within a planning ho
prediction accuracy by the fusion of the sparse feature priors. In addition, without any retraining, our method achieves similar d
l scenes. All objects and scenes are made available in an open-source dataset that includes annotations for object detection an
fer superior long-range visibility compared to cameras, their measurements can appear different from varying viewpoints. This
ges content features, and latent codes more effectively, leading to refined prediction of uncertainty levels. We further introduce
mate complex backscatter effects in water by employing several approximations, enhancing sampling efficiency and numerical
sed model that is computationally light and offers near real-time execution in addition to achieving state-of-the-art performance.
nds, where it was able to autonomously discover the favored habitats of snapping shrimp (genus Alpheus).
projection matrices, which enhances the correlations over two octrees's attention feature embeddings. Experimental results sh

ged and while interacting with the free surface. The results of this preliminary study suggest that the proposed hybrid approach
 action from stability recovery and uncertainty. The average transition time and energy cost are saved by 59.2% and 26.1% com
engine for small time intervals, they do not disturb the environment and can also be used for passive acoustic observations. In
BFs), which increases the disturbance-rejection abilities. We verify our method on an autonomous surface vessel in simulation
mized the minimum difference between the forward and reverse thrust of the propeller-leg to improve the stability of the robot m
s additional parameters to address the camera offset variability. The versatility of the NS model makes it applicable to various l
fine alignment and is better across all metrics at extremely low overlap (10%). To the best of our knowledge, this is the first wo
ol architecture and demonstrate the ability to perform station keeping with a mean error below 0.25 m and waypoint tracking wi
manned surface vessels (USVs). The results from our preliminary field experiment show the collective behavior is robust even w

capacity of each AUV. The planner utilizes a time-varying power profile of irregular waves incident at WECs for dock charging a
rtificial Potential Fields (APF) and Reciprocal Velocity Obstacles (RVO), in simulation experiments, which show that the propose
uncertainty of each map cell using a multivariate Gaussian. The virtual map is then used as a cost map in the planning phase,
se of most current unmanned surface vehicles (USVs) in terms of speed and payload. To the best of our knowledge, we are the
n a high-fidelity side scan sonar simulator. Our results demonstrate that our approach is able to surpass the state-of-the-art in c
on a disturbance observer and sliding-mode control (DOB-SMC) is proposed. A prototype of the proposed underwater manipu
cans, based on the current state of the graph. In subsequent steps this leads to improved registration results. To this end, a com
Stonefish simulated environment.
h on real data collected from two distinct vessels, where the proposed method is able to produce reliable water clearance for dis
rmine marker color. Experiments conducted in a pool and lake show our method performs 10 times better than static color thres
ormance and practical utility of the proposed method were demonstrated through field experiments using a small cruise boat, m
the transition between media and the estimation of the varying refractive index online, thus allowing computer vision tasks acro
arying magnetic fields, mainly outdoors and underwater. We have focused our work on the underwater domain, especially in un
perimental results on the ten sequences of the LBL-AQUALOC dataset in challenging underwater scenes show that our propos

nt-lock mechanisms, where there could potentially be 2^m distinct configurational dynamics, we require solving only n+1 traject
esign simulations and drill resistance experiments, highlighting the effectiveness of the proposed approach in detecting internal
. Our work also holds promise for inspiring future innovations in under-actuated systems control using dualquaternions.
ass position estimation method named as the MOCE (Model-free Online Center-of-mass Estimation) algorithm, which is inspire
ctively replace or complement traditional force/torque (F/T) sensors. Compared with only using the F/T sensor, our approach o
ipod-like structure that consists of two wings and a jumping mechanism. The flapping wing is directly driven by two independen
s between free-flight and perching is proposed. Lastly, we propose a simple yet effective strategy to ensure robust perching in t
onnection mechanisms with full system dynamics, planning, and control algorithms. Furthermore, it includes a novel general mo
ontinue the development of autonomous perching methods.
x ring instead of offering sufficient power to fly with it. In this paper, in order to achieve an efficient rotorcraft for rescuing in mou
ethod to act on the load position error in the presence of the uncertainties. Eventually, we validate the results through an extens
mentally determined the minimum counterweights required for stable perching under various conditions. Building on this, we dev
ator can directly counteract the floating disturbance by following a modified reference trajectory. Hence, a learning-based predi
ontact with the environment. The results show that we are able to identify the drag coefficients of a rotary-wing MAV through on
light tests. Furthermore, we further validate phi-theory through the computation of flight features in simulation and their subseq
d for deployment in human-proximal and indoor environments. Three specific effects (ground, suckdown, and fountain lift), eac
ory, the main result of this letter shows that the optimal path belongs to a set of at most 21 candidate paths, each comprising of
 self-sealing cups. Multiple cups in different directions are integrated into a wheel-like mechanism to increase the tolerance to a
s computed are fused into a cascaded nonlinear controller on the craft to ensure its effectiveness in tracking trajectories. Lastly
and the work surface. To preserve full alignment amidst these uncertainties, we propose a control strategy which selectively de
iments conducted in the Gazebo environment, leveraging the utilities provided by the RotorS simulator. These experiments not
e efficiency of our particle filter are confirmed both by simulation and by real drone experiments.
network that commands thrust and body rates to an INDI inner loop controller, both in simulated and real-world flight. E2E show
n in the cost of transport brought by rolling, equating to a 15-fold extension in the operational range.
ight mission requirements, such as energy consumption and mission completion time. We show that co-designed morphing dro
es. The ADRC scheme is performed when the disturbances are within a range; alternatively, if the disturbances surpass the pre
cts the performance of a UAV during high-speed rotational flight, from both the power and control-error perspectives. To test an

sment process. This process involves the prediction of the saturation level of each actuator during pushing across variable surfa
uxiliary ones for enhanced lateral positional accuracy, ensuring full actuation around hovering. In a demonstration mimicking th
nergy-efficient trajectory for the aerial manipulator. Addressing the challenges of precise control when performing aerial manipu
ovide a solution encoding that avoids deadlocks and formulate an appropriate alternating minimization scheme to solve the prob
dule is interchangeable within its assembly. We also propose a general controller for different configurations of assemblies, cap
rs. The effectiveness of our proposed system is validated using both simulations and experimental studies.
ough simulations and real-world flight tests. The results show the capability to generate real-time trajectories spanning a broad
d using a differential evolution method, and a dynamic model was formulated to specify its dynamic characteristics. Finally, a co
ch and demonstrate how large angle (θ > 2π) designs increase efficiency over planar or small angle (θ < π) clutch designs. We
h foldable wings has been able to overcome gravity to achieve flight, which provides a novel concept for the research on flappi
providing 2D input control, field amplitude, and frequency. A variable number of legs on each side of a bristle-bot’s body wa
e coil diameter, to achieve the desired range of stiffness variation. We test the proposed design on six configurations; these sho

a highly dynamic response, surpassing the torque bandwidth at low torque amplitudes showcased in state-of- the-art literature.
that, in the best case, the gravity compensation device can reduce the energy consumption by 41.15% and the maximum mot

he performance of this gripper compared to the existing ones in the bin and shelf environment is conducted. In this section, the
ontacts by the small robot.
d capacity experiments were also conducted. By integrating a pneumatic soft hand at the end of the proposed robot, an object-g
des were analyzed and experimentally verified. The results indicate that the developed robot can perform versatile locomotion t
ues are intentionally coupled inspired by human anatomy, enduring the high propulsive force to the ground for dynamic movem
n tow floating objects to sail with or against wind. Repetitive experiments have been conducted at the test site equipped with mo
eriment was conducted for water, powder, soil, and mud. As a result, although there was a change in the number of pointcloud

pe is also presented, including its structure, an embedded hardware system, and an upper computer software. The results show
imentally validated. The Lissajous curve-based vibrational orbit control is performed using a prototype actuator. Furthermore, a
out relative movements between the body and environment, and avoids contact between the object and its surroundings as we
emented the model and controller in a numerical simulation and experiments to show their feasibility and the capabilities of our
tubes, thus improving the performance of eSKT. The eSKT with multiple strips exhibits a significant contraction force and has b
me assists in movement and ensures balance, and thus helping subjects with greater difficulty. In this study we present the arch
e proposed control and tracking method has been verified by conducting in vitro experimental studies for catheter steering of a
proposes a data-driven-based solution. A neural network is constructed to efficiently estimate the motion model. Then, an appro
ery. Experiments first demonstrate the feasibility of the proposed models and ex vivo trials validated the usage and effectivene
ing the procedure can lead to excessive forces between the instrument and the trocar, potentially causing patient trauma. Prec
and 2.1%, respectively, for each controller. Both PVIC and PVI-HVC produced good agreement with able-bodied kinematic an
novel visualization concept for improved human supervision of autonomous robotic subretinal injection that integrates uncertaint
e colon wall, enhancing maneuverability. The tracks are constructed from de-molded polydimethylsiloxane (PDMS) and incorpor
uted process, common technical issues have been identified that are buried deep within the dVRK research and development a
rameter estimations. Safety and passivity constraints, including an energy tank, are also integrated to minimise potential risks d

roduces a method to compensate for breathing motion in 3D ultrasound compounding by leveraging implicit neural representat
on method in a Unity simulator and verify its depth estimation capability using a specially designed eyeball phantom. Both simu
g tree is employed to generate a tree graph to connect all extracted key points. To appropriately characterize the rib cartilage o
to the left ventricular longitudinal axis in the acquired images. Finally, the pitch angle of the probe is estimated so that the probe
Input to the first encoder is a two-channel duplex image representing the combination of the grey-scale Doppler and B-mode im
teotomy experiments based on the resulting registration, the average osteotomy error was below 1.5 mm. The proposed US 3D
ensing are proposed. Furthermore, combined with the multichannel one-dimensional convolutional gated recurrent unit (M1-DC
eye. To address these issues, this work proposes an intra-operative vision-and-force-based compliance control method for a po
ents conducted by the automatic robotic cranium-milling system indicate that when using the proposed control algorithm, the fo
space between the patient/surgical bed and the surgeon. To satisfy this constraint, the proposed controller is integrated with a
e virtual fixtures through haptic assistive modes. To validate our system's performance and effectiveness, we conducted initial
APF (SA-APF), which can predict the movement trend of pedestrians and plan the path in real-time. Both simulated and real-w
they operate the surgical instruments to resect the lymph nodes surrounding the blood vessel. Ten human subjects were invite
wrench, generated from visual information, and an interaction wrench, obtained from the tactile handle, into the robot's dynami
pe. To verify the accuracy of the proposed model, bending simulation and experiment are carried out. The accuracy of the prop
e multi-modal information fed back from the follower side, makes judgments, and actively adjusts the surgical operation. On the
size of the robot and compromise motion precision. This study presents a miniature 1R1T precision manipulator with RCM, OR
displacement up to 86 µm. The demonstrated amplifier is expected to apply to compact microsurgical robots or biomedical im
reprogrammed in real-time and in-situ. Magnetic reprogramming capabilities enable the RMCR to achieve complex shape tran
drug treatment concentration and treatment effect precisely in the lesion site against the intestinal peristalsis. A series of simula
recision. Phantom testing substantiates its considerable potential for deployment in neurosurgery.
o address the critical issue of uncertainty in neural network predictions. Neural network predictions are unreliable when the inpu
orkspace estimation. Approaches utilizing the proposed mapping are at least 14 times faster (up to 176 times faster), generate a
sed on the colon-rectal morphometrics, and then tested in an ex vivo environment with porcine intestines. We verified the perfor
he swimmer's position. We also provide an algorithm that computes the placement of the robotic arm and a controller that keep
rgical workflow and generate logs to summarize the procedure for teaching and providing feedback to junior surgeons. The use
simulated test data showed that the SEs outperformed their composite base-learners, however, SVM as a base-learner showed
ventional Ultrasound image sequences. The network architecture was inspired by the Attention in Attention mechanism, tempor
e error of the robot is 1.39mm. The phantom experiment of lumbar puncture validates its feasibility.
vely tracks undesired movements (motions) at designated locations and predicts upcoming motions using neural network-based
 motion scaling and haptic feedback, require a teleoperation control framework. In this work, for the first time, we implemented a
em achieve better teleoperation performance. Comparisons are made with and without using the adaptive motion-scaling algor
ed intestinal tracts. Further, the FCE demonstrates consistent reptile-like peristalsis under the influence of an external magneti
degrees in bending angle and 16.4mm in displacement. The insertion and deployment took less than 2 minutes on average.
maximum intensity. Crucially, these units can be remotely manipulated via a magnetic field, facilitating both translational and ro
time, which is named excessive force protection function (EFPF). Thirdly, the FTG prototype which meet the surgical requireme
d for a robotic-assisted flexible ureteroscope with large deflection up to 270 degrees and a bend radius under 10 mm. We introd
nomous surgical workflow, using the glossectomy as a case study. Porcine tongues are used in this study, and marked using co
orward findings that establish a causal relationship between these two important parameters and the success or failure of the p
movements. Furthermore, participants performed a simple anastomosis task with the robot and compared it to manual control.
e robot. This paper presents an analysis of an expansion strategy for the robot's accessibility and visibility area. The outcomes
vance, usability, and operator satisfaction are also measured using a user experience questionnaire, and interviews and showe

slit design doubles the stiffness properties compared to traditional rectangle slit designs. The path-following task shows that the
man navigation performance in terms of trajectory smoothness, trajectory length, and incorrect branch entry counts. For dynam
hance the control and precise navigation during procedures, a robust mathematical model and image processing techniques are

ustom-designed multi-layer tissue phantoms as well as heterogeneous chicken breast tissues result in less than 1mm in-plane e
d to state-of-the-art reported results which didn't consider self-collision and tool calibrations. Furthermore, physical Human-Rob
ke full use of this system to fabricate multiple magnetic soft microrobots. The experiment results demonstrate that this system c
ive high power actuation achievable with thin-film NiTi.
s help produce lighter microrobots with density less than 1 g/cm$^3$, enabling buoyancy-based self-propulsion. Buoyancy and
p. It holds promise for applications in microsurgery, micro-assembly, and microscale exploration.
S works by aligning the magnetic rod in a controlled magnetic field direction. The prototype was designed and fabricated using

capability and has the potential to be used as an actuation unit for multiple types of functional micromechanisms.
ucted by experienced scientists. Moreover, we demonstrate the usefulness of our platform for inexperienced experimentalists, s
following three motion strategies. The human attention during the task is explored according to the coordinates of the eye gaze
natural advantage of nanorobotics and microscope, the proposed system is capable of tying consecutive throws with micro-sutu
images and sparse depth data obtained by contact detection as input to generate a dense depth map. A two-stage data augme
dynamically adjust compensation parameters under different lighting conditions. Our system has successfully completed comp
l (MPC). The performance of the framework is verified in paraffin tissue sections dissection experiment which shows proposed
manipulator can be equipped with various effectors to accomplish different kinds of tasks. Experimental results validated its high
gularity of the transmission mechanism. The experimental results prove that the singularity of the previous work has been solve
o the order of 10^-1 N and 10^-2 Nm, respectively, considering realistic actuation paradigms and constraints. The attainable for
clinical impracticality. To address these limitations, we introduce a novel segmentation network named the cross-scale guidanc
s, using only a single high-magnification target view as a reference, ultimately enabling rapid and precise repositioning of the m
on between the sample and the glass capillary. Using the system and proposed technique, we demonstrated the automatic cap
and Y axes respectively. The result is a coherent 3.2 μm × 3.0 μm drift-free long range AFM topography without significan
inical platforms. Based on the analysis of the sperm head orientation, an adaptive tail tapping planning strategy is established t
are part segmentation network is designed to reconstruct lost contexts outside bounding boxes and to fix distorted features, by
 nt regions of interest (ROI) and detailed experiments have been conducted on both simulation and in-vivo settings. Results sho
translational operations precisely and controllably. We tested the process of trapping, rotation, and transportation of the mouse
5 mm in diameter) within a bifurcated lumen, filled with blood. This navigation is achieved using controlled rotating magnetic fie
world experiments for picking up all objects demonstrate that Sim-Suction-Pointnet achieves success rates of 96.76%, 94.23%,
s using only observed objects; if that fails, it generates a target object that, if observed, could result in a feasible plan; finally, it a
whole-body safety, agility and dynamic feasibility of mobile manipulators. Moreover, task constraints are applied to ensure that
ce navigation points. We perform extensive experiments in simulation and the real world that demonstrate that, with accurate p
nd hardware experiments on a velocity-controlled mobile manipulator with up to seven balanced objects, stacked objects, and v
GMMs) and Gaussian Mixture Regression (GMR) to predict the dynamic model parameters of MPC based on proprioceptive me
ping poses. This assessment can subsequently serve as an observe-to-grasp reward, motivating the agent to prioritize actions t
ntrol has improved the safety and interactivity of legged mobile manipulators. We conducted experiments on physical robot to c
ene reconstruction and the task-oriented objective, e.g., grasp success, by maximizing grasp reachability. We show the efficac
n an uncertainty-aware model predictive controller that is heuristically conditioned to the aleatoric (data) uncertainty of the learn
s to further improve the efficiency of model-based planners. All data and experiments are conducted entirely in the real world. O
use of Feedforward and Recurrent layers is employed to learn the errors of the first principles model. By combining the NN with
e cost function design which penalizes sampled trajectories with unstable motion, unsafe interactions, and high levels of uncert
ation or zero momentum conservation). We showcase the capability of the adaptive model to efficiently accommodate varying te
ecursive regularised least squares algorithm is demonstrated using the MIT Cheetah 3 leg swinging experiment dataset and co
d trajectories. Output dynamics are then approximated with polynomials and found to contain well-known rigid body dynamics te
icted states of the system, we propose the use of a physics-based term in the loss function, which includes a discretized physic
osed motion model can be used for local motion planning approaches, such as stochastic model predictive control, leveraging m
he context of motion retargeting of bimanual tasks with three different humanoid robots.
und that our mechanical system generated positive and conclusive quantitative outcomes that are similar to the human-adminis
ehensive evaluation of prediction performance, especially in time-sensitive emergency events of subseconds. Interestingly, as
thetic Smarthome, mirroring the Toyota Smarthome dataset. SynthAct generates a rich set of data, including RGB videos and d
force sensors are fixed to the claws, we introduced a new variable, called "grasp depth", to represent the grasping effect on the
ks to the resulting suitability maps where generated behaviors are optimized for each morphology w.r.t. ergonomics and perform
We validate the effectiveness of our model and algorithm in simulations with real human locomotion data. The musculoskeletal m
are performed to extract more informative but low-resolution features. As the output is mesh of human body, we construct a ke
gets for the MHE. The upper-limb musculoskeletal (MSK) model was composed of 10 degrees-of-freedom actuated by 31 musc
oretical results, a series of experiments using our OctRobot-I continuum robotic platform was conducted. The results show that
n of operation demonstrated larger deviations due possibly to unmodeled dynamics, which will be investigated in future work.
als a new direction for advancing control in soft actuators.
e maximum position/rotation error of the proposed model is verified 6.7$times$/4.6$times$ lower than the PCC model consider
tives are designed to encourage low-rank minimizers. CIDGIKc is enabled by a novel distance-geometric parameterization of co
constraints. Our experiments show that the AutoMPC controller reaches targets with significantly higher accuracy and reliability
umber of the key points and the interval between two neighboring points. We demonstrate the robustness of the method when
adaptability. The learned knowledge of robot's dynamics can be obtained and stored in a constant RBF NN model. Based on th
nonlinearities. Finally, we validate the proposed approach through simulations and experiments involving multiple systems, tra
proposed to identify and control a fiber-reinforced FEA following the voluntary movement of hand. The active tremor suppressio
restraining the dynamics to a planar or quasi-static motion while applying it to a real system. Real-time capability is achieved w
uccess rate of almost 65% of the throws in two actuation modalities (i.e., partial, keeping one module of the soft arm passive, a
tion and GPU picking. We demonstrate the utility of our pipeline by doing real-time, open-loop control of the real-world 3D soft r
ed to the complete model to obtain the corresponding actuation inputs. Our methodology is illustrated and analyzed on two clas
and trajectory-tracking experiments are conducted to evaluate the performance of the control architecture. Experimental result
gy-minimization method to reconstruct a physically admissible high-resolution mesh of POE. We evaluate mesh reconstruction
s accurately capture the dynamics of the system shape and body movements of a simplified swimming robot model. We furthe
neral class of underactuated soft robots. After evaluating the controller in simulated cases, we then develop a simple hardware
ment of robotic hand actuation and therapeutic interventions.
several experiments. Furthermore, we did not observe any significant weight shift in effort reduction between the upper and the
ull Earth gravity.

on on the base. We are able to cast our motion planner as a nonholonomic rapidly-exploring random tree (RRT) by representin
ers and free-falling Cassie - with up to 11-dimensional shape variables. The resulting optimal gait in the high-dimensional syste
methods to deal with possible failures during the execution of the navigation/mobility plan. The framework was evaluated in mult
ies to maximize the STL robustness score in training. In testing, similar to Model Predictive Control (MPC), the learned controlle
tion. We present a framework for multi-query TDSP that finds an optimal path that passes through a defined sequence of wayp
vity of the environment. Searching such a dense roadmap for multiple paths is computationally too expensive. The majority of e
MPPI algorithm to find a closed-form ``mode-seeking'' solution that covers only the target mode, thus preserving the fast conve
s the evaluation of visibility for an entire grid much faster than typical algorithms. We provide a practical application of our propo
of the curve, and the clothoid spline decoder effectively accelerates the clothoid tree exploration by relaxing the problem const
n and real-world experiments.
ent. Our approach derives an abstract battery model for future state-of-charge estimation and extends canonical ergodic search
ng space as Informed RRT* does. After the random trees are transformed into RRT* structure by the reconstruction process in
ns is a smooth manifold diffeomorphic to a submanifold embedded in the product space of the CR actuation and base internal w

rmed RRT*, and feeds into PointNet++ for refined guidance state inference. In addition, we introduce Neural Connect to build c
f-freedom. Moreover, we show that our approach does not require high-power hardware by also evaluating on a low-power sing

s that span both spatial and temporal dimensions. As a result, it achieves a remarkable computing complexity of O(T^ 2.4 ) +O(

applications for boustrophedon path plans means that this work has the potential to improve path planning efficiency in numero
me allotted for planning than the online portion of CTMP requires, time that can be used to improve the solution. To this end, we
presence of entrapment including vegetation, and narrow passages in dense environments. We demonstrate our method's capa
d deep-learning-based motion planning. Our diffusion-based network is trained on a set of diverse kinematically valid trajectorie
ce, with fewer polytopes, and in a fraction of the time compared to previous methods.
ds in the motion planning literature and inherits its asymptotic optimality from those techniques. In addition, we describe a meth
e environments. During the forward and backward vertices reduction, a thorough shortcut strategy is put forward to improve the
erial manipulators. Infeasibility proofs in 4D are two orders of magnitude faster than previous results. Infeasibility proofs in 5D c

* with an informed sampler to accelerate collision-free path generation. Our iterative dynamic programming algorithm enables t
h an Nvidia Jetson Nano in the loop to confirm the pipeline's performance and robustness in a low computing power hardware.
one through tracks with dozens of different gates. Extensive simulations and experiments confirm the effectiveness of the propo
pen-source software stack that can be used to combine a wide variety of two-stage planners seamlessly. We perform experime
nts show that our method is capable of generating high-quality trajectories online, even in highly constrained environments, and
with a simplified expression of uncoupled hitting actions. To obtain the best hitting motion, a trajectory optimization method is pr
d to the ground target and, consequently, a condition is checked at each time instant that triggers the opening of a gripper, loca
. The purpose is to address vehicles' motion constraints by using model predictive control on receding horizon and fully exploiti
ned to interact naturallyÂwith the operator for safer and smooth collaborative payload transportation. Successful outdoor experi
ning up an entirely new frontier in additive manufacturing. However, it's essential to note that the feasibility of this pioneering co
, a dual-motor linear active disturbance rejection controller is effectively developed for load tracking task of the DMJ. In the pres

location is used. In addition, the closed-loop stability is analyzed under up to three rotor failure. To validate the proposed unifor
e acceleration commands, which, following the resolution of a Quadratic Program (QP) based on Resolved Acceleration Contro
edictive control-based path following, and high-precision ultrasonic-based real-time positioning. The proposed JetRacer-4WS c
R-NMPC problem, we reformulate it as a finite-dimensional nonlinear optimization problem. To cope with the bilinear matrix ineq
mulated and real robot experiments both in single and multiple robots cases. The obtained results constitute a stepping stone to
ate the effectiveness and reliability of the proposed ILC approach. The results demonstrate the efficacy of ILC in mitigating track
ntal validation is conducted using both a simulation model and an industrial manipulator. Test results indicate that the proposed
he top-level planner. The whole-body motions to adhere to the optimal GRFs while ensuring the friction cone condition at the e
se components, enables autonomous and safe locomotion across multiple layers. We incorporate reduced-order and full-body
arameters of the end-effector under the influence of noise. Kalman filter and a neural dynamics-based method are employed to
real precision-motion system utilized in semiconductor industry applications and compare it to the built-in auto-tuning routine.
ress their patterns in the frequency domain. Our PS scheme aligns the patterns according to the footstep phases to generate n
ility of the closed-loop system in theory, and conduct real hardware experiments on a robot manipulator with seven degrees of
ctuators, which adds to the complexity of modeling, control, and synchronization processes. To overcome such a burden, in thi
motion. The effectiveness of such designs is validated through the real-world experiments.
ile a combination of MV and PT controllers is developed in high-efficiency surgical scenarios, which relaxes the requirement for
n both flat and curved surfaces demonstrates the performance during highly dynamic desired trajectories and confirms the theo
ller accordingly, thus simultaneously achieving stability and performance. The proposed architecture is successfully validated th
sition control in other directions. Additionally, we implement joint angle closed-loop control in joint space to reduce the impact of
two bodies of work on a bipedal platform. We formulate a model-based reinforcement learning problem to learn a reduced-ord
ation of the lifted state and control input. Then the bilinear Model Predictive Control problem is approximated by a quadratic pro
made by the tool. Furthermore, according to the grinding resistance theory, the cutting-surface-deviation model does not require
ontroller. We demonstrated the effectiveness of the framework in synthetic examples, and applied it on hardware to solve conta
edictive performance over common human-driven data-gathering protocols. Furthermore, our protocol converges with 46 s of t
ds. The approach provides a better insight into the topological properties of robot kinematics and control problems, allowing a g
T significantly outperforms the next best asymptotically-optimal planner by 4.01 to 6.58 times improvement in final action cost.
ach with state-of-the-art RRT* variants. Our analysis shows that such an approach has desirable characteristics and warrants fu

defined by the base and path controls, while the latter smooths the control parameters based on velocity and acceleration limita

hod generates a diverse set of Inverse Kinematics (IK) solutions for every waypoint on the reference trajectory and utilizes a dyn
es the Eikonal equation on the constraint manifolds and trains neural function for CMP without expert data. Our results show tha

sults show that successful attacks need only to make subtle changes to the real world, in order to obtain a drastic increase in fa
rst, we propose a method to construct the target information spatial distribution for ergodic trajectory optimization using clarity,
unsolved problems; second, we outperform state-of-the-art algorithms on certain problem types; and third, we present a rigorou
ues during execution. This paper fills this gap by formulating the trajectory deforming and tracking (TDT) problem of a tethered r
ut and point-cloud input from LiDAR. In the latter case, we also study two different settings: fully and partially observed environm
, collision avoidance, and base velocity constraint; which ensures that (2) a manipulator trajectory is computed subsequently ba
method uses pre-trained neural network models and requires no additional data for training and finetuning the model. We also
es, and coordinating many degrees of freedom. Our hybrid planner has been applied to three different robots: a quadruped, a w
oposed to actively search a safe reference path for the UGV, which introduces terrain risk information into the process of travel

nvironments without external supervision. Qualitative experiments show practical applications with a waiter robot.
n has been verified before for single person persuasion, but it is yet to be tested on groups. Thus, the goal of this project is to v
dder, and ensure that object is adaptively delivered based on human preferences. To address the occlusion problem that arises
n constraints and control objectives. Next, a reinforcement learning agent is built to handle multiple tasks with the received info
s mechanical vibrations to facilitate haptic feedback, delivering diverse mechanical stimuli. Notably, the dual-modal e-skin is ca

multi-view interactive hand pose and mesh reconstruction. Specifically, CAMInterHand extracts local pyramid features and glob
best knowledge, this is the first work to utilize 2D skeleton heatmap inputs and the first work to explore 2D skeleton+RGB fusio
r a touch is intentional or not with an F1-score of 86%. We demonstrate that multimodal intention recognition is significantly mor
w state-of-the-art performance by a clear margin. Moreover, we conduct HRI experiments on the carefully selected 150 challen
ion and communication. Experimental results show that the performance of interactive agents that can leverage user feedback
Pragmatic Object Grasping (PROGrasp). PROGrasp performs Pragmatic-IOG by incorporating modules for visual grounding, q
anism to focus on salient features, addressing location dependency issues. Quantitative metrics show that MM-CNN outperform
ot, enhancing the reliability of interactive robot agents. We present a dataset for robotic situational awareness, consisting of pair
n both members are talking or silent, looking at the least spoken person when encouraging her. Results of user study with 40 su
d operational activities while wearing the XoTrunk exoskeleton. The benchmark consisted of a comparison of the presented dev
mbination scheme. Moreover, the necessity and rationale of online estimating system gravity are illustrated. The effectiveness o
nd kinematic constraints to optimize odometry accuracy. Our system's performance is validated using a collected dataset, resu
s compared to previous works.
github.com/MISTLab/Swarm-SLAM
graph convolutional neural network as a feature extractor. It can acquire geographical location information solely through image
moval of redundant data to ensure an precise and robust collaborative state estimation. Extensive evaluations of our system usi
n simulations and experiments to verify the theoretical findings and effectiveness in practice.
and to overcome FOV difference. The proposed method has been validated on both public and our datasets and is compatible
enario. AutoMerge utilizes multi-perspective fusion and adaptive loop closure detection for accurate data associations, and it us
mitations and collision avoidance constraints. The proposed system is validated through challenging field experiments in indoor a
e graph, and integrates both local and global graph information to estimate the reward matrix using graph attention networks an
The robots have a limited-range perception to actively track a limited number of targets simultaneously, of which their future co
 applications that are configured for optimal system size, the supposedly high-performing system may actually live on borrowed

hive. Rewards and diversity allow us to use the diversity of positive optimization in the secondary phase, where performance d
es, we instead provide semantic scenes, as input for decision-making instead of raw images. This approach also helps minimise
lity. We experimentally demonstrate the effectiveness of our approach — it achieves high success rates across all tasks in Ro
chanism to attend to the neighbor robots and obstacle interactions - the first successful demonstration of this mechanism on po
F. It introduces graph-based representation to discover more features on the relevance among robots, and takes advantage of
mprovement with respect to the true state value, 3) applicability to any number of agents, and 4) an action space that grows line
perform ones that are trained with SoTA algorithms in single robot environment.
proposed visually cooperative behavior enables agents equipped with limited field of view (FOV) cameras to efficiently compreh
tion on undetected neighbors. This technique was tested with virtual robots required to complete a coverage mission, analyzing
adjust the size of a robot's region, thus adjusting the amount of space a robot is responsible for serving. Our simulations show t
non-linear factor graph solver while operating with high distributed efficiency of computation and communication. Via the use of
pansion incorporates previous states from immediate neighbors. The broader and more comprehensive information gathered fr
ative pose estimation can be achieved when time drift between robots is limited. To enhance the application range of time offse
shelf (COTS) communication hardware. We further explore the system’s scalability in simulation, evaluating the performan
ment. We perform numerical analysis on real-world coverage applications and show that compared with a greedy approach our
build various superstructures. We report locomotive ability of some of the structures we built, and discuss the dis/advantages o

me auction system to dynamically allocate tasks to the most capable robot in the current team. The system leverages utility valu
ss of coordination among robots in simulated crowd navigation.
is process. We substantiate our methodology through a series of experiments, which corroborate the viability and relevance of
h reduces the representation size of the planning space-for multi-manipulator object rearrangement, the number of hypergraph
or graphs, iSAM2. This estimation method is demonstrated in an example scenario with uncertain dynamics and sensor blindin
s for intelligently scheduling observations with slewing restrictions and changing utility, including a greedy algorithm and a dept
n array of metrics, including place recognition performance, the structural similarity of the reconstructed point clouds, and the b
can reason about their need for coordination, through analyzing over which horizon in time and space their routes overlap. The
me these challenges, we present a meta-learning approach that provides a decentralized planner for each robot to improve the e
e-of-the-art methods in dynamic environments with different numbers and shapes of obstacles. Finally, the proposed method is
ntralised coordination. We present results for a generalised multi-robot orienteering domain that show improved plan convergen

-and-tracking with asynchronous inter-agent communication. Our proposed algorithm DecSTER uses a sequential monte carlo

roduce and assess its performance in random scenarios. Our method, which is released as open-source code, represents a firs
el is used to learn how to weight task/robot pairings (edges) in the bipartite graph that connects the set of tasks to the set of rob
e include: geological assessment, mine planning and development, production drilling and assaying, blasting and excavation, tr
gestions, collisions and deadlocks. The coordination algorithm exploits a novel deadlock prevention approach based on time-ex
tion distance.
servability constraint, we leverage two main tools: (i) the weighted Observability Gramian with a forgetting factor as a measure
able motions. A classification of singularities based on graph substructures is developed, and it is shown that arbitrarily large for
ate distribution entropy maximization problem, and employ multi-layer perceptron (MLP) to model the relationship between eac
 d resiliency of a swarm without a centralized command. We provide a rigorous analysis of the proposed strategy and validate it
ing a fitness evaluation that combines the swarm's classification decision accuracy and decision time. We use our fitness meas
A prior approach mitigates this by decomposing the structure into smaller components that can be planned for independently, b
re precision grasp based on the differentiable Q-distance metric. We validate our method in both a simulated and a real-world e
perimental validation of the efficiency of the platform with the execution of complex teleoperation and telemanipulation tasks. Th
implementation of tactile sensing capabilities, such as force and capacitive touch, simply by adding off-the-shelf sensors (air-pr
nd pose is natural and reasonable throughout the trajectory. Finally, we present a joint target optimization strategy to enhance
ing wearable Robotic Hand(HIRO Hand), which integrates expert data collection and enables the implementation of dexterous o
ked fingers with a person. The performance for object grasping was also evaluated. We demonstrated that our proposed hand
egration enhances the hand’s tactile sensing capabilities.
ning the provided datasets, it was apparent that each mixed dataset contained a significant proportion of expert data, which sho

(CBFs), the multiple UAV-USV landing control problem is formulated as a constraint-based optimization one. Therein, slack va
ints. Leveraging distributed control barrier functions, our approach maintains global graph connectivity while optimizing the perf
that these constraints are sufficient for the followers to maintain formation. Moreover, we present globally asymptotically stable
important. This is the first time that a distributed GBP multi-robot solver has been proven to enable intelligent collaborative beh
rmore, we introduce a novel framework for the dynamic selection of distinct 5G QoS data flows that is autonomously managed
the mechanical design of the system, its initial software stack and evaluate the repeatability of motions executed by the propos
uate cost and latency models and confirm that FogROS2-Config selects efficient hardware configurations to balance cost and l
experiments showing our MID system maintaining the delivery of critical mission data in a situational awareness setting and ena
xperimental validation of the results conducted on a 7-DoF Franka Research 3 robot demonstrates a substantial enhancement i
the embedding of parameterized rays, thereby matching rays pertinent to the image. Through synthetic and real evaluation set
ance, comparison of VeloVox with several SOTA detector based baselines is evaluated on our in-house dataset and synthesize
RadSSD outperforms those state-of-the-art approaches in terms of real-time requirement and detection accuracy. In detail, the
ust the 3D object localization perturbation along the depth axis. In practice, the DE is explicitly realized by using both the absolu
conditions by addressing these two gaps separately. Our method resolves the style gap by concentrating on style-related inform
keypoint detection including a new objective function. Our SyMFM6D network significantly outperforms the state-of-the-art in b
a modulator for calibrating uncertainty bounds derived from CI based on a weighted loss function. Our simulation results show
RANSAC-PnP algorithm to robustly solve for 6D object pose. Extensive experiments have been conducted on both synthetic an
nd CAMERA25 demonstrate that FAP-Net achieves significant improvements (14.7% on 10°10cm and 11.4% on IoU50 on th
en exhibit unstable performance. The purpose of this study is to achieve stable performance for detecting unknown objects by
n the widely used Waymo Open Dataset, our PVTransformer achieves state-of-the-art 76.5 mAPH L2, outperforming the prior a
t dataset with aligned grayscale images as a teacher modality. We show that object-centric distillation allows to significantly imp
earning. Second, we propose Size-Adaptive Local Attention (Local-A) with adaptive attention regions for localized feature aggr
urning cases. On Waymo Open Dataset, our FrameFusion method consistently improves the performance of various 3D detect
e that with sufficient representational power, pillar-based methods can achieve performance comparable to other representation
module for extracting axial details and a lightweight temporal alignment module for associating keyframes and non-keyframes o
design by shifting partial channels from neighboring clusters, which enables richer interaction with non-local regions and thus e
al results on the Cityscapes and MS-COCO datasets demonstrate that our approach achieves state-of-the-art performance in a
e experiments on the View-of-Delft (VoD) dataset show that our proposed method outperforms the state-of-the-art (SOTA) meth
labels with this prediction, we provide weak supervision of pseudo-label accuracy. To further account for low quality pseudo-lab
 rity scores for each associated pair during feature aggregation. Starting from the BEVFusion baseline, we further introduce (1)
sed on the current and past observations. The features predicting the current objects from branch (i) is fused into branch (ii) to t
d CBAM is applied at the end of the backbone, thus not only decreasing the computation cost compared with previous works b
ing its ``U-shaped'' encoder-decoder architecture to generate predictions with sharper boundaries and finer structures. Furtherm
vel coarse-to-fine detection strategy based on vision transformers to achieve precise drone detection. We assess the effectiven
d end-to-end trainable framework, termed AsyncDepth, to effectively extract relevant features from asynchronous LiDAR travers
out requiring access to OOD data while enabling compatibility with established pre-trained models. Our resulting approach can
an be used to identify objects and accumulate corrected object-level segmentation masks. By introducing motion to regions of s
t obstacle tracking, our approach integrates implicit tracking. Sharing trajectories between agents can cause potential collisions
how that our approach outperforms previous works and adapts to new affordance labels and unseen objects. Notably, our meth
k based on MTrans, which not only generates pseudo labels but also quantifies the associated uncertainties. However, applying
utilizes the differentiable priors as dynamics and observation models, executes faster. We demonstrate that HandyPriors attains
ng using the publicly available Radarscenes dataset.
rse data, we indicate change rules of separate classification models for object representation. Subsequently, based on ensemb
arious items, doors swinging or even balls rolling around. We thoroughly evaluate our approach on real-world data sets, achiev
h different object and environment characteristics. Finally, we evaluate the performance of three leading OD models on the div
point refinement scheme is devised for proper aggregation of features in 2D and augmentation of point features in 3D. We eva
of center-based tracking to associate instances reliably. Our approach shows an improved performance on the moving instanc
nt capturing such multi-level motion in a BEV. To address this problem, we propose a product group equivariant object detection

s uncertainty estimation and Out-of-Distribution (OoD) techniques on this difficult benchmark.

To enhance the supervision signal, group query representation is proposed, leading to better performance without affecting the
ecifically, the Historical Frame Prediction (HFP) module is introduced to make compensation for information loss with temporal
be used to group points into approximately convex parts whose convex hulls can be used as compact scene representations in
descriptor on a 2.5 fps dataset. We also demonstrate strong results with our algorithm on DanceTrack dataset at 20 fps and pr
od is able to generate more robust and consistent pseudo labels that would benefit the training process on the target domain. E
-the-art approaches on normal data and outperforms them on misaligned data.
path with 18.28% faster inference time.
lization error by accounting for the characteristics of the radar point clouds. The experiments on the public nuScenes dataset de
ly analyzes its contributions to performance enhancement. Additionally, we introduce an OVA-based semi-supervised learning

se a new acquisition function that takes 3D point density changes into consideration which arise due to location changes or ego
ss has a characteristic size. By sampling object clusters according to their size, we can thus create a size-balanced dataset tha
obustly predict the target trajectory even in the presence of unknown disturbances. Finally, the Linear Time-Varying Model Pred
Predictive Control (MPC) controller formulated as a mixed-integer quadratic programming (MIQP) problem for the 4WS vehicle.
desired trajectory by solving an $n$-dimensional Sequential Linear Program (SLP) iteratively until convergence. Lastly, we eva
a more clustered eigenvalue spectrum. Numerical experiments with typical trajectory optimization problems reveal that as comp
pulator, MPCGPU is able to scale to kilohertz control rates with trajectories as long as 512 knot points. This is driven by a custo
via a time-adaptive model predictive control scheme whose allocated tracking time can be adaptively adjusted on-the-fly, there
epresentations that do not depend on the calibration of reference frames or sensor locations. The tuning process for the optima
ed by minimizing squared error), which may extend to arbitrary objectives with iteration. Our work extends prior works, which fin
 nimization of computational costs, but also continuous optimality and constraints feasibility even under sharp transitions. The ap
elds, interactions, and terrain parameters.
on pre-defined modes. In this paper, we introduce a novel contact-implicit scheme designed to optimize both control input and c
od to compute the forward dynamics and the analytical derivatives of series elastic actuators /variable stiffness actuators and un
d with standard estimation tools enables to achieve state-of-the-art performance in force control while preserving the benefits o
everal experiments using a Franka Emika robot. The presented algorithms were implemented in c++20 and resulted in the publ
behavior needs to be collected to maintain safety. We demonstrate our formulation on a case study of a planar multirotor attem
me cost is a CBF. We demonstrate the effectiveness of our method in simulation on a variety of systems ranging from toy linear

d over- approximations of the value function with low-degree polynomials, which are used to provide stabilizing controllers for co
vely. This allows the robot to handle unexpected contacts in real time. Real-world experiments show the effectiveness of the pro
rm granular particles into a target shape. The model and results are shown to be relatively fast and accurate. The control comm
e trajectory optimization process less prone to converge to poor local optima. Our method is validated on several reaching prob
parameter tuning and data scarcity. Incremental and safety-aware exploration toward the limit of handling and iterative learning
sary, where the controller plays first, choosing the control input, whereas the adversary plays second, choosing the noise's real
portantly, a necessary and sufficient condition for the existence of a robust controller leading to the desired bound is characteriz
Without relying on the knowledge of the system’s dynamics, the external disturbances or the complete state, the implemen
™s robustness through a time-varying Lyapunov-based region of attraction (ROA) analysis. This analysis provides a formal gua
extends HEDAC to applications where robots have multiple sensors on the whole-body (such as tactile skin) and use all senso
m. These benchmarks indicate that ReLU-QP is competitive with state-of-the-art CPU-based solvers for small-to-medium-scale
sing a simple, but yet effective, concurrency control scheme. We run a series of experiments in simulation and with real robots
trajectory to guide the follower using receding horizon planning. We use simulations to elaborate on the effectiveness of our ap
e. The service vehicle is permitted to terminate its path at any vertex. The goal is then to find a pair of paths so the convoy reac
earns a weight distribution over all spectral coefficients. Finally, in the third method, each agent learns weight distributions as p
the complex interactions between robots and humans, thus promoting cooperative tendencies in multi-robot systems. Our exte
he performance is maintained under different vehicle arrival rates, lane lengths, crossing speeds, and control horizon. Through
ed in a large close-to-real map and a maze map and compared with two alternative approaches. The results confirm the effectiv
a new formation planning algorithm that focuses on adjusting the start times of motion plans to avoid collision. Our experimenta
d-of-view sensing shapes, and for both single and multi-robot IPP. We demonstrate that the proposed approach is fast and acc
mework that is based on state-of-the-art open-source navigation and simulation frameworks and can use different types of robo

ance queries from the Octomap, providing an implicit non-convex decomposition of free space. As a result, our approach is sca
ractions, and scales better with higher numbers of robots without compromising the solution quality across a variety of environm
mization. The procedure is repeated with a reduced discontinuity bound. Our approach is anytime, probabilistically complete, as
let agents sequence local decisions based on the full current state of the system, and avoid such myopicity. We further allow ag
that our algorithm scales to hundreds of robots in large complex workspaces and consistently outperforms a state-of-the-art on
o robot teams of size up to 25 times larger than existing methods in this class of problems. By leveraging knowledge of the topo
remain complete for RMMTC problems. Additionally, we explore the use of model optimization warm-startup to further improve
g capabilities of the N-MPC to effectively avoid collisions in cluttered environments. The associated code is released open-sour
ier curve in order to approximate both the cable angle and angular velocity. Our method is initially validated in an indoor environ
, essentially unlimited, operational endurance.
monstrate the effectiveness and time-efficiency of the proposed method in both simulation and real-world.
bal exploration by balancing the exploitation of a known map with the exploration of unknown areas. In addition, the online path
verages two distinct feature pyramid backbones, with one incorporating dilated convolutions to address large-scale objects and
model predictive control (MPC) is designed to achieve active vision on our UAV. According to our performance evaluation, our
PU to enable Jetson-SLAM achieving high accuracy and speed on embedded devices. (iii) Jetson-SLAM library achieves resou
lision information into general LiDAR-based sensing and planning frameworks. Our simulation and experimental results show th
of risk within the FOV. We incorporate psi_d into a safety-critical flight architecture and conduct a numerical analysis using mult
atial hash map for rapid GMM submap extraction combined with an approach to determine relevant and redundant data in a po
wing the vehicle to safely map an unknown environment. Extensive experimental and simulation results show that our approach
nto independent sub-problems, enabling parallel planning for each subspace. The carefully designed global and local planning s
bots, and enables a methodical exploration of rooms. We evaluate the performance of our pipeline using a team of up to 3 aeri
ting the navigation of a UAV fleet. In this spirit, this work proposes an exploration strategy that enables multiple UAVs to quickly
s are conducted and demonstrate the efficiency of the emerged behavior and its resilience in real-life deployments.
e gap with aggressive tilted angles. The policy trained only in a simulation environment can be transferred into different domain
d measurements, and a CBF is constructed to parameterize the (un)safe region in the 3D space. Given the CBF and state feed
edictable for two whiskers based on the percent of occlusion. We then show that a 2x2 spatial array of whiskers responds asym
multiple computationally efficient but low-accuracy detectors to achieve real-time high-accuracy obstacle detection. Besides, we
fferentiable and accurate visibility model that allows decomposition of the trajectory planning problem for efficient optimization r
with artificial intelligence which help us build a functional model of the plant. The same approach is used to program different pla
reliably detect updrafts solely based on measurements of the aircraft's position and the local vertical wind velocity. Nevertheles
model size of only 4.5M. This network can achieve the state-of-the-art performance on three challenging datasets. Apart from
everal baselines, obtaining a retrieval accuracy of 88.2%. Our method generalizes to a farm with 100 robots, achieving a 12.5x
king. In the second stage, pose tracking is still performed to provide object poses in real-time, enabling our approach to transfo
d in detail. We also present a custom calibration pipeline where accurate digital replicas of conformal e-skins are generated for
ction cup MIGHTY, and introduce GEM-C, a novel Gravity, External forces and Motion Compensation controller, that constantly
from the simulation and real-world scenarios throughout the training process. Due to such designs, the proposed method subst
such as a kitchen. By first learning a background NeRF of the scene without transparent objects to be manipulated, we reduce
eriori (MAP) inference. These sparse points are then used to generate edge candidates and the belief of each candidate is upda
rent liquids and fill levels. We show that understanding liquid properties can facilitate effective liquid manipulation. More importa
onfigurations more accurately than comparable methods. Moreover, we demonstrate that object reconstruction from HandNeRF
a robotic manipulator equipped with an eye-in-hand RGB-D camera showed a 100% clearance rate when tasked to pick all trus

pervised low-light feature enhancement strategy is adopted to reduce the dependence on paired data as well as improve the alg
on to existing priori-information-based approaches. Experimental results confirm the efficiency of the proposed approach and d
ching the object, while also being able to update the model quickly from kinematic sensing during the interaction. In this paper, w
oduce a method that uses machine-learning to anticipate slip by computing the so-called safety margin of the grasp. This safet
ondence. The alignment module then computes the cross-pose as the SE(3) transformation that aligns these set corresponden
experimental results demonstrate that our approach achieves a level of pose estimation accuracy that closely rivals that of sup
es. The estimation strategy is validated through simulations and real-world experiments, demonstrating its potential in handling
orally-smooth grasp poses and works robustly against large depth-sensing noise. Using AnyGrasp, we achieve a 93.3% succes
o harvesting experiments conducted in commercial greenhouses, our proposed robot achieved a harvesting success rate of 86.
nner finds the most stable local grasp on the reconstruction, and finally the user is guided by shared control to this grasp. In exp
commercial robot gripper. Coupled with a customized impedance sensing board, we collect data on liquid manipulation with a sw
ard assignment, providing supervision to continuously shape and improve grasping behavior as the robot gathers experience in
of almost 95%. In addition, the implementation of a real-time majority vote during manipulation further improves recognition. Ad

bject rearrangement system that employs multi-view fusion. By observing the current scene from multiple viewpoints before ma
ves comparable results to a state-of-the-art DNN technique that fuses frames, depth, and events. We demonstrate smooth, on
ured from a single RGBD camera placed egocentrically on the wrist of the arm. This approach allows fast training of the real-wo
y in an optimal fashion with best approximation error. We demonstrate how this approach performs on multiple robotic platform
ork for further collaborations in the community to explore practical aspects of existing reinforcement learning methods. Dataset
s that evaluate both the single components of the system and the system as a whole, introducing an evaluation metric that allo
valuation framework, researchers can more easily compare different techniques and algorithms, leading to faster progress in de
we propose 1) a method for extracting driving scenarios at varying levels of specificity according to the structure of the roads,
sed detectors can outperform popular LiDAR-based detectors with our new metrics past at 10% depth error tolerance, suggest

s, RobotPerf establishes a standardized approach to robotics benchmarking. As an open-source initiative, RobotPerf remains c
elieve that, in general, results on the manipulation of clothes should be reported along with a better description of the garments
h safety and task performance. Experimental results with a collaborative manipulator validate the presented framework.
he Martial Art of Iaido. We built and deployed in real-time a ML-based Iaido strike recognition model trained on expert and begin
pect to the direction of the load. Furthermore, we propose a hybrid stiffness/position generator for online task execution. This ge
ble top tasks where a robot supports impaired users by predicting their posture and proactively re-arranging objects.
the virtual environment and human arm force is proven by employing the concept of virtual stability. Additionally, the Z-width of
g manikin. We show that our method generalizes well in the real-world opening insertion tasks achieving an overall success rat
cquired surface electromyography (sEMG) of five major lower limb muscles for seven subjects to evaluate the significant reduc
eleton (CDLE). A swing phase gait intervention using three different CDLE cable-routing configurations that impose varied dyna
e survey technique that emphasizes direct comparisons between items on the defined criteria. By using the NASA Task Load In

bjects than previous work. In our experiments, we show that our method trained purely with synthetic data is competitive with st
sponding to different levels of commitment of the robot to its desired goal configuration. The control architecture is loosely base
musculoskeletal and sensorimotor control models provide novel insights into the human response that robots can utilize to impro
on to performance of each input.
man operator’s experience and cobot control parameters. Measurements include the human operator’s physical effort, t
d with the Maestro hand exoskeleton, and repeatability of the interface power estimation is confirmed with pilot human testing. S
s of a variable admittance controller. An optimization technique based on users' preferences, i.e., Preference-Based Optimizati
observation can be established, which reduces the overall exploration time efficiency. To intelligently balance the robustness of
is compatibility with state of the art trajectory predictors, e.g., RNNs and LSTMs, while making no assumptions on the underlyi
O, our method enhances navigation safety by significantly reducing the infeasible motions in presence of uncertainty. Moreover,
ning using a much smaller number of parallel simulations on regular CPUs. The novel Shield-MPPI algorithm is tested on an ag
robust optimization (DRO) techniques. The safe halfspaces are used in a model predictive control (MPC)-like safety filter to ap
aged to guide the growth of hypotheses trees alongside the belief trees. We then evaluate our approach in highly aliased simula
ta associations. Our key contribution is to derive bounds between the value function based on the complete set of hypotheses a
 cations. We then integrate these shields into the POMCP algorithm for safe POMDP online planning. We propose four distinct
r a given edge should be added to the roadmap. We test our approach in several environments, demonstrating that our uncerta
ounds the risk of the path, offering trustworthiness. We empirically validate that our method produces safe motions and accura
rwise, COBeTS will guide the search towards a safe sequence of option primitives, and hierarchical monitoring can be used to
al scene content. To quantify the uncertainty on the learned surface, we model a stochastic radiance field. Our experiments dem
xecuted transitions is different from the expected one (e.g., when the robot unexpectedly collides with an obstacle) so that futur
aluate exploration planners in such settings we propose a new enhanced benchmark suite with several dynamic environments,
, along trails, and off-road terrain, using a Clearpath Husky and a Boston Dynamics Spot robot. In practice, our approach result
lgorithm. With an increasing number of batches of samples added to the graph, the algorithm finds improved plans. IBBT is effi
nstraint, providing an interpretable measure of robot safety. To enable real-time evaluation of chance constraints, we consider c
our framework is capable of both predicting planner failures and recovering the robot to states where planner success is likely, a
still achieving the efficiency benefits of determinization. We evaluate algorithms for dynamic replanning on a large real-world da
ct a perceptual map, which quantifies the aggregated influence of the environment on candidate viewpoints. This information is
tegy that optimizes long-term rewards while accounting for temporal constraint. We validate the proposed framework and algor
eclude application of existing methods. We first provide an extension of the ASP formalism that captures the unique challenges
proposed in APFT to adjust the planning horizon based on the expected information gain. Simulations and physical experiment
ntered online and preprocess a database of policies. Due to the computational complexity of constructing this database, we pro
e probabilistic perception in a scenario-based path planning approach. A velocity scheduling scheme is applied to the path to en
rs, in addition to characterizing the performance of the planner as a whole. Results show that CC-K-CBS scales up to 30 robots
of uncertainty over time steps. The uncertainty metrics are used to assess tree extensions of the sampling-based search algorit
ating constraints on collision probability as the robot approaches obstacles. Subsequently, we introduce the concept of the traje
kov Decision Processes (MDPs) with a novel state and action space, we investigate how RL can solve it in two paradigms: First

nted space, which becomes extremely time-consuming. In this paper, a topological tree is proposed to simultaneously leverage
WCS. In this work, we establish that computing an LWCS is NP-complete. We further develop optimal and near-optimal LWCS
sband coalitions. While simpler problems can be approached using heuristics or optimization, these methods struggle with more
ions of failures and calls the inner-level search repeatedly to find the optimal task allocation sequence, given the known outcom
ate sector-level paths that consider travel distance and traffic flow equilibrium simultaneously. In road-level coordination, we imp
building consisting of multiple rooms.
que as needed. The implemented prosthetic leg is evaluated with three intact subjects. The experimental results shows that sm
e used to train the proposed CasAtt-UNet to predict bone location automatically and robustly. The ground truth bone locations a
assisting users in various tasks. Through extensive experimentation, our intention-based model demonstrates robust performan

ps in the impedance parameters—a strategy that has previously been shown to improve model accuracy over classic linear in
lation. An AFP prototype is built to validate the design experimentally. The results demonstrate that the AFP mechanism has th
heir movements before and after exposure to the exoskeleton force field were compared. The study focuses both on spatial and
terns from the discharge timings of the motor neuron pools. We realized a real-time implementation of the EMG decomposition
chanism. Taking advantage of the active backdrivability of the D-SEA, we proposed an intelligent assistive strategy that automa
eight at only 50g and can perform functional ranges of motion, including 53° for flexion, 50° for extension, 40° for radial de
adapts to the natural 3 DoF ankle motion and the range of motion (RoM) during walking. The actuation in plantarflexion is evalu
 e developed finite-state machine is validated using data collected from three different measurement systems. Moreover, experi
ment. This real-time online optimization framework enhances the adaptability of exoskeleton users under unforeseen forces the
uction when walking with the torque profile optimized to them and a 46% (range 15%) reduction when walking with the other pro
ew coordination strategy. One able-bodied subject walked using a bypass adapter while prosthesis torques were commanded b
t that employs computer vision to collaboratively offer tailored assistance to the elbow, alleviating joint stress during interactions
els trained with evenly distributed distance candidates. Models are also released as hardware-accelerated versions with a new
different types of cameras that are evaluated with respect to calibration and within a visual Simultaneous Localization And Mapp
eep learning model with 2.25× fewer parameters. Additionally, RadCloud utilizes a novel chirp-based approach that makes o
Value Decomposition (SVD) and RANdom SAmple Consensus (RANSAC). We demonstrate that such representation is sufficie
elf-build robots were conducted to evaluate the methods. The results exhibit that our proposed method achieves higher accura
atic or dynamic tree structures. The experiments show that i-Octree outperforms contemporary state-of-the-art approaches by
sses, which facilitate self-supervised learning on real data of lidar depth correction models. Specifically, the models exploit multi
DAR scanning. The results show that GroundGrid is outperforming other state-of-the-art systems with an average IoU of 94.78
mage captured from arbitrary perspectives. The proposed method is evaluated on both large-scale synthetic ShapeNet and real
d-resolution RMP variants and CHOMP. Finally, the RMP formulation allows the seamless fusion of obstacle avoidance with ad
method for pose estimation, it further aids in navigating unknown environments. Moreover, by integrating navigation with wall-fo
e obtained even though the large sampling noises are used. However, a vanilla MPPI algorithm still does not work in a case wh
sage in non-stationary environments. Extensive simulation analysis on a 7 DoF collaborative robot manipulating a common tool
oth numerical simulation and real-robot experiments in indoor and outdoor environments, including scenarios with cluttered obs
y-to-use calibration method for head-mounted display, improve the effectiveness of human-robot interactions. To assess the fe
traints on principal manifolds, their associated local frames, and the movement primitives that are then needed for the task exe
occer robots. Furthermore, we demonstrated the capabilities of our motion planner during the successful participation in the Ro
igh efficiency and surpasses the performance of other planners. The implementation of our method, including the supplementa
mal solutions than brute force in a significantly shorter time. This allowed us to improve the optimization by increasing the numbe
the wearer, using data gathered from walking trials with healthy participants. Our evaluation method revealed inefficiencies in p
nical transparency, we created a model for calculating the required joint torque, accounting for gravitational dynamics, and subs
r way to optimize such a goal due to the physiological variability among patients, the condition of reversal is based on imitation
e primary aim is to improve the user experience by producing a secure and comfortable walking pattern. To validate the effectiv
he subject's arm, placed on a frictionless supporter, is stimulated to perform point-to-point reaching. By using 90 seconds of rea
robot in a hierarchical order depending on the patient’s performance and their muscles’ fitness. The performance of thi
with a hand, foot, or knee. For each exercise, target locations are customized for the individual via a visually estimated kinemati
fetching environment with a randomly positioned obstacle and use expert or random policies to generate an offline dataset. We
he method's real-world applicability is demonstrated through a safe and effective deployment in a real robot task of box-pushing
. Through extensive evaluation of a diverse array of 59 Atari games, our method not only surpasses the efficacy of prior DRL al
earning. By directly interacting with the other agents, ASRL learns to cope with diverse agent behaviours and maintains the cos
copter drone racing through a track. The experimental results show that our initial state buffer achieves higher task performanc
he performance in terms of either safety or optimality due to strict search space constraints. Through our theoretical analysis of
coding schemes to compress the replay buffer without impacting training performance. We evaluate our approach with three sta
a Sawyer arm.
bust temporal dependency. For the refined features of different frames, we design a bi-level feature aggregation module (BLFA)
system operates on magnification design, and is highly sensitive due to its multifunctional nature involving reflection and refracti
 dal perturbations and query for estimated Q-values as the response signal. Based on ESC, we then dynamically improve the s
e of using gradients for the attack is motivated by the fact that commonly employed federated learning techniques solely utilize
tor and a Single-Agent Motion Planner, MAexp can work with arbitrary numbers of agents and accommodate various types of ro
distribution of the offline dataset. Given this, we collect data from the inaccurate simulator starting from the distribution provided
odule, i.e., adopting adaptive adversary, the policy is then able to handle extreme cases when the velocity of the drone is imme
ments. Our empirical evaluations show that ReactHD achieves robust and accurate learning outcomes comparable to state-of-th
on to train our flexibility estimation module. During the manipulation, the robot interacts with the DLOs to estimate flexibility by a
ile avoiding task failure due to the unnecessary event-detection query. Our framework simultaneously learns 1) an embedding
ce how we interact with objects. In MOSAIC, we leverage this profound insight by distilling knowledge from multimodal foundatio
s such that they do not block the robot path homotopy to the target object, thus also aiding an underlying path planner in quickly
of MPI and show that MPI can synthesize high-quality novel views of complex scenes with diverse camera distributions and vie
es of geometric collision checkers. We validate our approach in simulated and real-world robotic experiments, and demonstrate
ulti-view surface sampling, and utilisation of volumetric signed distance functions (SDFs). We showcase our approach on a new
o randomly occlude some local patches of point clouds and establish the supervision via inpainting the occluded patches using
y appropriately combining the outputs of off-the-shelf pre-trained models, with no new training. On various simulated and real r
t learns to generate SKT by deforming a template trajectory based on the task-relevant geometric structure features of the sup
mentally collect task fingerprints, and demonstrate the usage of the representation for task classification. This representation mo
sence. Our SlotGNN, a novel unsupervised graph-based dynamics model, predicts the future state of multi-object scenes. SlotG

™s relative pose transformations. This approach enables us to incorporate information from the robot’s perspective instead
during training. Taken together, these results suggest reasoning about the 3D scene through point clouds can improve performa
environments with fine details. HIO-SDF combines the advantages of these representations using a hierarchical approach whic
d choices of generic pre-trained vision transformers PVT, we demonstrate in each case that policies trained on SOFT(PVT) far

bsequently, we arrange the catego- rized points using an octree, followed by an in-depth analysis of nodes to finish multi-resolu
pth values. By predicting for normals instead of depth directly, our method allows for robustness to changes in depth distribution
geneity. Experimental results on the public datasets of LineMOD, YCB-Video, and Occ-LineMOD show that the proposed meth
curate and efficient. The experimental results on benchmark tests show its superior performance.
into existing methods, showing their effectiveness.
fic vantage points. Additionally, MMAUD provides accurate Leica-generated ground truth data, enhancing credibility and enabli
to small patches based on geometric cues and learns to merge the patches in a deep agglomerative clustering fashion. SupeR
phere. To illustrate the effectiveness of our method, we apply MSMFormer to unseen object instance segmentation. Our experi
sing the sensor motion and design a novel loss to ensure temporal consistency. We evaluate our approach on the SemanticKIT
rate contrastive learning into different curricula. By drawing reference knowledge from the previous course, our strategy establi
ovel spatiotemporal graph neural network based on heterogeneous graph learning, which analyzes spatial and temporal connec
motion of all dynamic objects. We assume a correspondence-based setup where we have putative matches between the two p
seven different DLOs, we demonstrate that the proposed method is an effective and robust 3D perception pipeline solution with
iable rendering pipeline, which simulates the transient imaging process, allowing direct optimization of scene geometry to matc
's representation with a VLM learning a shared visual and text space from training on large-scale image-text pairs. By transferr
ion. The resulting representations generalize to novel semantic classes, without the need to collect large 3D datasets or finetun
roved real-time DLO tracking under mid-section occlusion, tip occlusion, and self-occlusion is demonstrated experimentally. The
 ularize the training, it is sensitive to hyper-parameter choices and has not been yet explored for the 3D domain. To address the
”comprising links, joints, and a gripper—can fold into a flat state, deploy with precision and minimal weight, and effectively man
more compact form-factor it would be difficult to densely lay them out on a roof without the trackers substantially shading each
d sim2real transfer to physical machines in future.
ally characterized. Moreover, we conduct experiments to demonstrate the capability of the gripper in two working modes. The g
n. One such robot arm for which the conventional singularity analysis approach fails is the new Link~6 collaborative robot by Ki
vironments to test how it was able to function outside of laboratory settings.
erated the robot, untethered, in various environments to test the robustness of the design and analyze its motion and performan
s. A CAD model of the robot and computer animations are provided to demonstrate the large workspaces and the gripper mech

perimental results that provide an example of effective cross-robot policies. We assemble a dataset from 22 different robots col
on large point clouds. We complement our results with the rigorous mathematical analysis providing deeper insight into the phen
ates the potential of soft optical tactile sensors for dexterous manipulation and grasping.
ic manipulation. To the best of our knowledge, our method is the first approach leveraging large-scale multisensory pre-training
n-conditioned robot manipulation policy, and allows following humans plans for generic manipulation tasks in a zero-shot manne
optimizers, and techniques that plan through convex inner-approximations of collision-free space.
class of CBFs that we term fault tolerant neural control barrier function (FT-NCBF). We derive the necessary and sufficient cond
es and varying sensing conditions. Our results show improved safety compared to traditional state-based approaches while allo
on of the algorithm is indeed control-invariant. Furthermore, we prove that the system is indeed able to move from any state wi
utationally efficient, which makes it applicable for real-time robotic control. We validate our approach, first using a 2D mobile ro
ve, physics-aware regularization terms are designed for better results, including trajectory smoothness and elastic potential ene
y satisfy safety constraints. We include theoretical arguments proving our approach generates reachable sets that bound the fu
nvolving the humanoid robot Digit. Three locomotion controllers are examined, including a commercial one, using a walking-in-
wns. The anomaly detector highlights issues that transcend individual modules and pertain to the safety of the overall system. W
es of a quadrotor drone, which is subject to complex aerodynamic interactions between the aircraft and the environment or whic
anding pad. Experiments show that with the learned IPC, the control algorithm safely landed the quadcopter despite the error f
e completion of each exploration run, with the robots avoiding collisions with each other and the obstacles. We test our approac
ve Control scheme that can guarantee recursive feasibility and/or safety under weaker assumptions than classic methods. In pa
his allows us to learn directly the set boundary, therefore learning in a smaller dimensional space. Compared to the state of the
potential scene changes occur. We test the system in simulations and real-world experiments to demonstrate the impact of sem
rative robots can be increased by a factor up to 3 only by changing the conventional servo-geared actuator architectures for MR
rce implementation as a ROS2 package and evaluate the solution in simulations and hardware experiments involving high-dim
on of prediction sets. Experiments demonstrate the utility of our approach in robot motion planning around humans and robot m
d by using multiple backup controllers, determining when to switch is an open problem. Consequently, we develop a broadcast
g mechanism that leverages the cross-context importance of nodes to pool the graph representation. By repeating the feature e
method is enabled by three key contributions. First, we simplify the problem of learning a transformation to the training data den
d on our new permuted-QR (P-QR), which fully utilizes and properly maintains the upper triangular structure of the square-root
s novel approach ensures robust tracking by leveraging location priors from historical boxes, even in scenes with sparse points
f explicit, dense, and interactive surfaces. Experiments were conducted on both synthetic and real-world datasets to evaluate th
ntation. Operating purely on visual demonstrations without auxiliary information, UVD can effectively extract visual subgoals em
apple and one suction cup on top. Even with ±15mm offset, at least two suction cups engaged with the apple 80% of the time,
 rawberries with different heights and angles. Additionally, we propose a strawberry weight estimation method based on plane-c
rnels. The framework employs deep convolutional neural networks to process these captured images to classify the kernels as
on GCPs, at the beginning of the season. The method utilises LoFTR, a state-of-the-art image-matching transformer. The orig
better than state-of-the-art computer vision-based methods) and R2=.29 for HC vines (not previously attempted with computer v
as this allows us to alleviate the influence of segmentation errors on the phenotypic trait estimation. To achieve this we propose
monstrate generalizability of our method on previously unseen trees. Videos can be found on the project website: https://kantor
y in the face scenario, the sum of changes in distance, calculated using the proposed algorithm was negative, meaning that traf
all leaves allows us to count them and compute relevant traits as their areas, lengths, and widths.We propose a novel self-supe
ign and integration for autonomous leaf retrieval with a mobile manipulator and a custom-made end-effector; and 3) learning-ba
Furthermore, the cumulative error compensation reduces the time for the harvesting process. The experiments were conducted
scene parsing, we propose specialized pre-training using a public benchmark dataset for fruit transfer learning. By leveraging pr
dling very thin and delicate branches. We also address an overlooked problem of connecting skeleton points in the form of a gra

ategies which enhance the transient performance during flight mode transition. Then, the stability of the proposed control law is
ve model in the form of a Conditional Variational Autoencoder (CVAE) from expert demonstrations to warm-start our optimizer.
ome modeling uncertainties and external disturbances. The experimental validation is carried out with a UAV that navigates from
ene graph builder for agricultural environment. We first propose a structure of the hierarchical graph for agricultural environmen
ce of the semantic environment representation SEEREP. We further present that the maximum sensor data rate which can be s
e precise cutting. An end-effector grasping pose control algorithm of the manipulator is proposed, which can control the end-eff
ol and inverse optimal control are applied to transplanter. The experimental results demonstrated the effectiveness of these pro
d while the robot is moving over the plowed field before the planting process starts. In order to achieve the best performance, w
ork for a fully autonomous system in the future. We validate the performance of our system through field trials in a sweet cherry
surrounding trees and jointly estimate their spatial and vegetation-based characteristics. Realistic simulated tests are used to ev
ng, a robust estimation and graph clustering approach is introduced to associate fruit detections across images. Through simula
ffer from high computational costs, ineffective view selection due to poor candidates, or non-smooth trajectories due to inefficie
ing, it arises during extraction from the growing tube. We present experimental results to show that using RGBD camera image
bjects seen in one dataset while observing them in conjunction with similar features from the disjoint dataset. This allows the ne
desired new field or robotic platform. Additionally, we propose a label refinement method to generate labels corresponding to th
ary elements to achieve a completely autonomous loading cycle, where the crane picks up and loads all logs within its reach on
ference, the Extended Kalman Filter (EKF) can modify the shifting between the GNSS signal and the true waypoints of the mow
manual measurements collected by foresters during field trials. Our experiments show strong segmentation and tree trait estim
diately exploited for future plans. The performance of the proposed method is evaluated against synthetic data and is shown to
n with other robots within the sensing range are the only requirements for the robots to effectively surround and herd the anima
orces of the connectors being lower-bound. The results are also explained in the supplementary video.
rage for both simulated and real-world experiments while being 50% more efficient in terms of trowelled distance.
lattice material, and planning and controls methods. We demonstrate the system through a set of fundamental assembly tasks
al-world dataset captured using a LiDAR system mounted on a quadruped robot monitoring an industrial facility containing static
ware experiments and simulations over large datasets show that the proposed technique is scalable to large areas.
woodchips for an unknown gravel material, requiring 40 bucket fillings. We also proposed a novel reward function for the materia
y varies according to the task parameters. Correspondingly, we develop a framework for learning a policy that can be generaliz
5i camera for localization, an impact-echo (IE) sensor, and a ground penetrating radar (GPR), to perform automatic NDE data
n. The results of action execution in environments provide feedback for subsequent coarse-resolution reasoning. Such coarse-t
o implement this solution. Grasp-Anything excels in diversity and magnitude, boasting 1M samples with text descriptions and m
evaluate our framework's abilities in realistic scenarios in the VirtualHome environment and demonstrate a 31% reduction in exe
at PCBC is able to achieve strong performance in a variety of few-shot regimes, often achieving task generalization with as little
ons while guiding the robot to perform the required actions. We also use a fine-tuned Llama 2 model via self-instruct and compa
o guarantee optimality, we design an LTL heuristic function that is provably consistent and supplements the potentially inadmiss
petitive executability. Our improvements transfer to a Boston Dynamics Spot robot initialized with a set of skills (specified in lang
GPT outperforms existing TOG methods on different held-out settings when generalizing to novel concepts out of the training se
formance metrics for system optimization. DOS significantly increases the reliability and maintainability of the deployed robotic
op is guaranteed through Lyapunov theory. Finally, simulation results demonstrate the capabilities and performance of such a c
tive law based on the Lyapunov synthesis method. The tracking and adaption errors are shown to be uniformly ultimately boun

ipatory computation stores optimal control based on current nominal states, thus mitigating the delay effects. Additionally, we e

ing limits. we present a unified framework that not only learns the tire model in real time from collected data but also adapts the
omputational complexity as an EKF and can be used for real-time control. The paper evaluates the risk-sensitive behavior of the
ntly. Finally, simulations of the proposed control on TSS is shown to validate the theoretical analyses.
to be applied to any kinematic-controlled robots under practical assumptions. We assess our approach through experiments on
ults demonstrate the capability of the integrated impact safety module within the proposed digital twin approach to limit the velo
ions surrounding a reference trajectory. This involves learning a stable DS on SO(3), extracting time-varying conic constraints f
e a batched, fast and precise distance checking method based on precomputed link-local SDFs. Our method can check distanc
o calculate shoulder-elbow angles is applied. Secondly, we extract the refined object's segmented mask. Thirdly, we compute tw
r to incentivize legible and proactive behaviors and an inconvenience budget constraint to ensure equitable collision avoidance
present a Stochastic Model Predictive Control (SMPC) for the autonomous vehicle to safely interact with the human-driven veh
y on key input features. This approach significantly reduces computational time while preserving a similar level of accuracy. Exp
ernal C^2 collision boundary functions. We theoretically prove constraint satisfaction and show that the robot is passive when f
ring adversarial samples. Our method improves the speed of model training on CIFAR-10 by 68.67%, and compared to other d
g UHRI that utilizes three-dimensional reconstruction from stereo image pairs and machine learning for localizing human joint e
ere recently field validated onto LineRanger, and mechanical design insights shall be presented for each module.
head nozzle unit posture against the acceleration of the mobile base. Then, the robot parameters were identified from tensile tes
ses that, taking as input the area size to be explored, the robots fleet size, their energy profile, exploration rate required and tar
ed exploration methods, the proposed approach efficiently re-uses sub-graphs built in previous iterations to construct the globa

space, respectively. The proposed approach is empirically evaluated in a realistic scenario, and based on the reported results, i
ists of a methodology to interact with a remotely located human operator based on an inexpensive long-range and low-power c
ntial step forward comparatively to classical perception systems in perceiving objects in these scenarios, but escape the burde
gion labels at each location. This mapping procedure is paired with a trained navigation policy to enable autonomous map gene
data and can generalize to novel 3D scenes and arbitrary text queries. We evaluate LLM-Grounder on the ScanRefer benchm
e from an image using a one-class classification algorithm. Additionally, we supplement the limitations of self- supervised labels
ouds since the size of an object is distance-independent. This can be addressed in cubic voxelization which has a fixed volume
 ed with only ten annotated images, predicts high-quality pseudo-labels that can be used with any existing panoptic segmentatio
semantic segmentation network especially suitable for handling low-light scenes. Experimental results demonstrate that the pr
ving efficiency by adaptively modeling a 2D key/value set for a given 3D query, thus enabling multimodal fusion with higher flex
ular RGBD SLAM backend, and uses these poses to progressively optimize a Language Embedded Radiance Field (LERF) for

ents likely to belong to the same semantic classes, relying on surface normals. We also describe a novel neural network design
method is lightweight, resulting in the method's online execution being both rapid and computationally efficient. Moreover, the p
nism. In particular, Open-Fusion delivers outstanding annotation-free 3D segmentation for open-vocabulary query without the n
metric properties of each semantic tracklet in the sequence. In an in-depth study, we discover that it is critical to promote spatia
with their semantic classes and instance IDs that are consistent over time without any post-processing like clustering or associa
gocentric semantic map-based Local Scene Priors (LSP) and the commonsense relational graph-based Global Scene Priors (G
e of 3D scene graphs exists for the training of the novel CECI model, we present a novel methodology for generating a 3D sce
-world underwater data that the proposed GGDRF outperforms the previous GDRF method both quantitatively and qualitatively
rk SNE-RoadSeg. A key contribution of this paper is a strategy for sharing encoder representations that are co-learned through
cture to learn discriminative prototype features, we design a triplet contrastive loss to increase the compactness of the graph. W
ing. Our instance transformer head incorporates essential information to enhance segmentation but also enables reliable, class
aining real-time capability and adding only 525 learned parameters to the pipeline. We further illustrate the importance of map c
ical fusion module, which includes feature-level fusion and BEV-level fusion. The feature-level fusion, composed of a mask gen
ment and classify objects even when one modality is partially available. Also, the proposed self-distillation loss encourages the
agents and their panoptic observations of other vehicles. Based on the connectivity of the lane graph, we segregate the environ
BEV fusion between the spatially synchronized features, we suggest ColFusion that applies self-attention weights of LiDAR and
uently employs a query-based method for low-latency updates of prediction results to the grid map. Finally, based on the updat
d the scene. mmEMP first devises a dynamic 3D reconstruction algorithm that restores the 3D positions of dynamic features. Th
nctions with the CARLA simulator. To show the correlation between detection performance and our surrogate metrics, We cond
s. Buttressed by it, we first introduce an Event Temporal Attention (ETA) module to learn the high temporal information from ev
r one of the single fusion schemes. Second, we distillate the 2D knowledge from a 2D network (Camera branch) to a 3D netwo
tral Unit (GSU) is also proposed to combine information from both modalities. SK-Net aims to extract valuable semantic informa
ts show a significant performance improvement in all tested baseline models. In fact, benefiting from the plug-and-play charact
he fusion of LiDAR and camera data at multiple levels. We use LiDAR depth to improve image depth estimation and use image
formance over state-of-the-art discrete-time schemes. We release the source code and own experimental data at https://github
entation and a comprehensive land vehicle dataset. A factor graph optimization-based multi-sensor fusion framework is establis
vents using inertial measurement unit. In spatial reasoning stage, we project the compensated events into the same image coo
ch fine-tunes the fusion of LiDAR and camera details; and the Volumetric Grid Attention (VGA) for a comprehensive spatial dat
ar points and image patches, and local scale refinement of dense depth using a scale map learner. Our proposed method signi
tion capabilities with semantic information from a robot's perspective. Besides RGB videos, semantic annotations, and depth in
hat our approach surpasses recent methods in terms of performance on the synthetic dataset Flythings3D, as well as generaliz
ed calibrations using a camera and a sparse multi-beam LiDAR or one-dimensional LiDAR.
ropose a novel Spatial Reliability-oriented Fusion Network (SRFNet), that can estimate depth with fine-grained structure at both
n both simulations and experiments.
corrupted representation of the world. Regarding this problem, we propose a method to estimate the angular velocity using ac
rchical attention mechanism adaptively. Considering the violent shaking on robots and rugged terrains, a lateral Human-ware P
 iability of temporal encoding on depth information and modality fusion, even on differently encoded data, to be beneficial to net
rameter design is divided into impact reduction and contact impedance control. There is no abrupt alteration in the contact force

parameters and surface normal. Simulation results are provided to discuss the robustness of the task with respect to these mo
sponse. Experiments show our acquisition system's effectiveness and performance improvement for VO tasks in complex illum
ls. We evaluated the proposed MagicTac in the tactile reconstruction task using the deformation field and optical flow. Results i
n network (HiFTA-net). This allows it to effectively learn from various situations in multiple datasets, including those involving si
with the Robot Operating System and physically demonstrated on Turtlebot3 Waffle Pi. It is observed that the speech recogniz
es a given calligraphy glyph to obtain a coarse stroke sequence. Using an RL algorithm, a simulated brush is fine-tuned to gen
ural network to learn how to deform a template. Our experimental evaluation demonstrates that our approach yields accurate e
of our approach.
es the development of a procedural terrain generator to create diverse training environments. We present a series of experimen
ally confirm that the system with the self-recovery mechanism can accomplish tasks by resolving various failure cases.
ntroduce an enhanced loss function that translates the voxelized quilt state into a more representative one. Both simulation and
pondence and strong modeling ability from neural network. Other techniques including realistic data generation, feature clusteri
ical IBUVS algorithm when facing large initial camera pose offsets under the calibration disturbance. Supplementary material in
ce. The prototype robot can generate an online summarized report using its integrated language interaction and IoT-based vita
ematics and statics are verified. The results show that the cable-driven spherical parallel continuum manipulator for nursing rob

o model task-specific perception processes. The framework can combine state-of-the-art computer vision methods to satisfy the
e morphology design is evaluated by simulating the robot navigation in an environment of interest. Generalized pattern search,
he context of a simulation and two intricate real-world scenarios: blocks world, barman and pizza preparation.

e observations and employs a generator-discriminator architecture to closely mimic pedestrian behavior. We also propose a se
ion in C++, and rigorous thread pool utilization, MuRoSim achieves high computational performance. We apply MuRoSim for tr
lation of rigid and deformable bodies with focus on usability, stability, robustness and easy access to state-of-the-art deformatio
accurate representation of hydrodynamics in fish swimming. In contrast, the idealised travelling-wavy undulations tend to apply
ck instance segmentation to show the effectiveness of our simulator for image-based perception. Trained on our synthetic data,
e-art. We also illustrate how SceneControl can be used as a tool for interactive traffic scene generation.
ntiable physics simulation over a variety of contact-rich tasks. Supplemental materials and videos are available on our project w
meters is presented. To bridge the sim-to-real gap, a pipeline is developed that utilizes a Motion Capture system to compare the
esign a keypoint detector and descriptors-assisted LiDAR odometry and mapping framework (KDD-LOAM), whose real-time od
Campus Map, a novel large-scale multi-camera dataset with 2M images from 6 mounted cameras that includes GPS data and
world sonar datasets, showing that it outperforms the existing geometric-only method on localization accuracy and achieves stat
isation problem tailored for CURL. Experiment evaluation shows that CURL-MAP achieves state-of-the-art 3D mapping results
eracy, partial information to be conveyed to the graph along the non-degenerate axes. The proposed method is evaluated in rea
r method in terms of accuracy, robustness, and run time on indoor and outdoor datasets that include challenging real-world sce
cy. Our experimental results on both public and self-collected datasets demonstrate that Ground-Fusion outperforms existing lo
ts validate the effectiveness of our hybrid approach, demonstrating its superior performance over existing implicit field-based m
ation and detailed object-level mapping, even in dynamic and illumination-varying environments. Our evaluations, conducted on
e from all the landmarks. The second step marginalizes the pose of interest with pure pose-wise constraints without affecting a
point changes. The utilization of dense local maps allows us to reduce the complexity of features describing these maps, as we
nage the map in voxels and update the distribution of points in each voxel incrementally while maintaining its consistency with t
using the View of Delft and NuScenes datasets, which involve 3D and 2D radar sensors. Our findings illustrate state-of-the-art p
consistency of each loop closure measurement through a consensus-based procedure, possibly applied to a subset of the glob
uring the early stages. Our second technical contribution is the integration of a patch descriptor, which extends the applicability
irstly, we introduce an improved observation model and a novel data association method for dual quadrics, employed to repres
litating its adoption on edge devices without compromising accuracy and performance. The crucial innovation lies in the multi-fe

ent the degeneracy issue of camera pose tracking in recovering the yaw component of general 6-DoF motion, we introduce as a
obot’s pose and its map, simultaneously constructing and leveraging high-level information of the environment. To extract t
m types regarding input, data processing, and output. The tutorial also discusses the subtleties behind the evaluation of VPR a
motion because of the discrepancy in the FOV and resolution of the radar sensor. In this paper, we address Doppler velocity me
he implementation of our algorithm can be found on Github.
he-art methods on several public datasets, including TartanAir, EuRoC and ETH3D-SLAM. Specifically, when dynamically fusin
ry estimation. Furthermore, we show that the proposed approach is also suitable for large scale keyframe optimization to increa
to retain prior scans' constraints by exploiting the local support property of the B-spline. To further improve the odometry accur
s function converges faster and leads to more accurate geometry reconstruction than other loss functions used in depth-superv
his way, we can substantially reduce the parameters to tune for a given type of environment. The experimental evaluation sugge
alidate our system's performance through extensive experiments on various challenging sequences. Multi-LIO emerges as a ro
thorough analysis of various Dynamic SLAM formulations, identifying the best approach to address the problem. To this end, w
her plane. After merging, the father plane 3DOF representation will be more accurate than the kids plane and the uncertainty wi
by a compact set of parameters, the point cluster coordinates. We derive the closed-form derivatives, up to the second order, o
ding Horizon Sampling (RHS) mechanism proposed by us. Then DORF gradually recovers more ambiguous static points, guide
re, performs fast finding of voxels observed by new scans, and incrementally reconstructs triangle facets in each voxel. To the
formance criterion. While this formulation is NP-hard in general, we show that a combination of greedy sensor selection and fas
e data association within the same class of features. Moreover, we design a Hybrid Bag-of-Words (H-BoW) model to deal with h
with multiple initial values, reflecting to the uncertainty of an initial value. We formulate an invariant extended Kalman filter (EKF
is algorithm determines the appropriate number of edges since this is a trade-off optimization problem. Second, the edges of th
solutions as evidenced by the improved tracking accuracy on unseen test trajectories.
g collaborative SLAM systems. However, this requires more than a single robot and does not include mapping in the mentioned
the performance is still guaranteed and even improved, thus aligning with the "less is more" principle.
. We expand the application range of L-SLAM while preserving its strong performance, even in low-textured environments, usin
SOTA) keypoint and feature extraction algorithm, we evaluate our method as well as SOTA place recognition methods using d
centric place recognition. This allows loop closure searches to be concentrated to semantically relevant places. Our proposed s
s a solution to the challenges encountered in directly estimating relative poses between cameras, particularly in adverse enviro
to one another. Experimental validation demonstrates that the approach achieves state-of-the-art pose accuracy and furtherm
ven approach. The result demonstrates the superior performance of our data-driven approach when compared to existing meth
ng modes in view path planning, to reduce time and maintain superior reconstruction quality. Our method exhibits the highest re
vel times. The effectiveness of the method is illustrated through on-device testing in various floor plans.

uiring any grouping or approximation in the estimation state. It uses continuous-time trajectory estimation to maintain the temp
ansients upon state disturbances). Our approach is inherently consistent even during convergence phases from significant erro
ected position and prior pose from the odometry, we build a graph optimization model to optimize the pose of vehicle. Through
olicy. We evaluate the approach both on the ROS navigation stack with DWA and a DRL-based navigation policy, which identifie
. We propose an approach that leverages an LLM and vision language model (VLM) to help a robot actively perceive its environ

s of a robot with extreme personality traits. In conclusion, we deliberate on potential future research directions and the associat
he integration with RNN, the temporal graphs utilize past trajectory information to infer the future intentions of each agent. The s

n-human versus human-robot interactions. The results underscore the challenges and potential of designing social robots that a
ur results show the potential of the system to drive children's interest in communicating, sharing, and discussing cultural theme
osed ChatAdp, we use ChatGPT as a user simulator to provide feedback. We evaluated ChatAdp in a case study for context-aw
romising potential for advancing various fields, including wireless communications, medicine, the Internet of Things, and intellig
ments. Simulation results show efficient trajectory planning amidst multiple dynamic obstacles and provide a robust performanc
onse of the gripper when programmed to different states. The results show that in the original state, the gripper is capable of ca
the beads into the actuator after the active deformation. To reduce the driving pressure and facilitate the flow of the beads, an
g and research in this challenging environment. The binary actuation system that results from this combination, is modular, sca
), as well as novel material transport and locomotion capabilities.
agent to adhere to the symmetry. Results in simulation with five different symmetric peg shapes show that our proposed agent
ation model and then shows the hybrid force-position control using the presented model. From the results, it is concluded that f
ytunnels, and to evaluate the shelf life in comparison with berries picked by professional human pickers. Our results with ten ex
an move at an average speed of 3.85 mm/s on horizontal, vertical, and inverted walls and make transitions between walls with
raflex’s locomotion into predictable steps on a triangular grid, simplifying odometry and allowing the use of path planning to
perated and depth-compensation-free; 3) in result, the manipulator could be modelled kinematically in a simplified way for posit
skin (E-skin) to acquire information from grasped objects. With the built-in E-skin, the gripper achieved object recognition at an
o. To perform the critical unlocking task, an end effector with the function of loosening and unplugging the aerospace connecto
ns substantiate that the pupal-mode actuator has the capability to reduce resistance and enhance the safety of the endoscopic
, as well as shape-morphing of the body, enables the robot to navigate obstacles as diverse as confined spaces, inclined surfac
cting full ripeness and a standardized harvesting sequence for effectively picking berries. In an outdoor harvesting experiment,
ody of the fish and is also an independent energy control system. The electronic components are evenly distributed on the doub

a layered laser-cutting technique, such that the actuator can be tiled into an array to control the rotation of many antennas. We
various palm configurations. The resulting robotic hand is slightly bigger than a baseball and represents one of the first soft rob

ies, and the increased area of contact, allows for adaptive grasping on asperities down to 0.4 mm with an adhesive force of up
D structures from flat plates. To demonstrate the capability of these techniques, we built and tested an origami crawling robot a
ks, such as transporting objects, separating or assembling miniature parts, and operating customized tools. Furthermore, we su
k that evaluates the electromechanical behavior of the DEA. A graph neural network (GNN) is employed as an encoder to extra
combined twisting-bending-elongation deformation by lattice superimposition. The proposed structural design method paves th
n in three-dimensional space. The performance associated with the designed soft hybrid-driven manipulator is experimentally ve
st the hand in conjunction with a low-cost vision-based teleoperation system on different tasks.
thods, demonstrated the glove's superior extension capabilities using a model hand with realistic stiffness. Furthermore, the glo
d soft twisting actuators. Additionally, the composite-like SFTTA allows mechanical property programming through the alteration
procedures. This mechanism enables rapid and reliable retraction of soft growing robots without sacrificing their inherent advan
s the introduction of an extremely anisotropic, composite, wrinkled film with elastic moduli 400x different in orthogonal directions
quired when adding a new module. It resulted in a manipulator with continuum soft actuators, but independent variable stiffness
of the proximal segment when operating the distal segment, reducing deviations by at least 70.11%. Operation is validated wh
ctive mapping between signals and contact points by considering time series. In addition, we propose an azimuth prediction los
gripper can estimate an object's weight and location and offer new insights into grasp parameters like friction coefficients and
ent accuracy during the synthesis thanks to transfer learning. Through two case studies, the capacity of generating geometrical
r designed an array flexible force sensor, and studied the structure, signal acquisition and preparation process of the sensor. F
per. Simulation driven optimization methods are leveraged, utilizing quasi-static and dynamic finite element methods models, to
re’s layer resulted in an increase in stiffness, aligning with the analytical model’s predictions. At the most rigid stage (0
5 N. The sensing capability of the retractor is validated on an in-vitro model to demonstrate prediction errors of 0.06 N and the p

EMGs, the accuracy of continuous grip force control based on the optical waveguide sensor is higher. The R-Square for prosthe
ting at a known arbitrary location. Our model-based approach estimates soft body deformation with approximately 3% average
ipeline from human demonstration via the sensor used as a wearable device. This approach of using tactile sensing as a mean
MC) is proposed to minimize the viscoelasticity of DEAs. Its stability is also proved mathematically. The experimental results ob
nique deformation profiles in an actuated soft-bodied robot. We introduce a Convolutional Neural Network-based tip force estim
f the actuator based on image sensing from optical fibers packaged with the jamming fibers. The tip position sensing shows ac
visible light. Our simple design concept advances the capabilities of bio-inspired robots and lays the foundation for future “
mization to navigate the high-dimensional design space. Our design framework can readily regulate the surface quality during th
ressure, which in turn powers a fluidic soft actuator. Experiments were conducted to evaluate the response time of the hydroge
er than 2 N, and various locations across the sensor, as well as multiple indents spaced 2 cm apart. Furthermore, the sensor is
d as synchronized real-time threads on a central Linux-based multi-core computing unit. That unlocks additional computational
e Robot Operating System (ROS). Our tooling suite covers the automated minimal containerization of ROS applications, a colle

ealism is further increased by simulating relevant gas and wind sensors. Scene generation is simplified with the introduction of
tive actions that are executable by a manipulator in a kitchen environment. Cook2LTL makes use of a caching scheme that dyn
e code in a simulated workcell, and debugging syntax and control errors, such as collisions. We outline the architecture of this s

nge and improve outdoor robot navigation. Leveraging the multi-scale nature of the costmaps supporting Nav2, we developed a
he reliability and detail of lunar traverse planning and topographic reconstruction, while providing an estimate of the uncertainty
a resilient control framework tailored to the zero-gravity environment. Simulations and experiments in the Zero-G Lab validate
used to design the centralized trajectory planner that, based on the relative location of the target debris at the launch of the ne
ctory images show that OST can improve the NN performance on the target image domain given that OST is performed on ima
sk where a Robotiq 3F gripper is required to navigate towards a free-floating moving target, pursue it, and subsequently positio
. Moreover, the characteristics of the event sensor offer several advantages in space applications compared to RGB sensors. T
d with such analyses. It leverages a multi-resolution strategy adaptive to the range errors common in stereo vision systems, ma
within 6 hours (4 orbits). This performance substantially surpasses that of ground radar, demonstrating significantly faster and
and robust controllers. In both cases, using the identified parameters results in significant robust control performance enhance
 oposed, which innovates in environment map construction and locomotion planning. Specifically, the closedform Minkowski sum

designed for a robot driven by an internal 2-DOF pendulum. In addition, it presents models for controlling the pendulum and und
lication scenario.The maximum determined static load that can be carried by the further developed EMI is 2000 N. In further ex
zes a single-motor controlled tendon-driven mechanism mounted at the end of a robotic arm equipped with controllable stiffnes
th six retractable shell-shaped legs on each side and this innovative design allows both wheels to independently change diame
ty to SMS. An LPV plus Hinfcontroller is compared with a Model-Based PD, and a Model-Based PD plus Hinf controller, in the p

elocity controller to manipulate the grasped object to achieve a desired shape. To validate our approach, we conduct comparativ
The significance is that we demonstrate that high-performance bilateral teleoperation is feasible in clinical systems, without har
contextual video features. It achieves the real-time performance of 1.1-1.3ms for processing a 1-second input window by relying
ow that our method achieves the highest performance with a 91% average success rate.
ation to real and show robust execution of our vision-based policy with a 96.3% success rate. We analyze and discuss the succ
graphics software have limited tools to generate images depicting realistic instrument-tissue interactions. To address these limi
ection, previous works have focused on making existing algorithms more efficient. In this work, we focus on making the simulat
cators in order to enhance temporal reasoning. Moreover, considering to exploit a general understanding of dynamics inherent
navigating through narrow corners that are inaccessible to fixed-length bronchoscopes. It can also be shortened to increase its
our methods.
ased on the reliability of the estimates. Our approach, Ada-Tracker, enjoys both short-term dynamics modeling by capturing loc
-sim gap between a physics simulation and observations obtained 3D perceptions through estimating a residual mapping and (
erstanding. We demonstrate the performance of our proposed method by comparing it against 3 different methods in several a

OUO (Leave-One-User-Out) and our proposed LOTO (Leave-One-Task-Out) cross validation methods to assess their ability to g
hanges in geometrical and biomechanical boundary conditions.
. We demonstrate ORBIT-Surgical sim-to-real transfer of learned policies onto a physical dVRK robot. Project website: orbit-sur
ted least squares is applied to solve RKHS-based formulation efficiently. Experiments indicate that the proposed algorithm proc
ion-only data, in this work, domain randomization is proposed as a strategy for translation to hardware of a simulation policy wi
s a function of actuation and external loading. We experimentally validate this kinetostatic model on prototype CPPRs with two
. We validate our approach through experiments in artificial isotropic and inhomogeneous animal tissues, establishing ground t
e catheter tip to enable real-time catheter localization. To enable improved localization accuracy, a novel nested calibration alg
oller and navigator. Valves can be autonomously delivered with a 2.0±0.5 mm position error, and minimal misalignment of 3.4Â
ed 2D US - 3D US registration is proposed. To achieve fast and robust ultrasound probe calibration, a method based on an imp
measurements along a spatial direction delta s. This is sufficient for an agent to choose the best next action. We validate our a
AVs, and the sensing of the source happens only at the individual level and is scalar, making it a viable option for UAVs with lim
ated, serving a twofold purpose: (i) to study the interaction of multiple, possibly heterogeneous swarms, and (ii) to enable a grad
acterized the mobility of Morphobot to test its capability of moving and physically interacting with its neighbors. To demonstrate

he control barrier function method to optimize the non-reciprocal interactions for a desired spatial formation. This work address
e pre-programmed robots that populate its environment. We use automatic modular design and neuroevolution to produce the c
nterface to run external scripts that would be run on a ground station during hardware deployment. The plugin simulates a radio
sed Habitat transport task and on MultiOn benchmarks show that our method outperforms baselines and prior works. Further, w
utperforms baselines in terms of cumulative task planning time by an order of magnitude. You can find more supplementary ma
ntrol framework in geometric algebra that was derived in previous work. In the experiments, we demonstrate how to model diffe
eactive capability to cope with online uncertainty and external disturbances in the real world.
raveling salesman problem with precedence constraints (TSP-PC) is formulated by grouping frontiers and object proposals. The
such trees means to decide on where to invest compute or where to widen a branch. We propose a heuristic to balance branch
valuate the impact of both plume dynamics and gas sensor properties on the GSL performance. Secondly, we demonstrate that
actions in the continuous domain, hence, it reduces replanning by securing feasibility at the final full path trajectory optimizatio
f auxiliary subproblems in sequential manipulation planning.

historical RGB observations, bounding boxes, sub-goals, and high-level instructions, by effectively fusing modalities via our Mu
the Belief-Desire-Intention model we compared with.
lanning efficiency of the proposed algorithm, outperforming the synchronous approach in terms of makespan.
sk involving three Jackal robots, highlighting the approach's potential applicability to important real-world problems. Our benchm

c model is updated alongside. The probabilistic model is used to generate a new task plan if the feasibility of the previous one i
r proposed method for varying team sizes and in different environments, and we compare our solution with four different baselin
ion compared to a two-stage decoupled approach, while still maintaining a success rate of 100% for the task.
ide a full digital twin as well as rudimentary baselines to enable rapid progress. Our vision is for RAMP to form the substrate for
e provide the initial datasets on injury prevention for the human hand based on drop tests with surrogates for the human hand,
eld an affordance-directed, lazy-search planning algorithm, which can seamlessly compose strategies and actions to solve long
roach on long-horizon tasks in CoppeliaSim and show how learned action contexts can be extended to never-before-seen scen
s show its superior performance compared to previous works in task planning accuracy and efficiency.
ather than an optimal one, the method substantially reduces the time complexity of searching to a polynomial level. Moreover,
ficiency and correctness of detecting frontier points lower. Meanwhile, the allocation strategy of the greedy algorithm causes ea
al haptic feedback modality, the human operator gains full sensorimotor control over the robot, and the awareness about the ro
munication delays. We validated the proposed approach in two different in-contact tasks using data collected from 18 participan
using the pulse width modulation signal (PWM) along with the joint speed to estimate the interaction torque. The mechanical join
n intelligent mode-switching framework by jointly considering mode-switching and communication systems. User intention recog
commercial remote controller with a kinaesthetic rendering feature and a wearable cost-effective tactile rendering interface, ca
int cloud-based force rendering algorithm is tailored to use the OBM, thereby allowing the generating of force feedback for com
demonstrate that the proposed control framework offers improved visual feedback quality and drawing performance.
cified to lump iron removal task with two types (KF and VF). Both POsticks have rod-shaped end tip which is identical to actual
ly used to pick up an object, as well as adjusted for new environmental constraints, with our assistive robot EDAN.
olved a simple pick-and-place task, which was performed using two different robots equipped with two different control modalitie

g the robustness and reasonableness of the place task. Our proposed method is extensively evaluated in two simulation and on
average reduction of 4.67% of the area covered by the robot within the frame when compared to the case without the optimizati
omising new approach for intuitive robot control through smartwatches and smartphones from anywhere, at any time. The code
ficantly more quickly and reliably than those that do not.
predominance of the deep hierarchical model in the sense of prediction accuracy and early intention identification. We impleme
e validated through locomanipulation experiments and heavy box pushing tasks. Here we show two forms of DMM: grasping a
 er's desired motion. Moving forward in this work, we focus on synthesizing and completing the user-intended motion autonomou
max pooling for pre-trained point cloud Transformers. Besides standard metrics such as accuracy or mIoU, we evaluate the num
wild while avoiding introducing artificial artifacts that lead to trivial solutions. Meanwhile, our SAM consistency loss leverages the
e effects of transitions across two domains through a symmetrical optimization structure for learning these mapping functions. O
parameter and data efficiency of the proposed method. Meanwhile, we obtain new records on 4 public datasets and multiple 3D
to ease the domain shift accumulation, which consists of a Depth-Aware Teacher (DAT) and a Geometric-space Aligned Stude
additional data. Supplementary material is available at: https://gergely-soti.github.io/grasp
s are performed in coarse simulations to learn lower frequency bands with efficient convergence, while higher frequency levels
urbation and utilize Kalman filtering to reduce the discrepancy between source and target domain data. We refine the source do
ment and to control various personal mobility devices, including powered wheelchairs and standing moving devices to perform
gh the experiments, the DARFT-1 reduced GRF by an average of 9.8 % and 10.02 % in three trials when placing the obstacle o
n terms of adaptability, robustness and reparability. A demonstration that assisting the operator in holding wires and delivering
which depicts the stretch of the flexible sealing lip. For the 15 kg-weight climbing robot, the adsorption force can be kept over 10
ulator for inserting the plate into the groove by a conventional motion planner, RRT. Our proposed method accomplished the ins
a modular cooking cell composed of customized appliances, providing flexibility to easily alter the kitchen setup. Our goal was
p to a 180-degree field of view. Mono Wi-Fi Scene is ready for customization in many use cases. Because Wi-Fi can penetrate
ntal issues by guiding and supervising the robots, accessible anytime and anywhere, via computers or smartphones. This syne
human-machine communication can positively influence efficiency, productivity, and work experience across various industrial
uman motion analysis and interactions in mobile and fixed environments. However, recovering 3D human meshes from a mono
orks were trained on a dataset obtained from various users and were tested both individually and as a complete integrated syst

els that are learned and revised rapidly from limited examples to predict the behaviour of other agents. In addition, the architectu
that uses an external portable IMU for localization. The system is independent of any external setup or environments due to the

and pose. The AUV can track the target in 3D space based on the estimated states. This method has been validated through ta

hree activities: a Keep-with-the-Rhythm game, a Sing-Along, and a Dance-Along. In each activity, the robot guides users throug
rocedure. The ARTERY project focuses on the development of a radiation-free percutaneous procedure using autonomous rob
nd manipulation (MoveIt).
cted features enable robots to explore potential trajectories more effectively, significantly enhancing robot compliance. We achi
ach involves adjusting leg posture and utilizing foot mass to lean into a turn. By lifting one foot off the ground and maintaining a
market mechanisms, incentives, and governance frameworks for these autonomous economic interactions. Furthermore, we im
th an EMG-based artificial neural network (ANN) to continuously estimate desired hand movements. The shared controller feat
der physics that are managed by AGX Dynamics. The cobot cell includes a safety scanner, which detects motion on a specified
magnetic gripper at distant points. This 1.2 m limb greatly increases EEWOC's workspace and the size of obstacles it can trave
llustrates the proposed concept and vision in various scenarios within the aforementioned application fields. By doing so, we ai
ion at a depth of more than 3600m but also demonstrate its ability to lift heavy objects. This gripper has the potential to make th
ance and shape adaptability allow it to navigate spaces up to 70% of its maximum height, seamlessly switch between rolling an
and a mechanism that automatically locks the unrolled outer skin. The results of this research stimulated many researchers aro
rated 5g of buoyancy, with a total adjustment of up to 16g, equivalent to 11.4% of their own weight. We also demonstrated that
d compact and have better force interaction performance with wearer and environment. The shoulder and elbow flexion/extensio
changes muscle activation and perceived exertion of its user during collaborative elbow flexion. Participants were instructed to

T-shirt folding experiment.

hod which detects specular highlights and reconstructs the surface normals from it. It assumes that the light source is approxim
challenge: single-image reconstruction of robot poses under noisy observations. In this paper, we address the problem of tracki

opagation of tracks between frames. The main focuses of AgriSORT are efficiency, flexibility, minimal dependencies, and ease
osition and refine it in the factor graph, using chi-square test algorithm to identify UWB ranging outliers. Based on the above two
as removed from gyroscope measurements. We then leverage robust and accurate rotation estimates to enhance translation es
d up to a 31× improvement in distance field computation, and is available open-source.
d novel guidance, improves the state-of-the-art performance on the challenging object navigation task from 70% to 73%. Along
n a Clearpath Jackal robot, we show the global stability of the proposed system in both indoor and outdoor environments cover
our approach in simulation and real world, showing 5% and 16.67% improvement in terms of navigation success rate, respectiv
t the training set and boost the model performance. A robust point cloud registration uses the detected semantic keypoints and
dology, we incorporate a nonlinear MPC controller that accounts for the intricacies of the vehicle's model and states. The model
e error and misalignment on coarse maps, CMN extends the Bayesian filter and maintains a belief directly on coarse maps usin
tion, recognizing object pose using only a single blinking LED presents challenges. To address these challenges, we introduce
t encoder produces multi-scale features with a single shared block, which allows us to adjust the pyramid length at inference tim
t focus on relevant objects during training and understand the environment better. Our model outperforms previous methods in
tering, enabled by place recognition, further improves navigation performance by increasing the temporal consistency of subgo
al-language pre-training to align natural language description with visual perception. The integration of a continuous knowledge
te that an agent equipped with POLoNet significantly outperforms a range of baseline methods, including end-to-end RL techni
kinds of objects. Besides, our PixNav is a pure RGB-based policy that can reduce the cost of home- assistance robots. Experim
on approach. Comprehensive experiments show that GeoAdapt significantly boosts place recognition performance across mod
d-to-end based on the planning task objective. This optimization uses a differentiable formulation of a semantic costmap, which
need to be retrained if the domain or the OOI changes. We demonstrate this within simulation by deploying the same navigation
ht, we introduce Preference extrApolation for Terrain-awarE Robot Navigation (PATERN), a novel framework for extrapolating
res, facilitating the reconstruction of potential high-frequency and edge information. Furthermore, a prior-guided reconstruction
aph edges. Finally, each sub-graph represents an object instance. AnyOKP is evaluated on real object images collected with th
xiliary Ray method to tackle the issue of sparse viewpoints in autonomous driving scenarios, which leverages sequential frames
both supervised and self-supervised V2D. Extensive experimental results demonstrate that our method outperforms previous
Doduo illustrates superior performance on point-level correspondence estimation over existing self-supervised correspondence
framework, equipped with a BEiT-3 backbone, to address all visual perception tasks in the ARMBench challenge—a large-sc
es previous state-of-the-art approaches, and we conduct comprehensive ablation studies to validate the effectiveness of our pr
g the results from visual odometry. With this framework, robots and vehicles with only one well-calibrated camera can collect se
enhance the quality of generated event voxels significantly. We also propose a novel local dynamic-aware timestamp inference
es a sequence of images as input and performs joint object detection and pose estimation for all objects in one forward pass. It
ssive zero-shot generalization capabilities. Our method outperforms existing thermal-based depth estimation models, opening n
oiding misclassification of unknown classes as known ones. To support our assertions, we establish an open-set surgical activi
 to the real-world domain. This approach is achieved through a combination of 2D and 3D supervised losses during the synthet
oral response consistency for supervision, ensuring stable and efficient keypoint detection. Moreover, we use a spatio-temporal
content, and frame rate, a feature not provided by previous methods. Experiments demonstrate the ability of the proposed meth
sion images with salient objects using a specially designed CIoU loss. The two adversarial networks ensure that the fused imag
the training phase, our network can produce accurate image-level cuboid annotations with 3D properties from Lidar boxes. We
ular for depth prediction task.
merges as a promising avenue. In this paper, we introduce a practical framework for generating images from a BEV layout.Our
our private dataset, we demonstrate that our approach not only exceeds the current state-of-the-art in novel view synthesis qua
dataset for event-based 3D reconstruction, which contains 101 objects with various materials and geometries, along with the g
g these limitations, this study introduces a deep learning framework aimed at enhancing SLAM performance through the strateg
valuations on three photorealistic datasets and one real-world dataset. Extensive experiments on the benchmark datasets show
differential pose layer. Finally, we introduce a novel loss function and gradient descent method to better supervise the proposed
omotor policy learning via a state decoder and an auxiliary reconstruction objective. During training, the state decoder reconstr
-independent, and viewpoint-invariant, thus generalizing well to categorical objects in novel scenes. We evaluate our approach
perimentation, we have demonstrated that our method is capable of synthesizing realistic images of lunar environments.
dataset refinement and online next-best-view (NBV) planning applications targeting the task of 3D reconstruction. With this cont
ateral candidate regions using ellipses and nearby corners. The Reed-Solomon (RS) algorithm is utilized for both encoding and
d clock devices. The key module of our system,~emph{FreeAlign}, constructs a salient object graph for each agent based on its
that our method is able to generate clear 3D scenes and outperforms existing state-of-the-art (SOTA) image adhesive waterdro

nces, we propose a 3D consistent point mining strategy, which effectively discards unfaithful points reconstruted by NeRF. More
arts. Each group comprises five sets of joint nodes, focusing on the body parts, excluding the occluded ones. Finally, to enhanc
e enhance the original distillation scheme in multi-frame methods to better exploit the knowledge from a teacher network. MAL
cted, and then a pre-trained vector quantization GAN is employed to create a discrete codebook of well-detailed data patches. A
ure. Furthermore, we propose a novel and effective data augmentation method named conjoint rotation which reasonably augm
the best of our knowledge, OpenAnnotate3D is one of the pioneering works for open-vocabulary multi-modal 3D auto-labeling.
ions, a natural recourse is to adapt SAM to additional visual modalities without compromising its robust generalizability. To achi
such as disaster recovery and drone delivery at night.
alize unsupervised spike depth estimation by introducing simulated mediate source spike data. Specifically, we design a Coars
he information provided by OPP. We evaluate the performance of PillarGen using our proprietary radar dataset, focusing on en
guration each time. In this paper, we propose a novel deep optimisation framework for the photogrammetry camera position for
multiple morphological information (spine region, centerline, and boundary) and then concatenate the original X-ray image with
g large real-world dataset. Overall, our synthetic dataset contains 32,400 independent images (each with different imaging simu
prediction and pose estimation stages, our method is named ContourPose. First, an additional contour decoder is adopted to i
nts showed that the soft acoustic end-effector can change the traveling direction of microparticles and modulate the location of
to evaluate the efficiency of this method. It is simple, fast, and efficient to deal with many kinds of object which are random in s
dataset using a physics simulator, which incorporates 3D CAD models of objects and automatically generates annotations for b

ce implicitly within its architecture. Trained in a self-supervised manner on a large curation of endoscopic images, RIDE require
re data imbalance arising from the uneven presence of surgical instruments or activities during surgical procedures. This paper
control of robots. Moreover, a comprehensive dataset focused on ILM peeling during macular hole surgeries was established. E
ng the approach angle using 2D image features, demonstrating that this algorithm can be used for grasping practical objects in
d and the object is detected by using the contours of both, obtained from a camera image. Then, to avoid damaging the object,
mately, we show that recombining demonstrations leads to higher task-respective success. We present extensive simulation and
ork, which models the error in Cartesian space. Our approach involves extracting an acceleration control signal from the traditio

approach are the point cloud of the object’s surface in its rest shape and the point cloud captured by a 3D camera in each
me tracker with fixed parameters across transformer layers. Moreover, a gated feature aggregation mechanism enables adaptiv
mera and an inertial measurement unit. Through extensive validation in diverse challenging indoor and outdoor environments,
fast movements of drones to enhance the drone-based object detection and feature embedding for small-sized, blurred, and o
proach provides robust tracking in human-robot interactive tasks agnostic to measurement type, even in the presence of prolong
ated language descriptions, rebuilding the pipeline of Tractor with modern large pre-trained models for static image detection a
his strategy allows the feature learning of the search area to be aware of the template while keeping the individual characteristic
tead of better, which suggests that the benefit of seeing the new data is overshadowed by the noise in the training signal. Focu
uced to account for missed detections and graph sparsity enforced to retain real-time efficiency. We evaluate our SWTrack imp
estimates are meaningful enough to be leveraged effectively in object tracking. We introduce UncertaintyTrack, a collection of
al use-case is found in some large-scale manufacturing where bobbins are enrolled with various materials (cables, papers, rubb
and 2) pruning encoder layers from a heavy tracker to refine the balnace between performance and speed. As an example, our
e correction for keyframes within abnormal quantitative windows. Our system demonstrated excellent performance over long di
riables. We validate our method with real-world data in both indoor and outdoor environments with different terrain types and w
nse motion, the inverse distance of the new candidate point often fails to converge, leading to either scale drift or tracking failure
tart-of-the-art methods. Additionally, we propose to use a filter to sequentially initialize the reference frame transformation till co
the multi-basis depth representation, we can continuously optimize the 360 degree depth with the traditional optimizer to achie
th trajectories.
ph optimization framework. Experiments in the public benchmark datasets show that our method can achieve superior performa
ll optimization process.
system in challenging scenarios. The practical validity of our approach is supported by our experiments on public datasets. Ou
e integrated the methods of sampling and spline curves, addressing the planning challenges of curves. Meanwhile, we utilized
y without multiple actuators and complex controllers. Furthermore, real-world experiments confirmed that the grounding point of

cle (PAM), which stores the compressed air, that can then be released into a pneumatic cylinder to generate torque. All electron

e motion trajectory via probabilistic movement primitives (ProMP). It then learns from the patient during the training process to fu
s tested through lifting experiments, demonstrating that it significantly corrects lifting posture, reduces stress on the lumbar spin
grated with machine learning, demonstrating its accuracy and usability in various motion scenarios. The results confirm the effe
exploration of human-preferred parameters and acquire more satisfying results for naive wearers via the HELIE acquired from
e, neglecting alignment in the other dimension due to inherent design constraints. For the first time, this study introduces a knee

gh experiments involving five healthy subjects.

ng-to-leaping maneuver endows the robot with an alternative and time-efficient approach to negotiate the step obstacles.
ons and validates the method with physical experiments, showcasing the feasibility of concurrent ground contact state and veloc
 gram in < 3 ms which shows that our formulation is suitable for real-time control algorithms. We also applied this framework to
thod to the specific case of the wheeled bipedal robot Ascento, for which we derive closed-form expressions of the dynamics eq
e the developed robot prototype on level ground and over obstacles. Additional simulated tests are performed for uneven terrain
prediction, yet does so for the slip prediction. For this, we propose novel optimistic and pessimistic methods, together upper an
d to bridge the transformation of the bicycle and self-balancing vehicle modes. A general kinematic model is developed for plana
egion. The design of the controller is mainly divided into two steps. Firstly, the linear velocity and steering angle of the robot are
ain a successful landing in all directions, even in the presence of significant horizontal velocities. The method is demonstrated t
ajectory is executed via a linear MPC controller incorporating the desired manipulation forces into its prediction model. Our app
ep reinforcement learning (RL) and the Student-Teacher approach. Finally, we evaluate the performance of these policies on a

upper body joints online. The effectiveness of the proposed method is validated in simulations and experiments on the humano
can be seamlessly integrated into whole-body manipulation tasks. We conduct various experiments in both simulations and wit

real world, indicating their performance potential for force-critical tasks.

oid function (TDSF) is implemented to generate more intuitive haptic feedback from obstacles. Comprehensive human experim
hich achieves comparable or superior performance over existing text spotting architectures for both regular and arbitrary-shape
ss-wise prototypes in the shared feature space, the soft pseudo labels can be adaptively denoised, which benefits the decision
lenging cases of occlusion and outfit variations. We compared our framework with a state-of-the-art Multi-Object Tracking (MO
oint clouds and therefore can be used to identify the correct candidate among potential matches retrieved by global similarity se
ue to their natural pairings with images. The open-vocabulary zero-shot performance improves with the strength of the pre-train
lars, which further reduces the model size, and theoretically prove their equivalence. Experimental results demonstrate that LP
d two tracker versions, namely RBLT and DeepRBLT, based on the depth of the features. Comprehensive experiments on four
s of objects and for establishing local correspondences. These local correspondences enable deriving geometric corresponden
oposed duplex attention mechanism leverages optical flow and RGB to address the lack of temporal information. The method e
ation theory, and its stability is proved. The feasibility of the proposed VTSEA prototype and the precise output torque performa

stiffness motor. SVSA-III has the advantages of compact structure, wide-range stiffness regulation, reduced control difficulty, an
he feasibility of the concept by conducting a series of hardware experiments using an open-loop control signal obtained by traje

cs of power density, mechanical efficiency and frequency response, have been conducted. In summary, the proposed actuator

imilar to motors. In this paper, we introduce a new type of variable stiffness spring design which is energetically conservative de
e optimizer takes into account aspects not considered in the path planning, like temporal constraints of the motion imposed by li
elay LOS links between communication nodes for complex mountain environments is presented. This study considers different
nd localization errors. Simulation experiments in a photorealistic environment confirm that the proposed planner can address th
es for replicating close-proximity operations between chaser and target satellites employing fully-actuated multi-rotors that allow
); textit{ii}) methodologies (eg fine-tuning all model parameters vs. only a subset); and textit{iii}) self-supervision strategy. Our a
ts encompassing six aerobatic maneuvers, a time-optimal drone racing trajectory, and an airshowlike aerobatic sequence for th
ework is thoroughly evaluated by 1) trajectory prediction of an irregular free-flying object subject to aerodynamic drag, 2) agile fl

ToF deck, featuring a total flight mass of 35 g. The algorithm only uses 0.3% of the on-board processing power (210 μs exec
hlight the effectiveness of our agent in terms of acquiring policies for high-dimensional continuous action spaces, enabling the s

n a Chinese Guzheng.
bots remains challenging due to the cost of hardware and the development effort involved in adapting existing hardware for a ne

ng-based Adaptive Controller (HMAC) to learn and adapt to such multi-source disturbances. Within HMAC, we develop two tec
or estimating high-probability regions of target object presence. Then, we propose a probability searching module that uses the
kway for robot navigation without the label. In the meantime, we design the trimming strategy that works collaboratively with dis
aIn, we introduce a novel dataset called the problem description generation (ProDG) dataset. The framework is evaluated with
euroMHE on real quadrotor flight data for disturbance estimation. Our approach demonstrates highly efficient training in under
ents on texture surface samples from the MVTAD dataset confirm the efficacy of MCDEN with a pixel-level AUC score of 96.6%
rajectory waypoints, supporting various object manipulation. To effectively prompt LLMs with the kinematic structure of different
ing Domain Definition Language (PDDL) formulation. In the planning phase, an LLM planner formulates an initial plan, which is
are the energy consumption of our model with a non-spiking machine learning algorithm under identical dataset and operating c
atures, thereby strengthening model versatility and robustness. We validate our approach through extensive experiments on bo
hat our network has better performance and efficiency compared to others. The code is available at https://github.com/KTH-RP

nally, we demonstrate that our approach is effective in a simulated campus patrol robot domain.
based vector quantization, allowing the agent to generate suitable actions in response to unseen textual commands from LLMs
at training a bipedal locomotion policy with the inferred reward functions enhances its walking performance on unseen terrains,
experiments, our method indeed emits an alert). We demonstrate our approach in both simulation and hardware for a visual se
nd a maximal feasible return (e.g., Information Gain) in terms of Value at Risk and a corresponding action sequence, given a se
orms five baselines. It is also the simplest of the baseline strategies besides one ablation. Being easy to use, flexible, and with f
n task and a 6D pose estimation derived from videos. We posit that as robotic systems evolve with greater degrees of freedom,
we establish a comprehensive evaluation metric for test suite quality, considering factors such as diversity and naturalness. This
in the future. Such outcomes derive both from a proper coordination among the heterogeneous AI sub-modules in the system,
iently learning robot food-slicing tasks in simulation. Following a real2sim2real approach, our framework consists of collecting a
scene graph subsequently informs a generative model, which forms a fine-grained goal scene considering the shape informatio
s (centralized, decentralized, and two hybrid) as applied to four coordination-dependent multi-agent 2D task scenarios for increa
elf pre-trained MLLM into policy networks. Through a series of simulated and real-world robotic tasks, we demonstrate that rob
Our experiments reveal that a realistic simulation is essential for mastering complex robotic tasks, yet translating these accomp
r perform reasoning tasks. To assess our methodology, we built a dataset featuring real-world trajectories, capturing actions ran
ull robot APIs presents a challenge to context-length limits in LLMs. In this work, we discuss how to combine different LLM prom
ework to train and deploy a compositional RL system that successfully pilots a Warthog unmanned ground robot.
rs. Based on this insight, we then derive a new adaptive replanning strategy based on deep reinforcement learning, which can l
o improve the robustness of loop closure detection. We validate the effectiveness of our method through experiments conducte
to a wide variety of application scenarios, and illustrates a broader paradigm in affordance-based reasoning.
enerated and human-developed assets, we chain-of-thought prompt LLMs to map these to relevant task descriptions, tempora
ousehold tasks. Results show that our model significantly improves faithfulness and success rates for complex long-horizon tas
multimodal driving model, demonstrating unparalleled results in diverse tests. Our solution combines language with visual perce
e task descriptions to an intermediate task representation that can then be consumed by a TAMP algorithm to jointly solve the t
o lounges, one office space, one outdoor courtyard). In rigorous quantitative comparisons in two environments, ASC achieves
are different mechanisms, utilizing metrics such as data similarity, data age, and consolidation frequency. The efficacy of forgett
ttance control. This enables the user to generate commands based on virtual forces while ensuring compatibility with the enviro
n of UAV operators.
be solved as two small quadratic programmings (QPs) and deployed in real-time. The computational efficiency and dynamic c

h only needs to recover the cable at a constant speed until the tethered lifting-wing multicopter returns to its base. The feasibilit
ded. Furthermore, we propose a novel occlusion correction algorithm that removes incorrectly projected LiDAR point clouds into
nd that it is suitable to be used as the main source of relative localization for swarm stabilization.
The heuristic frontier regions are detected by applying a lightweight vision-based method to the different levels of the occupanc

ing attractors and their RoAs for data-driven controllers operating over high-dimensional systems, including a 67-dim humanoid
cro-nano-manipulation.
fast and accurate parameter estimation. Global exponential stability with parameter convergence of the closed-loop system is

he proposed method generates safety-critical scenarios of higher quality compared with previous approaches.
d estimates the target vehicle's pose with respect to the ego-vehicle. This information is then fed to a Multi-Layer Perceptron, wh

arch. Our framework, named TrafficRLHF, demonstrates its proficiency in generating realistic traffic scenarios that are well-align
nt, we propose a queryable safety constraint module based on linear temporal logic (LTL) that simultaneously enables natural l
driving scenario, offering valuable guidance for selective fusion of sensor data, thereby enhancing driving decision-making. The
ate the superiority of POP in providing reasonable and safe trajectory predictions.
w that our future interaction modeling improves the performance remarkably, leading to superior performance on the Argoverse
to produce certain situations in simulation even when the Ego Vehicle tries to evade them.
des knowledge, enabling behavior sharing. We theoretically investigate the properties of the KT-BT framework in achieving hom
roduce conflict-free assignments among agents in finite time. We compare the performance of DMCHBA with the consensus-b
ame prediction length.
Finally, we showcase the benefits of our approach via real-world experiments that indicate the superior performance of LGX in
other. The latter two variants yield a high information gain when evaluated on satellite data from the Netherlands. Additionally, th
ding to environmental changes, which is robust to external disturbances. Our method is valid with simulation in different practica
ction datasets: the ETH/UCY datasets and the Stanford Drone Dataset (SDD). The experimental results show that with DyHGD
LLM-based driving action generation in comparison to traditional behavioral cloning. We make our benchmark, datasets, and m
(TTHMS) with learnable hyperedge-type embeddings to model the intra-group and inter-group level interactions among heterog
urve fitting methods. To fuse the radar odometry estimation with the acceleration and yaw rate measurements handling non-line
red trajectories and employs a model predictive control strategy to determine the optimal steering angles needed to actualize th
ive performance compared to state-of-the-art methods involving hand-crafted rules and exhibit superior generalization capabilit
crease the cost. The proposed system is validated by real-world crowdsourcing data and the result is comparable to the traditio
lities and likely trajectories for occlusions, as well as forecast motion for observed agents. We explore common observability as
mantic features and map objects. We evaluated our proposed localization framework in the publicly available KAIST Urban data
odel to encode the full distribution, and optimising predictions using both probabilistic and closest-mode measures. These obje
level optimization algorithm that simultaneously considers multiple future scenarios and user-defined risk tolerance levels. Owin
changing trajectory planning method scenario-adaptive. Our proposed method is evaluated through simulations on the well-kno
ally challenging terrain. We make our platforms, datasets, and algorithms publicly available to facilitate future research on whee
that geometric systems can produce trajectory plans that align with the human demonstrations in a large number of social situa
thod better than the state-of-the-art methods.
enable further research in LVLMs for autonomous driving scenarios, we develop and release Talk2BEV-Bench, a benchmark en
multitask learning. Furthermore, the DualAT introduces a novel waypoint prediction architecture that combines the temporal re
avigation affordance, and estimates the future novelty via random network distillation (RND). A tailored cost function searches

embedding and local features, preserving the 3D geometric information of the given point cloud. Then, we aggregate them into
GV and show that it performs well in real-world scenarios.
fy new agents with limited historical data as a major contributor to error and demonstrate the complimentary nature of 3D skele
 quired. Additionally, to better utilize the provided guidance, we augment the RL regarding the fundamental architecture and opt
o the future. Extensive results on CARLA prove that our model consistently outperforms strong baselines at both training and n
evaluated both networks on our traversability dataset, achieving an 89% accuracy in terrain traversability classification with the
ough image-point-cloud registration. Accurate correspondence between the model and observation can be estimated in real-tim
egates the local observations (raw 3D-LiDAR data) and the expert guidance information (gas distribution), and directly generate
ust navigation performance in a variety of unknown environments and even 50% higher success rate over existing methods in c
found at https://www.openbot.org/
augmented dataset WOMD-LiDAR consists of over 100,000 scenes that each spans 20 seconds, consisting of well-synchroniz
n be improved by 96 % with the proposed method.
generalization compared to prior approaches in imitation learning settings. Open-sourced code and videos can be found on our
ands. Using merely 7500 demonstrations, we are able to train a single policy, RoboAgent capable of 12 unique skills, and demo

ne fine-tuning of a pre-trained policy in the presence of distribution shifts and leverage pre-trained vision language models (VLM
ble models that are very competitive on the Waymo Open Motion Dataset (WOMD) and Argoverse leaderboards. From these e
matches and present a taxonomy of eight high-level categories of how these errors manifest. We find that developers often mak
o shows a rapid climb immediately after takeoff) even when the time-averaged thrust-to-weight ratio is smaller than 1. This is si
the external environment and overcomes the inability of open-close flippers to cross a 90° corner, which shows the rationality
acking inside an indoor flight zone. The objective is to track circular trajectories of a radius 2 times greater than the wingspan at
maneuvers under the guidance of the general kinematic model. Through this systematic methodology, one can easily study th
endage entrapment, resulting in a large negative displacement in the robot body during the swing phase. A simple model succe
ements and body posture kinematics of real fish in fish-like robots, allowing us to measure the power cost, thrust, and detailed
evaluation and comparison. The results show that our ASE method performs better than other existing methods in straight-line
gh current and increase the initial stiffness by up to 57.4%. More importantly, we incorporate the cost of tunable stiffness into th
periments show that the swarm robots can collectively leave away from the threat and respond to sudden environmental chang
ks? 3) How do animals manage to tolerate high sensorimotor delays, yet still produce smooth and robust gaits? To answer thes

he robot off a ledge at various approach angles. As intended by our design, the robot recovered the upright orientation from all
ng motions. After optimization, which takes only a handful of jumps, we show that this control architecture is capable of diverse
ced unwanted roll, confirming that this appendage increases maneuvrability by increasing stability, rather than by directly contr
ate how the spine should be used in dynamic quadrupedal locomotion and achieve more possibilities in the future.
lel underactuated MPC controllers for each leg at the control level, contributing to heightened motion precision and stability. Ex
d operation is also presented, showcasing the potential for multiple applications in search and rescue, exploration, and monitorin
d to unlock the secrets of navigating in the subterranean world more efficiently.
lapping-wing MAV currently available, making this is the first report of a battery-powered tailless flapping-wing MAV with insect
body length of our jumper exceeds that of other robots with stable mechanisms by 30% and 33%, respectively. Meanwhile, our
onstrating autonomous multi-modal locomotion. This work opens up new pathways toward autonomy in multi-modal microrobots
can initiate both left and right turns, demonstrating high maneuverability along the air-water interface. In addition, the robot can
Bot+ can swim at speeds as high as 16.1 mm · s−1 (0.35 Bl · s−1), when driven with heuristically-optimized propulsors, an
me vectors in general stay well below the 90° required for successful homing, and below 24° if all images contain sufficient te
based on NSGA-II method. The ”force vector” produced by the proposed platform has been illustrated through both the
ess, and cost-effectiveness. Various experiments, including voltage-current measurements, ionic wind speed, thrust quantificati
 nals of the motor area of the BNN are recorded by calcium imaging, analyzed, and decoded to control the motion state of the m
nabled us to determine the magnitude and direction of the force with an error rate of 4.8 percent and 11.1 percent, even when t
ormance of the foldable arm, the variable-diameter wheel, as well as the traversal performance of the robot. Experimental resu
ntroller that incorporates programmable body compliance into the sidewinding gait template. Our experimental results highlight
s built, i.e., the tail that creates roll motion is attached to a wheel-equipped robot whose wheels act like a reaction wheel and ge
method of automatically selecting and supplying modules using environmental cues. More precisely, we developed a mechanis
we validate the stiffness and load capacity using a developed prototype.
ntrol is implemented on the robotic eye prototype to examine its performance on three classical types of eye movements: visua
perior performance in synthetic and real-world datasets, enhancing downstream manipulation tasks by providing precise camer
. The code is available at https://github.com/OpenCalib/CalibAnything.
four new evaluation metrics to quantify the robustness and accuracy of extrinsic parameter estimation, applicable to both single
ring methods is over 98%, whereas OCAMO monitoring can detect small decalibrations better, and LTO monitoring reacts faste
ompared to existing methods. We open source our implementation on https://github.com/joomeok/PeLiCal.git.
Our method's distinct advantage lies in its ability to perform accurate calibrations at close distances, owing to the efficient featu
le to quickly carry out calibration. Experimental results on real-world datasets demonstrate the reliability and accuracy of our m
ation with semantic information and geometric features to generate accurate extrinsic parameters with the initial value from the
seline. The 2D pattern can only be detected from one direction and needs to be of significant size to compensate for its distance
Through weight sharing between update iterations and by enabling the update operator to learn from data with varying degrees
nd identify collaborative robot joint dynamics, and the results are verified and validated through numerical simulations and expe
herein the selected parameters are only evaluated on a subset of real world evaluations. The approach was successfully evalua
nsic parameters, while the rigidity of LiDAR points is considered in each iteration to enhance regression robustness. Extensive
nsor mounting methods. In the experiments, we perform GBEC between the robot end-effector and an optical tracker's rigid bod
der to improve the accuracy of inferred inertial properties, an attention model is designed to generate the weights used in least
ers of camera and LiDAR image features to achieve high-resolution representations. A multi-head correlation module is utilized
rigger numerous events, which can also be captured by LiDAR. Our proposed method leverages the edges extracted from even
novelties are: (1) A new target is introduced: an eight-quadrant corner reflector enclosed by a foam sphere. The benefit is the ta
dge, this is the first example of sim-to-real learning for vision-based bipedal locomotion over challenging terrains.
s and pits, prompting corresponding reactions. We validate our approach in simulation and on the real quadruped robot ANYma
ator. Remarkably, VLFM achieves state-of-the-art results on all three datasets as measured by success weighted by path lengt
or prior methods based on latent variable models. We instantiate our method by using a large-scale Transformer- based policy
onducted on various challenging continuous control tasks on simulators and real robots, we highlight the significant potential of
tate-of-the-art learning-based algorithms using less support points, thus making our proposed method more sparse.
this guarantee to remain safe, PSR can incrementally improve its prediction and planning performance online. We demonstrat
obile robot and stochastic dynamic obstacles.

al LiDAR observations. Additionally, we customize a LiDAR-integrated simulator to speed up sampling and normalize the geome
d confidence of the trajectory rather than direct control inputs as the output of the neural network, it generates a more realistic a

eloped method and its key modules. The experimental results demonstrate the effectiveness and efficiency of our method. The

in a unified manner.
 the difficulties of training; (2) It takes multiple time steps to track a path instead of a single time step, which requires the path to
Planning) to resolve the representation mismatch. ZAPP unites a prediction-friendly coarse time discretization and a planning-f

ed that the proposed approach is comparable with existing local planners: it provides trajectories with outperforming smoothnes
pproach results in confident and safe robot navigation in tight spaces by smoothly slowing down the robot in low survivability en
the robot's preferred velocity direction. We demonstrate these improved navigation capabilities in real-world narrow, dense env
navigate environments with varying elevations and steep obstacles. We evaluate the performance of our proposed approach t

hrough 41,412 random benchmark tasks with 150 pairs of 3D objects, we show GDN-R outperforms state-of-the-art baseline m
to traditional LOS checks. We refer to our approach as Jump Over Block (JOB). To demonstrate the advantages of JOB, we co
rforms the existing controllers in simulation. We demonstrate our approach with multiple experiments in a collision avoidance g
time, our model utilizes multi-sensor inputs to predict the success probabilities of the trajectories generated by a local planner t
y in narrow environments, appropriate informed guiding speeds enable E-RRT* to outperform other methods.
own environments was validated by conducting experiments on unicycle-type robots.
by capturing human priors of these concepts from visual appearance. We incorporate this physically grounded VLM in an intera
RVNet, we generate supervision labels using the ShapeNet dataset. Simulated experiments show that our PRV-based view pla
observation. To train our OSVP network, we generate supervision labels using dense point clouds refined by implicit represent
e of the art in transformer-based LiDAR object detection on the competitive nuScenes benchmark and showcase the benefits o
Furthermore, we introduce the MMArt dataset, an extension of PartNet-Mobility, encompassing multi-view and multi-modal data
residual maps, thus greatly improving the performance of the Lidar-based MOS task. Remarkably, our MF-MOS achieved a lea
mance in a self-supervised paradigm. Experiments on synthetic and real-world datasets show the capability of reducing the inter
tion, motivated by acceleration techniques of neural radiance fields. However, we directly store the normal and density values i
nloaded from our project website: https://wenbwa.github.io/TRTM/.
ransformer decoder which facilitates recurrent hierarchical feature propagation and amodal semantic grounding. We demonstra

methods, especially in scenarios involving challenging conditions, the improvements of our method achieving at most 0.1094m
cts from different perspective directions, and multiple 3D constraints based on object invariance and semantic consistency are a
batch of events, and "TID" which iterates over time with streaming events in an online fashion. Our top-performing model (ID) se
queries to interact with multi-robot multi-camera inputs. Each BEV query can actively select relevant cameras for information ag
nctions, and spatial inductive biases. Leveraging our R2DM model, we also introduce a flexible LiDAR completion pipeline base
aring approach. To the best of our knowledge, the proposed model is pioneering in its capability to concurrently handle multiple
ne the segmentation and detection features in a shared transformer decoder with cross-task attention layers to enhance and in
e boundary geometry and structural compliance of unstructured objects. Moreover, sensor-based feedback control can be acco
achieve an average static measurement error of 0.17°±0.15° and an average dynamic measurement error of 1.34°±0.4
urfaces without an increase in complexity while reducing the time-consuming step of probing the surface. Experimental validatio
adient-based parameter fitting, online estimation via an extended Kalman filter, and online gradient-based MPC. The proposed
ng effect, we improve the Shape from Shading (SFS) algorithm to process tactile images, enabling a coarse but fast reconstruct
n and grasping.
s travelled distance by 23.8% and time by 77.4%. Furthermore, our method requires four fewer hyperparameters and decision v
o the tangential-to-normal force ratio, instead of a step function according to Coulomb's law. By modeling friction as a function o
EAGUE learns manipulation skills in-situ of the task planning system, continuously growing its capability and the set of tasks th
rfitting of the TTA model, we devise novel regularization which modulates the adaptation rates using domain-similarity between
her enhance performance. Experiments on diverse datasets with temporally correlated observations demonstrate our method's
utions before and after distillation, we further utilize a Fréchet Inception Distance (FID) loss as a regularization constraint. In o

abling human users to bootstrap new skills with minimal online interaction. LRLL continuously transfers knowledge from the me

CL methods, PSNN-VCL not only achieves superior performance in terms of widely used CL metrics but also achieves at least a
nes by over 11% in average success rates, showing its superior knowledge transfer ability compared to prior methods. More ex
d simulated, and flying poking experiments were also conducted on them respectively. On the other hand, the performance of s
e-tuning. Leveraging the deployed policy, we derive more real-world manipulation skills on ArrayBot to further illustrate the distin
er of possible compositions. We demonstrate that our approach outperforms a state-of-the-art baseline and is able to synthesize
amera, and design optimizations were performed to improve the robot's capabilities. The prototype WiBot 1.0 was developed, a
deformable object simulation, minimizing path tracking error while maintaining a predefined minimum distance from obstacles a
aths between locations of interest, while meeting a user-specified communication budget constraint. The effectiveness of the p
y areas. The UAV dynamically schedules waypoints to prevent UGVs from collisions, formulated as a Vehicle Routing Problem
vironment with obstacles. Both numerical simulations and real-world experiments are conducted to demonstrate the efficiency o
oduce a unique dataset, integrating force sensing data synchronized with OCT B-scan images, derived from a sophisticated se
l-world dataset. On these data, we compared manual labels, automatic labels and the predictions of a baseline deep neural ne
f the action. The results show that video is an important modality, but using force-torque data and gripper position help improve
e baseline results and the dataset is made publicly available to support the community at https://cea-list.github.io/CEA-IMSOLD
on and segmentation. PointSSC provides a challenging testbed to drive advances in semantic point cloud completion for real-w
https://baikaixin-github.io/structured light 3D synthesizer/.
s in video clips of dynamic scenes. Then, we introduce spatio- temporal OOI classification task which aims at identifying interac
NSS)-based setup. The dataset comprises around sixty experiments conducted in various conditions over a period of 17 month

on, semantic segmentation, and robot relocalization, is evaluated on our THUD dataset. These experiments reveal serious cha
bot dataset. We evaluate on a set of real-world collaborative human-robot manipulation tasks and show that our conditional mo
ataset, we provide a simulation and evaluation framework that enables a planner’s actions to be simulated in closed-loop to
he metric space. To the best of our knowledge, no prior data collection system has a combination of all three components. We f
or robotics-focused visual question answering. We also demonstrate how evaluating real robot experiments with an intervention
to generate corner case scenarios. Our model successfully learned to produce corner cases from input scene graphs, achievin
-rich robot manipulation sequences across diverse skills, contexts, robots, and camera viewpoints, all collected in the real world
antiating this principle requires training policies to minimize their effect on human behavior, and this in turn requires data that al
uth risks under different scenarios. We assess the ability of ten algorithms to (1) detect and locate risks, (2) anticipate risks, and

of the driving environment. We also introduce custom-designed deep networks aimed at multiple important object localization a
-scale annotated LiDAR point cloud under six groups of different sensors but with the same corresponding scenarios, captured
subset, focusing on challenging cases that involve partial occlusion and low visibility. Diverse recent methods are evaluated on
e seamless inclusion of existing datasets and a fast way to enhance the perception capabilities of various robots operating in un
free annotation enables the network to understand the context of a bin-picking task, where collisions between the gripper and th
ors, who utilize our annotation software to correlate the selected keyframes. Finally, we present a benchmark evaluation of seve
tasks that researchers may use to evaluate the accuracy of their algorithms. Finally, we run our open-source diameter estimati
ataset of LS3DPC with synthetically generated reflection artifacts to train the proposed network. Experimental results show tha
more, to filter out high-confident noises and obtain more reliable pseudo-labels, we investigate the geometric curvature propert
urrent state-of-the-art methods on the 3DCSR dataset, improving from 40.59% to 75.74%. Our code will be available at https://
rnatives for articulation estimation. Our research provides a promising solution for applying Transformer-based architectures in
e used to perform simultaneous outlier detection and semantic segmentation. Finally, to help expand the research field of LiDAR
educed by up to 97% without performance decreasing.
top S key points with the highest attention scores are selected, such that in the next layer, attention is evaluated only within th
minating the previous need for 3D point clouds and IMU input. Furthermore, we deploy our enhanced baseline algorithm on a re
ficient and practical in real-world applications. DAT adaptively selects and updates two small groups of trainable parameters ba
and slipping. Self-attention architecture with and without spatio-temporal sequences fusion mechanism (denoted as STNet and
eption errors. Our model applies a cross-attention mechanism to fuse the environmental point cloud and past human joint poses
f freedom (DOFs). In a 7-DOF robot test setup, the learned self-model achieves a Chamfer-L2 distance of 2% of the robot's wo
for the task of language-driven affordance-pose joint learning. Intensive experimental results demonstrate that our proposed m
ping, comparing our method with vision-based grasping network and RANSAC applications. The results from this real-world sce
ed from the training images, and labeled data is obtained to train markerless keypoint detection algorithms without the need for
data separately.
ts demonstrate that our approach successfully generalizes beyond the training domain and outperforms all baselines.
of optical flows in an indoor setup using standard BOS techniques. We then train a convolutional neural network (CNN) by apply
ich runs at 39Hz within 101mW onboard a Crazyflie nano-drone extended with the GWT GAP8 SoC. We compare our FCNN a
trategy to generate more potential proposals and utilizes a point-cluster agreement-based thresholding (PCAT) mechanism to f
tic quality improves from 56.65% to 70.08%. Furthermore, our approach is faster and can be tuned to improve inference time b
on three indoor datasets, including NYUv2, ScanNet, and InteriorNet, show that GAM-Depth outperforms existing methods an
mon deterministic regression methods may falter when faced with uncertainties from 2D detection errors or self-occlusions. Lev
yet effective automatic annotation pipeline for generating finer lane labels. To propel future research, we propose a novel LiDA
different environments. The experimental results show that our method outperforms the state-of-the-art PCP methods on the pr
function and, thereby, enables robots to tackle a much broader range of tasks in real-world settings. The resulting approach ca
en transformed into Spatiotemporal Risk Maps (SRMs). We verify our navigation system's abilities in simulation, validate it on a
observations and demonstrate superior utility for policy learning compared to other representation learning techniques. Furtherm
path-relative Frenet frame. We applied the PBP trajectory decoder on top of the HiVT scene encoder and report results on the A
cy while reducing inference time. The robustness of the proposed method is illustrated through a joint compliance and task com
ch in this direction could help democratize and distribute the process that forms multi-task robotics policies. Our implementatio
tory. The results show that our framework outperforms prominent models using five benchmark human motion datasets.
xisting reinforcement learning methods in terms of efficiency and performance.
g long-horizon planning tasks.

tion accuracy, whereas data augmentation and dropout rate play a significantly larger role. In a comparison study, while human
utonomous driving simulator to evaluate their performance under diverse conditions and in the presence of noise. The experime
 matically prune the lightweight model to obtain the final model. Through extensive experiments conducted on two widely-used
LP. The expansion of the voxel grid and the distillation of the shallow MLP are efficient in this case. Experimental results demo
simulated images. This demonstrates the effectiveness of our approach. We further validate our approach using the classifier
n forecasting tasks and obtain state-of-the-art performance by a large margin while being efficient and performing in real-time. A
hine vision tasks. Our solution consists of a base layer bitstream and an enhancement layer bitstream. The base layer bitstream
as new ones introduced in this work and find that our method significantly outperforms others in terms of the time to find a vali
oretical safety guarantees of SafeDPA and show the robustness of SafeDPA against learning errors and extra perturbations. C
m (0.5% of the simulated robot length) for each of the robot’s 20 disks.
amely the inference time of a neural network. This allows the robot to plan and replan reactively, making our approach suitable
ed framework. Furthermore, results indicate improved performance in terms of traversal time and success rate compared to bas

ying difficulty. The objects in the bin have various challenging physical and perceptual characteristics that can be characterized
s using a physical 7-DoF Franka Emika robot arm with a suction gripper.
oints between the robot’s coordinate system and the camera’s coordinate system, facilitating accurate pose estimation
its efficacy in an autonomous inspection, maintenance, and repair task. GANDER produces favorable results compared to bas
OS-TOG to generalize to novel objects and tasks more efficiently than existing alternatives. Additionally, the paper presents sui
oach generalizes better than prior methods while needing significantly less computational time for inference.

ined models. Moreover, we discuss rotation representations and associated loss functions that substantially improve policy lear
ng rigid objects to a target position with ropes connected to robotic arms. Furthermore, DeRi-Bot is a distributive method that ca
, regardless of the camera viewpoint. The efficacy and efficiency of the proposed method were evaluated both in simulation an

angled objects in the buffer bin planned by PullNet - a network that predicts position and direction for pulling from visual input. T
ng task, showing that our policy can reliably reach folded states from crumpled initial configurations. Videos, code, dataset and
ned with unified latent action space and further improve generalization, achieving real-time autonomous grasp control. Real rob
ty. Simulations show that the proposed framework outperforms contemporary counterparts, notably in grasping success rate an
hod achieves a median success percentage of nearly 90%. These results mark a substantial advancement in enabling shape s
n ordering of factors based on generalization difficulty, that is consistent across simulation and our real robot setup. Videos and
ethod significantly surpasses the baselines and has better generalizability, which caters to practical industrial needs.
be collected in minutes.
dynamics. In contrast, we learn a hierarchical policy model that operates directly on depth perception data, without the need for
% of the trials achieved a greater than 90% overlap with the 200 target shapes. By relying on segmentation images as a visual o
es. In addition, GAMMA adopts adaptive manipulation to iteratively reduce the modeling errors and enhance manipulation perfo
€™s robustness to noisy inputs. As a result, given a single top-down view depth image, our model can generate diverse grasps
experiences, which are arguably the ones providing the largest amount of information. We evaluate our proposed approach on
domization, which is based on GraspNet-1Billion. Our experiments demonstrate that ASGrasp can achieve over 90% success ra
e reinforcement learning and imitation learning algorithms. In both cases, we benchmark the proposed method using RLBench
ctively perceive the object from multiple perspectives during the manipulation process. This enables the estimation of 6D object
ralizability to novel categories and instances. Experimental results demonstrate the effectiveness of our framework in both simu
 s, on a real-world setup with a UFACTORY xArm 6 robot arm.
olutional neural network, and finally refines the predicted action with reinforcement learning. The proposed method shows signi
g from the high-efficiency encoding in the image space and focusing on points in local graspable regions, our framework can pe
del can be further improved in unseen scenes and the amount of manual annotations can be significantly reduced.
zero-shot success rate on previously unseen objects, an 85% mean zero-shot success rate on a class of previously unseen adv
erformance outperforming the existing methods. We also conduct a thorough ablative study of the proposed architecture and pr
mage patches, the feature maps extracted encode crucial class-wise statistics. To this end, we propose class-aware feature con
aining data problem. In CrackNex, both a support prototype and a reflectance prototype are extracted from the support set. The
the binary attention module undergoes a degradation of the self-attention module due to the uniform distribution that occurs af
erimental results show that our method outperforms the state-of-the-art baseline methods in 3D radar super-resolution tasks. Fu
dicate that our method can surpass the state-of-the-art in recognizing nouns and modifiers from images. Interestingly, our meth
d on the estimated confidence. Our evaluation shows that we outperform other self-supervised RGB-only methods, and approac
does not use prior information.
ow that COV-NeRF matches the rendering quality of modern NeRF methods, and can be used to rapidly close the sim-to-real g
ks, such as peg pick-and-place, unstacking bowls, and flipping slender objects, TAVI achieves a success rate of 73% using our
experiments validate the robustness and effectiveness of our contact estimation approach across various scenarios, significan
ts. The results from experiments underscore the superior efficacy of our method in both simulated and real-world scenarios, ou
ement noise, which is inevitable for current tactile sensors.
a simulation environment in Gazebo for robot training, applying the learned policy directly to our real robot. Notably, this work re
earns the system's inverse Jacobian through small exploratory motions. Our system approximates the underlying inverse Jaco
tile context and the other key module applying the attention and mask mechanism to the interaction of visual and tactile tokens
ts that are used in a very popular board game.
wrench on the screw via solution drawing upon dynamics. This wrench, adjusted based on position/force feedback, is mapped b

ext of transfer learning, our paper provides empirical evidence indicating that the proposed skill framework contributes to enhan
provide a thorough analysis of learned policies, reward functions, and reduced observations for a comprehensive understandin

an be used to linearly parameterize the closed-loop system. Thus, an adaptive controller with proven stability can be designed
esolves (iii), where we show that with smoothing, classical sampling-based motion planning can be effective in global planning.
our framework allows the incorporation of expert knowledge and customizable setups in task mechanics and models. It require

n multiple UAV modules are assembled. The latter provides guidance for the UAV agent to plan a collision-free and efficient tra
two latency sensitive case studies: (1) Collision Avoidance: a robot arm guided by visual feedback from consistent distance es
umber of agents, however, the memory and computation time per agent remain constant. Future work will investigate the ability
ness of the algorithms. Superior scalability to large-scale systems against other DDP and sequential quadratic programming m
hat are only perceived online, a heterogeneous graph attention network based heuristic is learned to select more appropriate pa
n, deadlock-aware guided map integrated with a gradient-optimized, centroid Voronoi-based coverage control policy, is construc
when new factors are added to the graph and is scalable with respect to both network size and time. We validate LC-ADMM an
while DHBA exhibits better reliability and lower overhead in harsh communication conditions.
ve control is employed along with physical constraints of the custom-built robot to enable real-time prediction and optimization o
rmal function is employed to approximate the control input sequences in the controller to alleviate the computational burden. In
aton. This neural network generates predictions that serve as guidance for random tree construction, directing the sampling pro
es us to apply the standard Q-learning algorithm. Theoretical bounds in terms of the approximations are also derived. We prese
e of our model is analyzed with a real robot (UR5e).

obtain O(n+m) complexity solvers. Our numerical results indicate that significant simulation speed-up can be achieved for high
h structural flexibility. A robotic model is employed as an example to demonstrate the co-simulation scheme, where a non-smoo
our approach leverages emulated rehearsals, where the robot learns within the latent space to predict its rewards and generate
ecution. The key insight of this work is that by adaptively adjusting the robot's compliance control gains during task execution, w
proposed method is experimentally verified by using the peg with various shapes and sizes.
ved network is dedicated to enhancing the accuracy of initial pose estimation, while the optimization technique focuses on refin
We further demonstrate the applicability of ASAP on both simulation and real-world robotic setups. Project website: asap.csail.

Matching framework for Multi-session Mapping, namely MM4MM, to estimate the relative transformation of multiple single-sessio
fense interactions between a patrolling agent and an attacker. Our results show that the data driven patroller can effectively pro
reas allocated exclusively to individual robots. This problem in formulated in terms of minimizing the maximum weighted idlene
sed framework outperforms distributed control methods by 40-100 percent in simulation and physical experiments.

assifier can be used for single-contact as well as multi-contact impact scenarios, such as those occurring in surface-to-surface i
e transferable to other robotics applications as well. Additionally, we present MVT-Flow as a new baseline method for anomaly d
propriately. In this work, we focus on detecting and classifying both manipulation and post-manipulation phase failures using th
ments that FT-Net learns to effectively perform recovery behaviors. The fault-tolerant locomotion can even be executed in variou

enough to detect LOC in more nuanced cases, where the quadrotor remains undamaged but nevertheless losses control throu
AVs violated 71% of properties during at least one test. Artifact available at https://github.com/less-lab-uva/SGSM

path planning to each frontier. The proposed method achieved higher explored volume generation than the state-of-the-art algo
ne. Once perched, rollers in the fingers of the perching mechanism drive the UAV along the powerline, providing a close-up pla
erceptual modeling framework using Gaussian mixture models (GMMs) as well as a novel systems contribution, which consists
traversal performance. Additionally, map sampling, terrain simplification, and tracking controller modules are integrated into the
ntrolled updates or leaving the degenerate directions of the optimization unchanged. The proposed method is thoroughly evalua
based on the lessons learned through several hours of field use. The methodology is validated through four field trials, including
roviding substantially more stable gradients than conventional concentration gradient-based methods. We demonstrate the sui
d speed and robot velocity, the tidal range also has a significant impact on the navigation of sailing robots.
path and optimizes the collision-free static trajectory. Finally, the static trajectory will be fed into the low-level local planner to av
experimentally validated in terms of its maximum elevation at different azimuths and its ability to balance different loads. The re
ed components. Our experimental results with images from operating commercial plants show that we can automatically detect
equired mechanical energy for robot motion. We evaluate the pipeline and understand the distribution of energy consumption u
igns the platform's attitude with the UAV for stable landings, even on small platforms under windy conditions. Secondly, the rob
amework is designed to iteratively identify the next best $T$ locations to sample, which prioritises the potentially high reward lo
 t monitoring strategy that provides the days and locations of when and where to sample with the robots by leveraging Gaussian
our results show the importance of a reference point cloud in the calibration method needed to estimate the vertical translation
smoke plumes and discuss broader implications for this type of system in similar applications.
navigable. By using a sequence of images, WayFASTER significantly enhances the robot’s awareness of its surroundings,
phasis between vision and tactile perception modes. We conducted a performance evaluation of the proposed sensor across th
me touch discrimination, and touch modality recognition. The experimental results indicate that the sensor performs well in dete
e system achieves classification accuracies of 94% on a dataset of 11 natural textile textures. Furthermore our system is capab
cs to analyze the vector field produced from 2D marker displacements. Despite the simplicity of our technique, we demonstrate
thout domain randomization. We experiment with the trained agents in harsh pushing conditions, and show that with significant
nsparency-optimized bilateral architectures.
nd reflecting it to the same camera viewing the elastomer through a mirror. As a result, the tactile and force sensing modes can
ls along the entire finger. Overall, GelLink is a compact robotic finger that shows adaptability and robustness when performing
s for accurate depth map generation when compared to alternative lighting methods commonly implemented in previous camera
struct a shared latent space for task stage classification, object dynamics estimation, and tactile dynamics prediction. Through e
e, the robust design of AllSight provides it with a unique zero-shot capability such that a practitioner can fabricate the open-sou
uxiliary markers or patterns on the gel surface. More specifically, we collect a dataset consisting of approximately 100,000 imag
ness of the GelFinger sensor in assessing large surfaces by using it to reconstruct curved crack outlines. Comparative experim
ensor’s internal lighting conditions. We describe the sensor’s fabrication and evaluate its performance for both tactile p

ments of marked pins, similar to the function of intermediate ridges in the human fingertip. We acquired images of markers' mov
eal-world settings with a zero-shot sim-to-real approach. Our analysis highlights and addresses some challenges met during the

rain attributes. The results demonstrate that the attributes extracted using this method strongly correlate with the actual softnes
n-hand object pose tracking. Experimental results demonstrate that TEG-Track consistently enhances state-of-the-art generaliz

up to 9.87 N, with a high sensitivity of 141.7 pm/N, good repeatability, and a hysteresis of 7.96%. Compared to commercial load

r method, showing 26.9% improvement in multi-touch force-sensing accuracy for the tactile skin.

maintaining accurate contact positioning. The generated images can be used to train zero-shot models for newly fabricated sens
libration processes, we propose a self-supervised photometric stereo method based on deep learning, which eliminates pre-cal
simulated depth image sequences into RGB tactile images with markers. Furthermore, the generator in the proposed GAN allow
exterity under visual occlusion, laying the foundation for closed-loop behavior in applications such as regrasping, insertion, and

t functions are explored: the visual cost defined by garment corner positions guarantees the gripper to move towards the corner
rough a series of experiments, we showcase the effectiveness of our method and its superiority over existing state-of-the-art se

ed on image feedback, where the converged orientation of IFP directly harvests the data for solving the hand-eye relationship.
 and asynchronous sensor-pair cases. Comprehensive numerical studies are conducted to verify those identified motions, show
control scheme is proposed to address challenges in the nonlinear control arising from the interaction with the deformable obje
and decoupling error in the other axes are 0.442 %R.O. and 0.660 %R.O., respectively, and the performance is comparable to t
ng method is robust to friction and can achieve a mean (±SD) displacement error of 1.04±0.43 mm in contact location with c
bance conditions.
ence on accurate dynamic models of the manipulator and the contact force. Simulation and experimental results demonstrate th
pose the quantified analysis metrics for the process by analyzing them as a typical example of flexible manipulation disassembl

. The case studies and benchmark results show that the algorithm is general and consistently outperforms the state of the art.

h for the use case of autonomous driving with predicates used in formalized traffic rules on a recorded dataset, which highlight
nterest onto a single tangent (Euclidean) space, over which an off-the-shelf learning algorithm is applied. This paper provides a
h avoids the state-space explosion associated with explicit product-automata construction, thereby efficiently solving the proble
work for agents to make informed decisions in choosing the most suitable edge attributes (travel durations and success probabi
pected and maximum network latency. Our shield is minimally intrusive and enables networked robots to satisfy key safety con
n Markov Decision Processes, where the navigation durations between physical locations are uncertain, but the time-dependent
task into learnable sub-goals, which not only converts non-Markovian reward decision processes to Markovian ones, but also im
ws us to make fewer modeling assumptions and leads to more natural and powerful models of interaction. We introduce the se
ely. Based on the established proposition graph and perceived environment, the agent can autonomously generate an LTL form
us multi-robot systems. It is shown that the solution time of finding a satisfactory task allocation grows almost quadratically with
only used novel tasks without violating any safety constraints. We deployed LTL-Transfer at the task-planning level of a quadrup
bution changes with the angle; a complex path with varying distances and angles was employed. The results confirmed that the

oncert to answer those queries. We demonstrate these ideas in a simple case study and present data illustrating how plan qua

sification accuracy by 39.1%. Furthermore, the Tactile Height Sensing (THS) module and the Tactile Position Exploration (TPE
joints, a very simple alternative to most variable stiffness mechanisms. The resulting modification of the geometry and kinemat
l is trained entirely in simulation and is evaluated on simulated and real-world objects that are not seen during training. Simulati
ributions are efficient to compute and perform comparably to particle-based distributions in a grasping domain. Finally, we depl
ely with over 70% success rate. Real-world experiments demonstrate that the model trained in simulation performs satisfactorily
arkably, our framework significantly outperforms several strong baseline methods in terms of density and reliability for generated
ops in algorithm performance, thus making them important attacks to consider in the cybersecurity of grasping robots. Source co
main- invariant image embeddings, image-text matching and fusing predictions from different sources. We perform experiments
SAGED enhances object feature information by considering global contexts. Secondly, the Graph Sampling Aggregation metho
0+1 N and contact locations with over 96% accuracy. We also demonstrated the performance of AST Skin technology for a rea
ransferability. Key challenges regarding the reality gap for grasping have been identified, stressing matters on which researche
tives trained on synergy trajectories of human grasping motions. The representation using synergy primitives allows for a straig
e of improving grasp performance for static scenes by refining the initial grasp proposal. We also show that it can enable graspi
cific class of Fractal Fingers and Hands.
ng direction as an input and generates an instance-centric representation for each partially observed object in the scene. This re
Gripper to enable training of learning models and to demonstrate the capabilities of the GRM. The dataset includes ranges of gr
and anticipate failures. Unlike prior work that cannot foresee future failures or requires failure experiences for training, MoMo le
ent learning method to learn generalizable grasping policies. The experimental results demonstrate that our proposed represen
the uninstrumented real world. Building on the latest algorithmic advancements in model-based reinforcement learning (MBRL)
Models and Gaussian Processes. This method resolves the challenge of discrete grasp pose selection, enabling smoother gra

he object. By leveraging the contact maps of different fingertips, the proposed optimization algorithm can generate precise and
grasp planners. Results show that jumping regrasping improves pick-and-place success rate. Haptically driven jumping – usin
y nearly 30% and achieved over 150 successful grasps with three different object categories. This demonstrates our approachâ
h allows us to carry out several distinct evaluations from a single set of experiments, without having to repeat the data collectio

urthermore, the validity of the analysis and methodology are experimentally demonstrated. The gripper performance is also eva
size and stiffness of the object can thus be accurately measured. Simulation and experimental results demonstrate the propose
vailable Festo MultiChoiceGripper, which features a Fin Ray structure. The total contact length between the fingers and the gras

theoretical model. A series of grasping tests with a vibrating-surface equipped parallel jaw gripper confirmed the design guideli
ode grasping and incorporates a liquid metal strain sensor (METT) for real-time monitoring of finger bending angles. Comparat
cessity of the magnetic gripper model for precise manipulation, verified its excellent sensitivity to magnetic fields, and demonstr

his system both virtually on several test parts and physically by destructively and non-destructively disassembling a model of an

aching direction of the gripper. DSG capabilities are verified with experiments conducted using real food ingredients within a pic
tual tools, we co-design manipulators, grippers, and multi-fingered hands, showing that environmental constraints are an effect

each chopstick. We also evaluate the versatility of HASHI through grasping trials of a variety of real and simulated food items w
n, design files, detailed assembly instructions, trained models, task videos, and all supplementary materials required to recreate
is effective for grasping various objects and harnessing tools.

nt datasets by simulating human rubbing movements with the thumb and index fingers of an Allegro hand. Our model also dem
e to calibrate the misalignment between the exoskeleton's and the human-finger's base frames and estimate the length of huma
nd able-bodied participants. The findings highlight the potential of dual haptic feedback to enhance sensory perception in prost
phase and amplitude modulation, determined using the simulated annealing algorithm, is employed with five actuators in a spa
aining curriculum improves performance during the training process and, crucially, promotes the transfer of the acquired skills t
articipants use Hapstick to teleoperate a physical tendon-driven continuum robot in a simulated colorectal cancer screening tas

roller, we train perceptual models in near-real time. We demonstrate our approach using a biomimentic sensor, exploring "tactil
istent with our display, with a mean RMSE of 6.68 mm or 3.85% of the display’s maximum vertical range for shapes we cou
nsmission and human-computer interaction to improve the overall system performance. The experiment results on real-world sc
nted by an Adaptive Frequency Oscillator (AFO) based position controller that leverages the periodic nature of human gait and i
addressing global challenges, particularly in labour-intensive and hazardous tasks. By improving the integration of humans and
essing realtime control of a wearable anthropomorphic robot hand. The offline results demonstrate that the proposed self-supe
 ming the desired trajectory. We then present a reinforcement learning approach that uses the human-robot coefficiency score a
Predictions with Intersection over Union (IOU) values below 0.2 undergo an assessment by human operators. After their evalu
me, despite having no access to the expert at training time. To the best of our knowledge, this is the first instance of using the v
schedule for novice users (in the Training study), and were able to determine the minimum training time necessary to observe
a simulated experiment in which a robot repeatedly carries an object for the human to work on, and the object Cartesian pose n
nal human's intents, more robust reward function representations are achieved. We compare our approach to recent state-of-the
o proceed autonomously or request expert assistance at estimated times of uncertainty. An implicit estimate of uncertainty is le
models trained with CARLA-VR improve displacement error and collision rate by 10.7% and 4.9%, respectively, and are more ro
a robust adversarial reinforcement learning framework. We further utilize conformal prediction to quantify human preference un
ible, and safe trajectories periodically. Additionally, a yaw planning that provides the pilot with the best suitable view angle is em
setup with 2 kg, and 3.9 kg, and a human subject with 0.5 and 1 kg. Next, the overall scenario including load-compensation an
mpliant contact experiments to validate the proposed method and demonstrate the unexpected contact response in humanrobo
ictions. The approaches are evaluated and compared with each other, by using a KUKA LBR iiwa robot. The results highlight th
ther analyses show, that even complex scenarios can be covered with our framework.
conducted experiments using standard benchmark datasets like ETH, HOTEL, ZARA1, ZARA2, UCY, and RGB-D pedestrian
(2) updating the value network to estimate and utilize pedestrian-unpredictability features, and (3) implementing a reward functi
eople's decisions more accurately, corroborating previous theoretical results that CPT is more expressive and inclusive. We als
als. Experimental results demonstrate the model's proficiency in understanding floorplan topology and human behaviors.
ware navigation methods using MAC-ID, we present a novel benchmark called BSON. It offers realistic and diverse social enviro
ue to their iterative nature and need for gradients. We present Interactive Joint Planning (IJP) that bridges MPC with learned pre
hest from the waypoint, an novel obstacle gradient is introduced to guide safer optimization. 3) By assigning edges with asymm
ased method is 4000 times faster and thus more attractive for interactive applications, even though at the cost of a slightly lowe

EGTrust is constructed. Experimental results indicate the effectiveness of the proposed approach, achieving an accuracy of 74
tion studies confirm the utility of key components of our pipeline. As shown in our open-source code, we achieve 97.2% accura
nd allows us to analyze group cohesion. A user study with 43 participants is used to validate group preference profiles.
ses capability-based human trust.
raining, allowing fast deployment in the interactive navigation tasks. We choose to use multiple traversable objects such as curt
model providing resiliency for HD-sEMG modules, which can be used in the wearable interfaces of neurorobots. The proposed 3
9.9% when less than 20% of the task remains.
ential element of ToM). We designed a dynamic trust-aware reward function based on different trust beliefs to guide the robot p

directions conveyed by the proposed guidance method. Subjects were asked to select their felt directions on a tablet while runn

rse range of food items with varying physical properties and can successfully adapt to different environments featuring multiple
d long-short term memory based networks. Results show the proposed network has better generalization performance compare
s in the LLM’s memory. It is created using prompts. In the explicit map, a topological map composed of nodes and edges is
MDP-based planner, optimized for target searching, is augmented with this reward map to generate an operator-constrained pol
s effectiveness across various users and defect types. To our knowledge, this is the first demonstration of the real-time applica
on Bluesky.
ch work cell and the practical limitation around running the experiments. Both work cells scored high in usability, expected incre
evels. The results uncover a strong correlation between sEMG and EoP. This paper also reports the effect of the direction of pH
and then iteratively selects and relocates the best view-blocking objects to provide the dense perception of the region of intere
in terms of spatiotemporal metrics. This is reflected by 0.1m/s increases in overground walking speed and 5% decrease in perc

by human-robot interaction dynamics, which considers the rules of social space. Furthermore, an obstacle avoidance compone

the belief predicates must be evaluated over the continuous (infinite) belief space. To address this challenge, we present an alg
ection and safe trajectory generation around human collaborators. Finally, we use our proposed proactive controller in a collab
mulated demonstration on the Virtualhome testbed show that the assistant can effectively improve team performance.

n scenarios where the robot is restricted from navigating in certain areas. To correct map imperfections resulting from problema
RGB videos that are leveraged in the training process. We show that utilization of the generated data during the network trainin
ore videos and generated behavior trees are available at:https://github.com/jainaayush2006/CoBT.git.
liant with the ISO/TS 15066. The proposed approach has been experimentally validated in two different scenarios: one compos
dy on a maximum entropy inverse reinforcement learning algorithm. The model generated enables to train an intelligent agent th
able method of underwater communication. TTS-Sonemes outperform Tonal-Sonemes at close distances but fail at further dist
eraction forces. System passivity is ensured by an energy-tank based task passivation strategy. The framework's efficacy is vali
ampling a care-receiver’s response in an adversarial manner during the training. We evaluated the proposed method using
ompleting the peg transfer task. By learning from the expert and collaborating with the human during the task, the intelligent ag
ond to a sequence of online language corrections that address failures in both high-level task plans and low-level skill primitives

e ergo-interactive control in terms of task performance and ergonomics adaptation is verified under different experimental cond
, and its effectiveness is assessed through user performance evaluation in a trajectory tracking task. Compared to direct contro
ncy and low-bandwidth communication. Through extensive experiments, we validate the robustness of our method and the res
commendations for subgoal allocation based on NL agent conditions. The framework, PlanCollabNL, is evaluated for a number
tions from our proposed framework, users are able to better explore the full space of strategies and collaborate more efficiently
een experimentally validated in both simulations and a real-world industrial task involving a Kuka LWR4+ robot.
oaching velocities by a safety control set. Second, a unified task control set is designed to achieve different robot tasks for diffe
ating the performance of the planned trajectories during physical interaction with the environment. Both validations confirm the e
perception, and completion estimation. Through comprehensive experiments, we demonstrate that discussions with domain e
ed results show that a robot connected to our proposed framework is perceived as a cognitive agent capable of manifesting pe
ble to greatly reduce the necessary amount of preferences, by permitting the highlighting of favoured trajectories, in order to re
can tackle both issues, allowing the use of pre-trained state-of-the-art text-image alignment models with robots to enable co-pa
g low cognitive demands and ease of use. Our MateRobot demonstrates the feasibility of recognizing material property through
ablishing a target arm end-effector position and dynamically adjusting the impedance parameters in real-time through a impeda
ing with the essential principles of incremental learning. We evaluate our approach to incrementally learn a VMP library from se
ace suit was constructed and tested with a SuperLimbs prototype. Results from the experimentation found that without assistan
that a mobile manipulator can adapt to real variation in human body shape, size, pose, and blanket configuration to uncover tar
uence of the serial-position effect on recalled competence (see overview at https://youtu.be/BgH2zhh1s48). Participants viewed
proach (LaCE-LHMP). Our approach is inspired by data-driven airflow modelling, which estimates laminar and turbulent flow co
 nvironment through human-robot interaction. The architecture builds on core cognitive principles of learning and memory for eff
parsing masks and 3D depth maps. Finally, we develop a humanrobot interactive drawing system with low-cost hardware, use
R-HMD). We trained 15 participants in mix of high and low noise virtual environments or on a simple training task, and found a

of a brush's translation, dipping and scraping. Unlike existing methods that concentrate on generating paintings from 2D image
m in a 3D mini map and drag it to a new desired goal pose. We validate the proposed system through a user study and real-worl
walking speeds and terrains in simulation and walking acquisition with hardware. All results showed successful on-site walking

zed humanoid JAXON in the simulator. The resulting control policy demonstrates a wide range of locomotion patterns, including
sses. To make the template behaviors extensible when the robot encounters possible different scenarios, the NC-MPC is propo
ts to avoid potential damage robustly with proper NWR, depending on the site and intensity of the stimulation. We also verify th
s and response times across different fault scenarios with a false positive rate of 0. The findings of this study hold significant imp
actuators drive the patches forward, which deforms the elastic skin layer. The tensioned skin layer inversely drives the bending
oCub, currently being developed at Istituto Italiano di Tecnologia. We also compare our strategy with the existing state-of-the-a
agent discriminators to reduce the Feature Gap. Our intensive experiments on the public multi-agent cooperative perception da
herent conflicts by preserving objectness clues that are lost in the fusion BEV feature.In general, our method maximizes the eff
icle features at scale, spatial, and channel levels to correct the pose error introduced by camera asynchrony. We also introduce
n-depth performance analysis of our method regarding different input sequence lengths, vehicle headings, and daytime conditi
lity of our method.
ur framework, we conducted various ablation studies and comparisons with prior methods on a newly collected dataset annotat
y, we present HPL-ViT, a pioneering architecture designed for robust feature fusion in heterogeneous and dynamic scenarios. I
ccelerating the model while keeping its accuracy. Targeted at improving this component, the time-consuming 3D convolution ne
amlessly integrated into a strong LiDAR perception network and benefit from the features extracted by the network. Experimen
and an image-conditioned cross-attention module to refine voxel features with visual clues, and (ii) employ a distillation modul
sociation, track lifecycle management, false-positive elimination, false-negative propagation, and sequential track confidence re
mpensation Module and Distribution-aware Statistical Consistency Module, both aimed at enhancing intermediate features to mi
nce the success rates of these attacks in a black-box scenario, we introduce three types of statistically sensitive natural discrep
pstream detection, tracking, and mapping modules). In this work, we aim to bring forecasting models closer to real-world deplo
query cooperation paradigm on transmission flexibility and robustness to packet dropout. We hope our work can further facilita
s in the estimation of the attenuation coefficient and the employment of fog-induced noise distribution for defogging. Experimen
equence-level to significantly reduce the communication cost. Furthermore, to tackle the redundancy of transmitted data, we in
theoretical optimality property of Kalman Filter-based tracking algorithms. In this paper, we propose a novel 3D multi-object co
Contrary to recently reported results, we show that emph{under proper training guidelines}, all vision architectures perform ind
d paired human-robot data for both semantic and temporal skill embedding alignment. Extensive experiments in a simulated kit
ocus on generating the optimal action at each time according to the maximizing Q-value principle. This motivates our looking ah

and actions. We do so by leveraging the videos to learn a reward function that measures incremental "progress" through a task
efinitions of similarity. Our experiments in a synthetic environment, an assistive robotics environment with domain transfer, and
enarios. As revealed by the experimental results of this study, the framework can successfully unfold fabrics of different colors
e fine-grain data about robot performance. Typically the performance of robot applications depends both on the quality of the c
l escape energy. We present 8 simulation experiments demonstrating the applicability of our approach to various complex quas

ameters of the spherical mechanism on the MAW and the transmission efficiency are obtained, and the parameters of the sphe
performance and better adaptation for real-world feature space. The reduced features are replaced with random signals from a
unded in nonlinear geometry. These artificial dynamics shift the uncontrolled dynamics of a robot via an appropriate control law

employ an adaptive torque-controller combined with a joint impedance regulator that guarantees high tracking accuracy while a
e application control in a modular fashion. The successful climbing of the robot proves the scalability of the proposed climbing r
rates of 92.3% for known cables and 88.4% for unknown cables, surpassing contemporary state-of-the-art approaches. Supple
uracy achieved even without trial-by-trial learning.

city scenarios. And the other is a navigation module, founded on the Heterogeneous Agent Reinforcement Learning (HARL) fra
of consistency enable our model to adaptively enhance image features, ensuring more accurate and visually pleasing results. E
ared to 54.4% without the proposed adaptive control.

e and maneuverability of the arm. The operator can also command the movement of the end-effector, sliding on the surface geo

egions of interest. Exploiting this representation, we evaluate a candidate trajectory around key factors such as collision avoida
essing hundreds of constraints simultaneously.
el consumption. The physical model free NPC outperforms the base PCC method with 2.41% and 3.45% more fuel saving in sim

ent. The efficacy of our approach is demonstrated through experiments involving simulated high-dimensional quadruped robot

nd real-world plant watering. Results demonstrate an improvement of approximately 80% after only four real-world trajectories w
date paths and speed profiles to produce time-optimal trajectories while maintaining a bound on the future expected risk. We as
andom Forest surrogate models that drastically improves the results.
y reduced by up to 89% with the proposed Motion Memory technique and with increasing past planning experiences.
ed decoder, we demonstrate the effectiveness of our methods in reducing causal confusion and improving performance in plan
ency and path optimality. For environments represented by point clouds, MNP first efficiently compresses point clouds into a la
ined dynamically based on a learned estimate of reachability, and subgoals are distributed to focus on challenging parts of the
h finds better solutions faster and leads to a higher success rate in execution.
ch bounds is made tractable by the use of a hierarchical grid on SE(3). Our method is evaluated both in simulation and on a rea
setting. Although being artificial and retrospectively constructed metrics, we interpret RUL and SOH as physical measurements
cept electronic task board with industry-inspired tasks used in an international competition in 2021 and 2022 to benchmark loca
onlinear PDE models with high computation speed. The presented controller is robust to parametric model uncertainty due to it
s, and an adjustable stitching interval ranging from 1 mm to 5 mm. The subsystems and components are interconnected using
ng between path lines. The approach was evaluated in simulation and practical experiments on the DLR SARA robot. The resu

experiments demonstrate the effects of different fallback strategies and show how our proposed approach is able to ensure saf
new conflict prioritizing rules and a heuristic value calculation method that derived from following vehicles. Finally, we introduce
tly improves the performance over current state-of-the-art approaches.
enarios at query time while eliminating scenario redundancy in the database (in our case study, given the same ODD query, the
adopted. These analytical results leverage queuing-theoretic analysis, order statistics, and a conservative, reachability-based a
ilities by conducting simulations, employing real-world driving data to validate the accuracy of our approach in modeling human
temporally align predicted trajectories via a masked attention operation. Using the common Argoverse benchmark dataset, we
rch strategy to determine the optimal passing sequence. It is noteworthy that the travel efficiency is enhanced and the computa
esy of driving behavior, leveraging marginal and conditional behavior prediction models trained from real-world driving data. The
f the driver. With the cognitive digital twin, the CDAS warns the driver when there is a significant discrepancy between the predi
during training. Therefore, we exploit the reachable set-based time-to-collision (TTC-RSB) metric from the motion domain along
ributions. In addition, a sampling consistency loss is introduced to further improve the segmentation performance and reduce th
two tasks in the wider spectrum of object states, we extend them with new metrics called S-MOTA and MOTPS that address t
the compression ratio based on the channel state information (CSI). Subsequently, we devise a judger mechanism to filter the

mate boundedness of closed-loop errors in the presence of time-varying exogenous disturbances, while guaranteeing asympto
hich share their data acquired on the road feeding a machine learning model to identify autonomous cars. We extensively tested
elease our code and model at https://github.com/LJacksonPan/RaTrack.
der image observations, a modality that has not been extensively explored for mixed traffic control via RL, as the alternative: 1)
separated pre-trained perception models. 2) Instead of directly letting LLMs drive, this paper explores a hybrid setting of letting
ing methods but outperforms other learning-based methods in planning quality, while operating more efficiently in terms of runti
lection of relevant agents from the scene and passing them through an attention mechanism to extract global encodings. These
change in minADE as compared to the original. In addition, we investigate techniques to improve model robustness, including in

em control. By analyzing the performance of MRTL across nearly 600 signalized intersections and 1200 traffic scenarios, we de
ithm that reduces computational costs, while still achieving a stable near-optimal policy. Our approach partitions the graph into
set of standardized sub-modules for flexible implementation choices: AV Evaluation, Scenario Selection, and AV Training. CLIC

al query scheme in context fusion for a more efficient and lightweight network. We evaluated our approach on Argoverse 1, Arg
enging CommonRoad traffic scenarios as well in the realistic CARLA simulator. The results demonstrate that our decision modu
ch leaf node of eMem-NDT is modeled by a neural network for aligning the behavior memory prototypes. By eMem-NDT, we in
mong the generated warning, the driver behavior, and the states of ego and surrounding vehicles on a long horizon. The warnin
roposed methodology through comprehensive numerical simulations conducted in both high and low-fidelity simulation scenario

generates high-quality lane-change maneuvers such that the vehicle merges into the traffic flow with a high success rate of 96%
model (GAVA) outperforms state-of-the-art baselines by at least 15.2%, 19.4%, and 12.0%, respectively. Our findings undersco
allenge, Lyft Level 5 dataset and Waymo Open Motion Dataset to show the effectiveness of our approach and its state-of-the-a
M model builds the relationship between sEMG signals and continuous movements. To verify the effectiveness of the CNN-LS
experiments demonstrate RT's effectiveness in both intent estimation and shared-control tasks. We will make the code and da
 intended motion solely from EEG data with 70% accuracy, triggering a robot to execute an assistive task. We provide a compr
on images to generate macroexpression images. We compare our single input based network with current two-frames-input bas
es alone. Furthermore, our models predict driver maneuvers more accurately and sooner than existing approaches, with an acc
g LLM-enabled methods by exploiting the ability of LLMs to interpret the user interaction in the context of the objects in the scen

uxiliary tasks designed to learn more robust representation of scene dynamics. We conduct empirical evaluation on common be
ntropy to gauge visual concentration, thereby adjusting the contribution of gaze fixation to the HMIP. Empirical experiments sub
per closes within 45 milliseconds after initial contact with the impacting object without requiring any controller or actuation energ
onal sensors. We assess the platform's efficacy by comparing it to a feedback control based on position data provided by a mot
actuator and allows for the determination of effective torsional stiffness. The experimental validation showed that the VS-SCA
bbing. There is growing demand for automated sanitization systems in hospitals, food-processing plants, and other settings whe
esearchers in the field of cable-driven continuum manipulators.
optimization method verifies the correctness of the proposed theory. Finally, an optimal design of a tendon-constrained synchr
promise between compliance and payload. An open loop control strategy is proposed, which can theoretically decrease the stif

is modeling attributable to cable material properties.

data from knee and hand joints validate the optimizer's efficacy. Additionally, our optimizer is multifunctional: it streamlines the
iments. The fabricated robot weights 91 g, and its length, width and height are 220 mm, 410 mm and 95 mm respectively. The
e precisely, the temporal region causing the largest position deviation is identified and updated at each iteration. In our evaluatio
ing to the general PCC model. Note that this study is grounded solely in simulation.

kinematics optimization. Moreover, to show its practical use, two opposing robot architectures are chosen: (i) Serial robot - Uni
f modular robotic parts. Novice users could easily apply a minimal robot arm to various tasks. Two test cases of common manu
uring motions. The results show a more robust detection of consecutive sub-motions with distinct features and a more consisten
s subjected to different boundary conditions, a non-isotropic helically-reinforced tube, and a not-purely-kinematic scenario with g
tia matrix at a desired position while following a motion plan constrained to the robot limits is presented. The system is tested fo
nts that its design, speed, accuracy, and energy efficiency are appropriate for this application.
c model. The quintic B-spline is utilized to generate trajectories in the virtual passive-joint space, while the Genetic Algorithm is
happens in rigid-link parallel robotics. The stability prediction results are validated in simulation by cross-validating them by usin

he maximum allowable MP speed map of the workspace. Finally, a varying speed trajectory is designed based on the metric to
hat the presence of a passive tail helps improve energy efficiency. The system design, controllers, and software implementation

rollers and energy functions. We conducted simulation experiments on Safety Gym and real-robot experiments using small qua
ng a single-step convex quadratic program. Our method is evaluated on the first and second-order systems in various scenario
timated risks is used for the robot to select the best control policy for safe manipulation. Our experiments show that the propose
ng. A novel aspect of our system that is critical to robotics applications is that it links the predicted actions to specific locations i
ng modality use. Furthermore, in contrast to the non-modality attention model, whose task success rate drops significantly unde
the weight and length of tools, enabling us to understand the characteristics of the grasped tool from the current sensor inform
sferring parameters to more complex tasks with a larger number of parameters and even between two different tasks, which is
arable with or exceeding state-of-the-art deep reinforcement learning-based quality diversity algorithms.
mitation learning based on the collected data. We implemented an overall system using two Franka Emika Panda Robot Arms a
its use with arbitrary DNNs and allowing it to learn complex motions with high accuracy. We validate it using datasets and a rea
rations. These demonstrations include the movements in the task scenes with various target positions and obstacles. After mov
rimental results show a high generalization capability of the proposed model in terms of the number, positions, height distributio

udy case since it involves complex two-agent physical interaction/coordination and because its periodic nature can facilitate rep
mproves the ability of the LfD model to recognise physical properties and generate desired wiping motions for unseen sponges
m those samples. Then, we recursively use the simple yet effective neural networks to march toward the start and goal bidirectio
at requires minimal safety data labeling. During deployment, we employ a lightweight online optimization to find actions that en
emonstrate that our approach, namely RLAP, is able to extract rules without explicit supervision of rule labels and generate goa
be used to compensate for varying payloads while maintaining high backdrivability. The proposed method is evaluated on a co
re standard approaches fall short.
echniques and their downstream effects on tissue. Here, we introduce a parameterised model of the elliptical excision skill and
egularizes the forcing term, and improves the interpolation accuracy of parametric DMPs.
0 minutes. Experiments demonstrate that our approach enables the model to learn a more general and robust policy across the
f varying severity, more accurately than state-of-the-art constraint inference methods.
acts time-varying probability densities from the 2D sketches, applies ray-tracing to find corresponding regions in 3D Cartesian s
ntary information can be found on the website: https://sites.google.com/view/tax-posed/home.
iterature, positioning it as a promising solution for scalable and stable policy learning in complex environments.

ntial data to learn demonstration-conditioned visual policies, allowing for in-context adaptation to a target domain that is distinct
while being robust to perturbations and safe around humans and obstacles.
ness of the proposed algorithm, rigorous assessments are conducted using the LASA dataset, further reinforced by empirical va
arned ADS by utilizing the substantial flexibility of the neural energy function. The effectiveness of the proposed approach is va
with accurate decomposition compared to the next best approach in novel scenarios. Further, we demonstrate the robustness o
ess of an existing ERTC planner when transferring it to the new problem. The results of three simulation experiments revealed
o predict "intent embeddings", which roughly correspond to subgoals, via planning. We show that augmenting state representa
o a significant improvement in latent-variable skill discovery for robotic manipulation while overcoming possible interference occ
hanical properties. We demonstrate the proposed method minimises process force and path deviations to a level similar to offli
er learning of new long-horizon tasks than previous pre-training approaches. Website at https://clvrai.com/sprint.
earning and RL. We refer to this approach as Reinforcement Learning with Teacher-Student Collaboration (RLTSC). RLTSC in
bot exhibits walking, running, sitting and jumping, much like the way humans interact with a pet.
matching dataset is collected of domain-randomised image observations. The teacher world model then supervises a student wo
robot is pushed to gather information when uncertain, predict intent from noisy measurements, and anticipate in order to interce
mechanism into the sequence modeling-based RL methods. By utilizing the EM mechanism, we propose a novel training object
endently determine its actions. Specifically, we incorporate a self-attention mechanism to enhance the cooperative behavior bet
iment with the Go1 robot, we demonstrated better control and dynamics of the bot with our reward framework.
mulated and real-world multi-robot scenarios validate the effectiveness of our proposed framework.
t the neural policies estimated by HyperPPO are capable of decentralized control of a Crazyflie2.1 quadrotor. Project website:
/sites.google.com/berkeley.edu/aprl
 y without sacrificing performance. Experiments on quadruped and humanoid robots demonstrate that the learned policy is robu
large-scale datasets, thereyby helping facilitate the advancement of robotics in areas such as self-supervised learning, multi-m
apabilities. Specifically, both the UIE and UOD task heads share the same network backbone and utilize a lightweight design. F
bility map. Through rigorous simulation experiments, the proposed method (GS-PKNN) demonstrates its remarkable effectivene
ce risk-aware traversability estimates, and refine our feature predictions over time. We find that our method outperforms traditio
rios, our proposed D-GLSNet can output high-quality matching, enabling GLN to function stably and deliver more accurate loca
s in a single model. We validate our method through real-world experiments involving four ground robots navigating an alpine e
d into a complete mobile robot system and deployed to a real robot. Experiments in simulation and the real world demonstrate
pretable objectives so that MATRIX can generate human motions based on diverse destinations and heterogeneous behaviors

a goal-conditioned RL algorithm combined with Disentanglement-based Reachability Planning (REPlan) to solve temporally ex
ber of objects, novel backgrounds, and unseen object geometries. Experiments demonstrate the effectiveness of our model in a
ethods highlighting its modularity, (iii) are fairly insensitive to different intrinsic scaling parameters, and (iv) maintain robustness
Our method incorporates a temporal contrastive model and introduces a surrogate loss to extract task-related latent representat
eves higher success rates (87.2% on average) and has a 75.6% shorter inference time. We also demonstrate that our method p
pts. Furthermore, we show robust zero-shot transfer to novel objects. Videos of our experiments are available at url{https://malte

a tightly coupled predictor which `observes' the state of the robot leads to significantly improved anticipatory action, especially w
the proposed method to contact-rich tasks where a three-fingered robotic hand manipulates previously unknown objects. Our
ate this approach within the Meta-World framework and compare it with other typical multi-task algorithms in terms of learning e
these trajectories, our offline RL method discerns skill termination conditions and coordinates skill transitions. Our evaluations h

g, significantly expediting the planning process. Our experimental results demonstrate that our approach outperforms existing m
rientation from the source demonstrations. In simulation, we evaluate our method in three manipulation tasks involving deforma
ss (TBRR). We show that TBRR-aided DRL training is robust to aggressive attacks and outperforms baselines on standard DRL
-to-real gap, achieving large 3D deformations up to 40 cm in real-world settings. Experimental results also show that MultiAC6
lete processing pipeline can help to incrementally update the place recognition ability, robust to the increasing complexity of lon
ound the robot provide a compact and informative state representation for subgoal selection. We study the effects of different c
ped robot, Solo12, and exhibit diverse agile behaviors with successful obstacle traversal.
errains to exhibit improvements in motion efficiency, evidencing the efficacy of tactile perception in terrain-adaptive locomotion
we perform metric localization by predicting pixel-level flow vectors that align the query radar scan with the LiDAR map. We ext
ames requiring optimization. Furthermore, the proposed method establishes local map consistency by repeatedly optimizing key
RL) where a distance-to-target-based cost (standard in e2e) is augmented with physics-informed reward shaping. The key intuit
specially in challenging scenarios with corners and dynamic obstacles.
tion-driven applications, especially in the dearth of training data. We share more qualitative results at https://raaslab.org/project
agent can outperform the expert, and (c) our agents negligibly overfit to training data.
behaviors. Our behavior-conditioned policies outperform existing works in these challenging scenes, reducing potential collisio
he well-established object detection model, YOLO, with a Proximal Policy Optimization (PPO) agent. The results indicate that th
d safe policy optimization, especially for challenging tasks. Numerical simulation and physical experiments demonstrate that FS
botics, where symmetry lies mainly in task execution and not necessarily in the learned motions themselves. In particular, we in
with the number of objects. Further, we design an attention-based network to model the dependencies between the input states
 tasks such as autonomous driving. To this end, we propose Guided Online Distillation (GOLD), an offline-to-online safe RL fram
of real-world training data.
ence. In this paper, we propose Multi-Agent Neural Topological Mapping (MANTM) to improve exploration efficiency and gener
nually labeled into rigid and non-rigid terrain. The depth estimation task is trained in a self-supervised manner by utilizing the fu
ss rate by up to 61 percentage points over the baselines. Simulation and real robot experiments suggest that our policy learns
soft quadrupeds, whose inherent compliance and adaptability make them ideal for dynamic tasks but also introduce unique con
s shown that the method is successful in generalizing to new tasks and is also transferable to other robot dynamics.
-of-the-art methods. Using domain randomization during training we ensure a successful transfer to the real world, where we a
deos, following natural language instructions, and physical interaction. The video is available at https://sites.google.com/view/bi
domain randomization, our learned policy can be deployed to a quadrupedal robot, leading to a quadrupedal robot that can tak
onditions in real-world indoor and outdoor environments.
nd stable gait skill transformations. Experiments demonstrate that the control scheme is robust in unexpected scenarios where a
skill library. We show that our framework learns joint-level control policies for a diverse set of motor skills ranging from basic lo
otion. In our simulation and hardware experiments, the compliant policy achieves more robust, energy-efficient, and safe intera
ery, and conveyor terrains that require careful coordination of footsteps. We also showcase that IFM can efficiently produce mo
motion with adaptive behavior. A hallmark of our approach is robust blind stair climbing through swing foot reflection. Moreover
can overcome imprecise sensing and actuation to output highly precise control behavior end-to-end. We show our robot can pe
eter, which enables to adjust the robot's behavior dynamically. Importantly, our approach removes the need for additional rewar
Aliengo robot demonstrate that our method is efficient in handling various external disturbances, and the resulting policy exhibi
he learned policy excels in task performance while conserving energy, all without requiring manual weight adjustments for both
rom goal-oriented navigation in confined 3D spaces. To address the inefficiency of tracking distant navigation goals, we introdu
especially in scenarios involving crossed-leg maneuvers. Furthermore, it demonstrates versatility in tasks such as locomotion on

sh recovery. We also present an efficiency comparison between the proposed control strategy and our baseline walking contro
kes the framework extensible and enables the efficient waypoint computation in matrix and algorithmic form. The feasibility of th
effectively tracks a wide range of walking speeds and demonstrates good generalization capabilities to unseen scenarios.
he best quantitative result among other LIO-based methods and biped robot state estimation algorithms. The dataset and sourc

e impact of spinal flexion on robot balance during trot gait. Subsequently, we develop an optimized controller for spinal flexion, d
videos into motion primitives that can be executed reliably by real- world robots. We demonstrate our approach by transferring t
safely descend a step height of 40cm, which is 80% of its leg length. To the best knowledge of the authors, this is the first time
all, we consider SpaceHopper an important step towards controlled jumping locomotion in low-gravity environments.
orphologies, a broad mass range spanning from 2 kg to 200 kg, and nominal standing heights ranging from 18 cm to 100 cm. O
mplex terrains.
ate load carrying of more than 2.0 kg at the foot. Additionally, the controller is robust to interaction forces at the foot, disturbanc
ment and subsequently fine-tune it using a physical robot. Simulated results show a tenfold sample efficiency increase compare
parameters or disturbances. The proposed controller exhibits the capability to partially eliminate the disparity between the mode
 enerated by MPCs. The effectiveness of the control approach is verified with extensive numerical simulations and experiments
generalize reference motions and effectively handle more complex locomotion tasks that may pose challenges for the simplified
analysis of the typical nonlinear wheel gaits, the difficulty of achieving the same walking speed as the linearized model is discu

is also demonstrated using a full-scale humanoid robot, Nadia, walking over both flat ground and rough terrain utilizing the gen
rbation analysis is used to explicitly reason about the tracking performance of a hybrid trajectory and is incorporated in an iLQR
ystem in real-time. To validate its effectiveness , we conducted experiments in various scenarios. The quadrupedal robot succe
ed experience accumulation requires significantly fewer offline trials to successfully reach the goal compared to regular experien
the hardware augmentation of quadruped robots to create avenues in the field of amphibious robotics and expand the use of q
eal-world deployment. Another key feature of the proposed approach is the incorporation of a boolean observation within the co
ce with the configurations of spine's shape, and found direction of spine design that have a positive impact on walking. These fi
wer reprojection errors (e.g., 4.73 vs. 8.03 pixels).
nes the RTK-GPS estimate with LiDAR point clouds through Bundle Adjustment(BA). All sequences divided in training and valid
demonstrate that our proposed SAD-Depth exhibits superiority in both model size and inference speed, achieving state-of-the-
ed to a particular neural network layer, therefore having access to high- and low-level concepts. This work shows how VDNAs

ce for the design of future VPR systems and contribute to the advancement of computer vision and robotics research.
n the Oxford Robotcar and MSLS datasets demonstrate that the proposed algorithm is not only effective in appearance change
ic Fusion (DDF) module to automatically select valuable information from different modalities for better feature fusion. Moreove
works well in practice), we systematically quantize each measurement residual into a single bit and perform maximum-a-poste
n Success Ratio and a 29.6% enhancement in trajectory accuracy against ORB-SLAM3, but also emphasize our ability to cons
forgetting when mapping incrementally. Furthermore, when a loop closure occurs, with the elastic pose graph based representa
on algorithms, we present novel evaluation metrics. Our experiments, conducted on challenging datasets including Oxford Rob
and onboard an aerial robot surveying a multi-storey mock disaster site-building. Website: https://ori.ox.ac.uk/labs/drs/nerf-map
e performance of the proposed method in real-world experiments and compare it against state-of-the-art algorithms. The propo
al and upward retrieval methods to eliminate them respectively. We validate our method on SemanticKITTI dataset and author-c
fline methods while achieving real-time performance (25Hz). The code and dataset will be available at: https://github.com/XiaoH
environments with dense and high-resolution sensor measurements demonstrate that OctoMap-RT builds maps up to 41.2 time
uced new evaluation criteria. Through a series of rigorous experiments conducted on this dataset, we substantiate the superior
wn datasets, covering a diverse range of outdoor environments and varied terrain types. We also demonstrate the compatibility
ing an FSM, which is a self-supervision module that utilizes the path traversed by the robot in pre-driving, also known as a foot
graph to the DSLP field. Experiments show that our SSL model outperforms two SOTA supervised lane graph prediction mode
30k and Pittsburgh250k. Experimental results demonstrate the superiority of our method over baseline models in terms of reco

work on environments with novel sounds and maps. Experiment results on realistic 3D scenes (Replica and Matterport3D) dem
sks are then used in the second phase, InstanceLift, to supervise the training of a neural label field, which interpolates regions m
ugh evaluation, we validated the saliency detection and localization performances to the real urban scenarios. Project page:http
 tors are being marginalized whether a robot is exposed to a BM or switching to another one, which maintains the accuracy and
n framework. We have implemented our method and tested it with the public CMU-GPR dataset. The results show that our algo
nt to effectively address the inherent scale challenges present in monocular camera reconstruction. To our knowledge, our meth
d Kalman Filter. The proposed approach improves accuracy and robustness on a public dataset. We additionally publish a new
ethod showed an extreme robustness to complete sensor occlusion (i.e., kidnapping), and enabled pinpoint sensor localization
cloud data becomes degenerate, there are momentary sensor interruptions, or the sensor moves along the map boundary or in
ng up to 91.7% recall at 100% precision in opposing viewpoint cases, while requiring less storage than all baselines tested and

g us to obtain a global probability map and achieve precise 3-degree-of-freedom (3-DoF) localization. Based on that, we develo
rimental results demonstrate that our system can predict scale-consistent trajectories from monocular videos and achieves stat
de and models are available at https://github.com/DLR-RM/UMF.
etry methods.
ucture information, the multi-branch Mixer utilizes a multi-branch structure to aggregate local and global position information, im
e datasets demonstrate that the proposed FF-LINS achieves superior accuracy and robustness than the state-of-the-art system
gistration performance by up to 20%.
of the localization system. We demonstrate the effectiveness of the proposed approach using real-world datasets, which are c
tions for absolute pose regression, resulting in more robust localization under challenging environments. We conduct extensive
etwork that aggregates local features optimized for camera pose estimation to create a global feature that enables precise ima
h epistemic uncertainty, which is used to evaluate information gain and scene coordinate quality. (3) Based on the three arts of
able to outperform previous state of the art while being faster, using 8 times smaller descriptors, having a lighter architecture an
n the retrieval ability of the past data while continually learning to recognize new places from the new data. We thoroughly evalu
sensor measurements and model simplifications through learning. The proposed methodology is evaluated in hardware using a
ons (domain adaptation). Further, an exhaustive comparison with the state-of-the-art shows improvements in performance gain
is able to autonomously build a map and localize against it in two important use cases: localization within the same colonoscop
We demonstrate position estimation with mean accuracy and precision below 1 mm, and orientation estimation with mean erro
-the-art methods. The code will be released publicly soon.
ts have been conducted to prove the effectiveness of the proposed method. Our code is available at: https://github.com/IRMVL
ment utilizing a particle filter as well as to perform scan registration by tracking the pose of a car based on matching LiDAR sca
s on large public and self-collected datasets show that our proposed framework significantly improves the performance of vario
well-known autonomous driving datasets, and the results demonstrate the accuracy and robustness in terms of vehicle localizat
nability, and eliminating handcrafted loss functions.
e methods, and improves accuracy while maintaining real-time performance, i.e., runs faster than the sensor frame rate.
fectiveness of our framework, reducing 27.4% and 15.2% of computational overhead on the 7Scenes and Cambridge Landmar
earn two 3D geometric representations (X, Y, Z coordinates) of the scene, one in camera coordinate frame and the other in glo
nding boxes. Our pose verification module evaluates the pose confidence of top-scored candidates and determines the most re
such that the local scene structure around the camera is online recovered to build the emph{Querying Point Cloud} (QPC). The
po is trained using Spike-Timing Dependent Plasticity and a supervised delta learning rule enforcing that each output spiking ne

rvised feature aggregation enables our suite of methods, AnyLoc, to achieve up to 4X significantly higher performance than exi
s at greater than 4Hz on an outdoor asphalt dataset.
of-the-art methods that use prior map information, and demonstrate its ability to perform real-time localization, at 15 Hz, on a res
alization with less than 0.2% relative error of trajectory length in four nocturnal environments.

ng only 2.5mJ per frame and executing in 38ms. In our evaluation, we globally localize in a real-world office environment, achie
otion capture (mocap) system is used to verify the results obtained by the localization algorithms. The main aim of this dataset i
e real world.

stem as ground truth, and was evaluated against a state-of-the-art method based on KFs. Experiments showed that the averag
hieves decimeter-to-sub-decimeter-level positioning accuracy using off-the-shelf sensors and decimeter-level tracking accuracy
sensor fusion with self-calibration capabilities and incorporates equivariant velocity-type measurements. We augment the filter
in a densely occupied real office environment. Our proposed solution can achieve temporally consistent localization with an ave
system integrates an uncertainty-aware neural network model for WiFi localization and a bespoken Bayesian fusion technique f
ments, we demonstrate that the proposed method enables more accurate global localization with fewer iterations compared to
and conducts weight updates and resampling iteratively in the particle update process, prioritizing the likelihood function spread
ures such as walls and doors. For that we apply filtering to remove clutter from the 3D input point cloud and then employ further
he factor graph, leading to improved early-stage hidden state inference for prediction. Our results show the trained TCN can pr
error of the mapper, but it does so at a comparable level of accuracy as an IMU would. All without the need to mount and calibr
from Doppler and IMU data alone give similar or better results than 3D point cloud registration. In our experiments, the position
on to resolve the substructure ambiguities by conducting a local-to-global consistency refinement. We demonstrate the capabili
gn the covariance recovery algorithm to improve efficiency by avoiding redundant computation. Remarkably, our approach ach

xisting navigation methods. We also perform quantitative tests on a real-world robot platform, TurtleBot. The experimental resu
ompound kernel from the first phase, we effectively address the inherent limitations of GPs in handling highdimensional problem
sence of previously unseen goals. Finally, we deploy DeformNet on an actual UR5 robotic arm to demonstrate its capability in re
ation and the real world with a quadcopter platform across various high-level tasks. We show that the proposed architecture can
ite of more dynamic merging scenarios, finding that our approach avoids the frozen robot problem which is pervasive in conven
e grid density of point clouds of real-world demonstrations and simulations in a differentiable physics simulator. Empirical valida
s is the first demonstration of a learned controller for such a task on real world hardware.
nables the design of an energy-shaping model-based tracking controller for rigid-body robots. We demonstrate dynamics learni
optimization algorithm directly via experience, specifically tailored to the needs of the control problem. We evaluate DMPO on a
, which provides an abstraction for the target dynamics. Conditioned on this, the ADM probabilistically models the system's dyn
od, together with frameworks for computing rewards and resetting the environment, high-quality controllers for a few common ro
on a physical differential-drive robot and a simulated quadrotor, showing improved prediction accuracy on OOD data relative to
RL than other approaches for learning from video data. Our system, called V-PTR, combines the benefits of pre-training on vid
wheels, allowing it to execute leg-like gait patterns. The high-dimensional actuation capabilities of the rover facilitate effective m
policy and a corresponding safety critic are pretrained on an offline dataset, in which the safety critic evaluates upper bound of s
g speed and PMMR producing human-like, high-quality solutions with a nearly 100% success rate. Source code and suppleme
diffusion model samples intermediate poses based on the representation using classifier-free guidance and then uses gradient
ignificantly from the training data. In a separate experiment, relying exclusively on visual servoing and ignoring force goals drop
um on terminal goals is designed to further accelerate the hardware-based training process. Comparative experiments and abla
t can be used to guide planning. We perform experiments in simulation on two DOO manipulation tasks to show that using the s
 ance on the noisy far point cloud which includes the whole object to propose the approximated place to manipulate. Then, we m
require perfect knowledge of the physical properties of objects (mass and coefficient of friction in particular). The improved vers

on ratio with fewer holes. Compared to two recent baselines, one implicit and one explicit, our approach requires only 10%–5
deo memory consumption. Validation experiments demonstrate our method's exceptional quality, efficiency, and adaptability ac
he process. Experimental results on several datasets with varying lighting conditions demonstrate that our method achieves sta
code is available at url{https://github.com/MichaelWu99-lab/QuadricsNet}.
monstrates superior performance in terms of precision and efficiency compared to existing methods. Furthermore, the techniqu
F-based mapping on edge computers of handheld devices and quadrotors in real-time. Experiments demonstrate that our meth
m. Moreover, the estimated shape and pose uncertainties, which we demonstrate can accurately reflect the true errors of our o
egments along with their intra-image neighbours. It further enables us to represent and update segment-level descriptors throu
the performance of SAMs for navigation, by building and deploying a behavioural navigation stack on a quadrupedal robot. Vid
c and dynamic obstacles simultaneously, an initial velocity estimation approach and a mixture model are utilized. Compared to
rms existing moving object segmentation baselines and even generalizes to different types of LiDAR sensors. We demonstrate
ts, we only consider points with the same label as neighbors. This allows us to avoid estimating normals in high curvature areas
process. Instead, we find optimal poses by directly maximizing the photometric consistency of LiDAR maps. In experiments, we
by a mobile mapping system. Compared to the performance of other methods such as Octomap, Gaussian Process Occupanc
with other state-of-the-art methods while maintaining comparable mapping accuracy and high memory efficiency. We demonst
ht back-end removal module maintaining keyframes is proposed for further DOR. As experimentally verified on SemanticKITTI,
strategies, free from data association and without making assumptions about the environment. In addition, we present the bene

swiftly and efficiently navigate through mode switches, enabled by its tunable bistability. This feature is essential for underwate
are printed in place, simplifying the manufacturing process.
or passive-mode. In passive-mode, the ROV body can pitch similar to how a dynamically-coupled underactuated UVMS withou
a 10-day field test, the vehicle showed good aerial performance and underwater control effect. Several full mission cycle tests p
s per second) and performing complexmaneuvers, such as a figure-8 and S-shape trajectory. Instantaneous yaw changes of 15
passive joint in the leg. More specifically, its swimming speed is fast as 0.34 m/s at the frequency of 1.4 Hz, showing a 30.8% in
zed, extracting information about the quantity and composition of MPs, in order to provide an overview of the complete procedu
everal previous walking cycles' parameterized biases to predict the current walking cycle's biases to generate the needed pitch
safety within a planning horizon and minimize the expected distance to the target position. Furthermore, path safety is consider
method achieves similar depth prediction accuracy on a downward looking dataset we collected with a diver operated camera r
ons for object detection and pose estimation. When benchmarking against similar end-to-end methodologies for 6D object pose
m varying viewpoints. This inherent variability presents formidable challenges in data association, particularly for feature-based m
vels. We further introduce a dataset of perturbed underwater images to benchmark uncertainty quantification in this domain. Th
g efficiency and numerical stability during training. The proposed technique integrates underwater light effects into a volume ren
te-of-the-art performance. Finally, we explore the design choices and implications of semantic segmentation for visual servoing

s. Experimental results show that our method outperforms the previous state-of-the-art works, achieving up to 8.2% Bpp saving

proposed hybrid approach may offer a simple and practical method for modeling UV behavior when submerged or interacting w
 by 59.2% and 26.1% compared with random take-off cases, and the locomotion is smooth, graceful, and low-risk. Compared w
acoustic observations. In this paper we present ReefGlider, a novel AUV that uses only buoyancy for control but is still highly m
face vessel in simulation and natural waters, both with external disturbances. Specifically, compared with the traditional MPC m
the stability of the robot movement process, and optimized the difference from 25% to 3%. This paper provides sufficient techn
es it applicable to various lenses, as validated with fisheye (FOV >120 deg) and wide-FOV (FOV ~120 deg). We validated our m
wledge, this is the first work to benchmark both learning-based and classical registration methods on an AUV-based MBES dat
m and waypoint tracking with an average final error of 0.398 m.
behavior is robust even with time-varying network topology and agent dropouts.

WECs for dock charging and generates efficient mission plans for AUVs by optimizing their time to visit the dock based on the
hich show that the proposed policy achieves superior performance in navigation safety, while requiring minimal travel time and e
map in the planning phase, and performing belief propagation at each step is avoided. A receding horizon planning strategy is im
our knowledge, we are the first to introduce a new type of AMV that can work in displacement mode, foilborne mode, and subm
ss the state-of-the-art in classification accuracy and efficiency. This framework is a promising approach for more efficient auton
posed underwater manipulator was created, and a flowing-water experimental platform was constructed to test its trajectory-trac
results. To this end, a combination of two different types of loop-closures is presented. Namely a consecutive loop closure, and

ble water clearance for distances beyond one kilometer. During harbor maneuvering 90% of the relative water clearance errors
etter than static color thresholding methods as background color varies. DS detection is possible at a range of 5 meters in clear
sing a small cruise boat, modified as an autonomous surface vehicle.
computer vision tasks across refractive media.
er domain, especially in underwater caves, as the narrow passage and turbulent flow make it difficult to perform loop closures a
enes show that our proposed approach outperforms state-of-the-art visual-inertial SLAM.

re solving only n+1 trajectory optimization problems for single rigid body dynamics sequentially, thereby rendering the problem
oach in detecting internal cavities within wooden structures.
g dualquaternions.
algorithm, which is inspired by the extremum-seeking control (ESC) technique. Experimental results are presented to validate th
T sensor, our approach offers two solutions: substitution with tactile sensing, achieving comparable flight performance, or integ
driven by two independent servos. Three carbon fiber springs are employed to build a lightweight jumping module with high ela
nsure robust perching in the presence of measurement and control errors and avoid collisions with the perching site immediate
cludes a novel general model accounting for the transient hybrid dynamics for aerial systems with cable-suspended load to clos

orcraft for rescuing in mountainous and plateaus, we break through the maximum-descending-speed of a coaxial rotors UAV. H
results through an extensive set of numerical tests in a realistic simulation environment including underactuated aerial vehicles
s. Building on this, we devised and evaluated multiple perching and disentangling strategies. Comparisons of perching and dise
ce, a learning-based prediction approach is leveraged to rapidly forecast the UAV platform motion online. Subsequently, an opt
ary-wing MAV through onboard flight data and observe the close correlation between the motion of the MAV, the measured ext
mulation and their subsequent observation in flight tests.
wn, and fountain lift), each occurring with a different magnitude at a different spacing from the ground plane, are investigated a
paths, each comprising of at most five segments. Unlike in the classical Markov-Dubins problem, the CCC path, which is a cand
ncrease the tolerance to attitude errors. The wheel mechanism also eliminates the requirement of matching the attitude and tan
acking trajectories. Lastly, the flight experiments extend to both sides of the wing to validate this method as being applicable fo
rategy which selectively deactivates angular motion control and enables direct force control in specific directions. In particular, w
or. These experiments not only demonstrate the feasibility of our platform but also provide new perspectives for future aerial pla

eal-world flight. E2E showcases a significant 1.39-second advantage in simulation and a 0.17-second edge in real-world testing

co-designed morphing drones outperform fixed-winged drones in terms of energy efficiency and mission time, suggesting that t
urbances surpass the preset range and the desired control is beyond the ultimate output of the quadrotor, the quadrotor compl
r perspectives. To test and demonstrate the functionality and performance of the proposed method, we present data collected

shing across variable surface orientations. Furthermore, the assessment results are used to plan and execute physical experim
emonstration mimicking the installation of antennas or cable-car masts, we show the capability of Geranos to assemble poles (m
performing aerial manipulation tasks, this paper presents a controller based on deep neural networks that classifies and learns
n scheme to solve the problem. On the hardware side, we integrate our algorithms with collision avoidance and drone controller
rations of assemblies, capable of tolerating rotor failures and balancing the energy consumption of each module. The functiona

ctories spanning a broad range of evaluation metrics, around 90% of the range occupied by popular algorithms. The proposed
haracteristics. Finally, a composite disturbance observer-based torque control algorithm was developed, which capitalizes on th
θ < π) clutch designs. We also report the first unfilled polymeric material, polybenzimidazole (PBI), to exhibit the JR-effect
for the research on flapping wing.
bristle-bot’s body was used to ensure that each side of the bristle-bot's body has a different frequency response. This asy
x configurations; these show hysteretic energy losses no larger than 35% over the stiffness variation and confirm stiffness to sc

state-of- the-art literature. The differences between current and torque bandwidth, two concepts often conflated in literature, are
5% and the maximum motor current by 33.56% of the driving element.

ucted. In this section, the gripper shows 32.912% improved performance in representative tasks. Finally, the practical aspects o

roposed robot, an object-grasping experiment was conducted to verify the feasibility.

orm versatile locomotion to accomplish object transportation in diverse environments by utilizing its foreleg-wheel mechanisms.
ound for dynamic movements, such as jumping. The implemented leg and toe mechanisms weigh 8.16 kg, and the height from
test site equipped with motion capture system to find the best strategy to control the sails and rudder, in order to increase towin
the number of pointcloud and a decrease in the intensity of the sensor data after contamination, it was validated that the numb

software. The results show that the robot has great adaptability in complex environments.
e actuator. Furthermore, an application to underactuated-type locomotor is demonstrated.
nd its surroundings as well. Further, a kinematics model is established and verified to precisely control its lengthening and steer
and the capabilities of our transforming multi-modal robot.
ontraction force and has been successfully utilized in a bio-syncretic robot to demonstrate its actuation capability. This work ma
 study we present the architecture of the framework, presenting and discussing some technical choices made in the design.
for catheter steering of a soft tissue-imitating phantom. Results show that an average steering error of 0.56 mm and signal-to-n
on model. Then, an approximate model-based optimal control scheme is developed to push the object to track a time-varying tr
he usage and effectiveness of the robotized dissector.
sing patient trauma. Precise and rapid localization of the trocars enables the automation of the insertion procedure and dynami
able-bodied kinematic and kinetic references. As designed, PVI-HVC results were similar to those of PVIC when the user used
n that integrates uncertainty information of the data provided to the robot. We design a focus and context visualization that rend
xane (PDMS) and incorporate micro-patterns on their outer surfaces. We have proposed a straightforward strategy for detecting
earch and development architecture, and were found to be common among dVRK user feedback, regardless of the breadth an
o minimise potential risks during human-robot interaction. The proposed method's efficacy is demonstrated through experiments

mplicit neural representations. Our approach employs a robotic ultrasound system for automated screenings. To demonstrate t
eball phantom. Both simulator and phantom experiments demonstrate an average needle-tip estimation error within [1.0, 2.0] m
acterize the rib cartilage outline to match the source and target point cloud, the path extracted from the tree graph is optimized b
stimated so that the probe avoids the connection between the mitral valve and the papillary muscle and chordae. To validate th
le Doppler and B-mode images to ensure anatomical spatial correctness. The second encoder operates on the pure Doppler im
mm. The proposed US 3D reconstruction-based registration method is non-invasive and radiation-free, and shows the promisin
ated recurrent unit (M1-DCGRU), a fast and accurate identification of seven drilling stages was realized. The final force sensing
ce control method for a posterior segment ophthalmic surgical robot. This method aims to achieve compliance control of the su
d control algorithm, the force error percentage can be maintained below 5.0% within 3 s and the maximal root mean square err
troller is integrated with a handrest and the controller's DOFs are strategically allocated. In this work, as the first step of addres
ess, we conducted initial feasibility experiments involving a medical student and two experienced surgeons. The experiment fo
Both simulated and real-world experiments are conducted to demonstrate the effectiveness of our disinfection robot system, esp
uman subjects were invited to perform an emulated lymphadenectomy based on the da Vinci robot in a dry lab. Experimental re
le, into the robot's dynamic model, which complies with remote center of motion constraints. The interaction wrench's gain is ad
The accuracy of the proposed model is compared with that of the kinematic model based on constant curvature assumption. T
surgical operation. On the other hand, the interventionist controls the grasping state of the instruments remotely to control the s
manipulator with RCM, ORIGANOID (dimensions: W60 x D120 x H30 mm, weight: 12.6 g). The robotic arm features flexure hin
cal robots or biomedical imaging apparatus requiring coaxial displacement amplification in confined spaces.
hieve complex shape transformations. Results show that the range of motion in the tip direction of the RMCR increases by 193
istalsis. A series of simulations and experiments are conducted to evaluate the feasibility of the developed robot. In experiment

e unreliable when the input sample is outside the training distribution or corrupted by noise. Recognizing such deviations is cruc
6 times faster), generate always valid joint configurations, are more interpretable, and are easier to extend.
nes. We verified the performance of the robot travelling through the entire large intestine with the maximum speeds of 54 mm/s
and a controller that keeps the swimmer within the ultrasound imaging slice. The position measurement and closed-loop contro
o junior surgeons. The user study of this system indicated a significant reduction of the number of clutching actions and mean ta
as a base-learner showed tendency to yield greater than 90% accuracy. The MLPs outperformed the SEs in accuracies and in
ention mechanism, temporal tracking networks, and introduced a novel 3D segmentation head that performs 3D deconvolution a

sing neural network-based approaches. The outcomes of our experiments exhibit promising prospects for real-time motion mon
st time, we implemented a teleoperation control mode incorporated with an adaptive sclera force control algorithm using a PHA
ptive motion-scaling algorithm. Detailed user studies indicate that a suitable motion-scaling ratio can be obtained and adjusted
ce of an external magnetic field, and possesses the capability for film expansion and disintegration under high-frequency electr
2 minutes on average.
g both translational and rotational movements. We propose an openloop manual control sequence that allows the formation of a
eet the surgical requirements is made and integrated into the ESR manipulator. This manipulator can measure the catheter for
s under 10 mm. We introduce and study three learning models by incorporating spatial strain encoders, and compare their perf
udy, and marked using color inks and near-infrared fluorescent (NIRF) markers to indicate the pseudotumor. ASTR monitors th
success or failure of the performed experiments.
ared it to manual control. The results showed that robotic control is superior to manual control in simple surgical tasks and the c
bility area. The outcomes of this method demonstrate the promising potential to enhance the robot's accessibility, as evidenced
and interviews and showed positive results. Compared to conventional, manual guidewire navigation, the magnetic navigation s

owing task shows that the position error was maximally 3mm in open-looped control. Test on a skull model demonstrates that th
entry counts. For dynamic environment navigation, dynamic time warping (DTW), a technique for measuring the similarity betw
processing techniques are also introduced and validated. Experiments are also performed to characterise and validate the mod

n less than 1mm in-plane errors for insertions reaching depths of up to 61 mm, demonstrating the validity and generalizability of
ore, physical Human-Robot Interaction (pHRI) admittance control experiments were conducted in a surgical context to evaluate
onstrate that this system can efficiently fabricate microrobots with excellent mechanical properties, which have application poten

propulsion. Buoyancy and magnetic forces actuation enables flexible 3D locomotion in fluidic environments. Experimental result

ned and fabricated using an insulating polymer layer, two conductive layers incorporating silver nanoparticles with anodic/catho

echanisms.
ienced experimentalists, supporting an important role for robotic microsurgery in broadening the use of such techniques in expe
ordinates of the eye gaze, and then a neural network with gaze-guided attention is trained to segment the visual Region of Inte
ive throws with micro-suture on mini-incision, like in vascular anastomosis or microsurgery. We successfully evaluated the knot
p. A two-stage data augmentation method is proposed to augment the sparse depth data, and the depth map is optimized by a
cessfully completed comprehensive scanning of the optoelectronic tweezers chip, achieving a 60% reduction in autofocus time
nt which shows proposed framework is robust and precise enough to obtain an intact single cell. The error of autonomous singl
al results validated its high precision and bandwidth, exhibiting its potential to perform intricate micromanipulation tasks.
ious work has been solved. The new robot prototype S2worm-G weighs 4.71 g, scales 4.0 cm, achieves a top speed of 75.0 cm
straints. The attainable force and torque magnitudes are on par with the requirements of surgeries at human head (e.g., brain,
d the cross-scale guidance (CSG) network for accurate and efficient segmentation of minute sperm objects. The CSG network
cise repositioning of the microscope. Experimental results demonstrate the precision and stability of the auto-calibration method
nstrated the automatic capillary insertion for TEVC experiment, at which the low insertion depth was preferable. The results indi
ography without significant artifacts. The method is tested using an AFM-in- SEM system based on a 3-DOF cartesian robot equ
g strategy is established to avoid the risk of touching the sperm head where DNA is contained. A visual servo controller equipp
fix distorted features, by refining preliminary segmented masks through merging features extracted by feature pyramid networ
 -vivo settings. Results show that our proposed method can achieve faster scanning speed than traditional raster scanning and
ansportation of the mouse oocytes, and measured rotational and translational speed with a range of applied voltage. The metho
olled rotating magnetic fields, guided by real-time Xray Fluoroscopy images. Incorporating a two-degree-of-freedom control sys
rates of 96.76%, 94.23%, and 92.39% on cluttered level 1 objects, cluttered level 2 objects, and cluttered mixed objects, respe
a feasible plan; finally, it attempts to locate and observe the target, replanning after each newly observed object. By combining
are applied to ensure that the trajectory can fulfill specific task requirements. Simulation and real-world experiments demonstra
trate that, with accurate perception, the decision making of HIMOS effectively transfers to new environments in a zero-shot ma
cts, stacked objects, and various obstacles show that our approach can handle a variety of conditions that have not been previo
ased on proprioceptive measurements and perform tasks with multiple unknown environmental parameters. The approach is ev
agent to prioritize actions that yield detailed observations while approaching the target object for grasping. Through extensive ex
nts on physical robot to compare the differences between decoupled control and independent control, and the results show tha
bility. We show the efficacy of our method in simulated experiments with a dual-arm TIAGo++ MoMa robot performing mobile g
a) uncertainty of the learned dynamics. This controller actively chooses the optimal control actions that (i) optimize the control p
entirely in the real world. Our experiments show the proposed system is able to successfully capture the dynamics of clay, and
By combining the NN with the first principles model, the LMPC is born, capable of accurately predicting the future with a compu
, and high levels of uncertainty derived from the model. We demonstrate the effectiveness of our approach through experiment
y accommodate varying terrains and iteratively modified behaviors within a behavior optimization framework. This provides both
experiment dataset and compared to the existing batch optimisation method.
own rigid body dynamics terms. A composite model consisting of a rigid body model and an online data-driven residual is develo
ludes a discretized physical model derived from first principles together with slack variables that allow for a small mismatch bet
ictive control, leveraging model uncertainty to make safe decisions. We validate our work on a benchmark real-world multi-terra

milar to the human-administered massage, decreasing participants’ mean systolic and diastolic blood pressure by 2.8 mmHg
seconds. Interestingly, as our ESP features can be described in human-readable language naturally, the application of integrati
cluding RGB videos and depth maps from four synchronized views, 3D body poses, normal maps, segmentation masks and bo
the grasping effect on the object. The robot grasping strategy diagram is constructed as follows: First, grasp quality is predicte
.t. ergonomics and performance.
ata. The musculoskeletal model, along with its control algorithm, will be made available to the research community to promote a
n body, we construct a keypoints-based 3D human template, with which the 2D low-resolution features can be mapped to 3D s
dom actuated by 31 muscles. The MHE relies on a series of overlapping trajectory optimization subproblems of which the follow
ted. The results show that the proposed model is capable of interpreting and predicting the dynamical behaviors in LJ-based co
estigated in future work.

n the PCC model considering gravity and external forces. Furthermore, we devise an inverse kinematic controller that is capabl
etric parameterization of constant curvature segment geometry for CRs with extensible segments. The resulting IK formulation i
er accuracy and reliability than state-of-the-art baselines both in- and out-of-distribution of the training data.
ness of the method when key points are partially occluded. The proposed method is also integrated into a simulation in Unity fo
BF NN model. Based on this, a novel knowledge-based controller is further proposed to provide desirable control performance f
lving multiple systems, trajectories, and in the case of external disturbances and model mismatches.
e active tremor suppression effectiveness of the proposed method was tested on a bench-top tremor simulator, and such metho
me capability is achieved with our implementation. The state estimation is tested in a Cosserat rod simulation and on the physica
 of the soft arm passive, and complete, with both modules active). This performance raises to 85% if one can choose the numbe
of the real-world 3D soft robotic arm Helix.
and analyzed on two classical designs of soft robots doing manipulation and locomotion tasks.
cture. Experimental results demonstrate the superiority of the mechanics-based modelling and control approach, showing an a
uate mesh reconstruction on simulated data and the POE-M pipeline with real-world experiments. Ablation studies suggest PO
ng robot model. We further apply our approach to a stimulus-responsive hydrogel simulator that captures the complexity of chem
evelop a simple hardware platform to evaluate both the models and the controller. Finally, using the hardware, we demonstrate

etween the upper and the lower body with respect to walking speed. The obtained results from this study can be used in a tool

tree (RRT) by representing the floating-base dynamics as pfaffian constraints on joint velocity. These constraints guide the ma
he high-dimensional system shows better efficiency compared to that of the reduced-order model. Furthermore, we provide a ge
ork was evaluated in multiple irregular and heavily cluttered simulated environments randomly generated and in real-world trials
MPC), the learned controller predicts a trajectory within a planning horizon to ensure the satisfaction of the STL requirement in d
defined sequence of waypoints and is computationally efficient. Performance comparison is provided in simulation that shows la
pensive. The majority of existing methods construct only a sparse roadmap which, however, struggles to find all distinct paths i
s preserving the fast convergence property of MPPI. Our simulation and real-world experimental results demonstrate that SVG-
al application of our proposed algorithm by posing the visibility quantity as a heuristic and implementing a deterministic, local-m
elaxing the problem constraints and reducing the problem size. Both techniques are integrated into our path planning algorithm

s canonical ergodic search to ergodic search under battery constraints. Empirical data from simulations and real-world experime
reconstruction process in RBI-RRT*, the algorithm continues to find the near-optimal path. A series of simulations and real-wor
tuation and base internal wrench. We then propose to perform path planning on this manifold using AtlasRRT*, a sampling-bas

Neural Connect to build connectivity of the guidance state set and further boost performance in challenging planning problems
uating on a low-power single-board computer. The planning speeds demonstrated are fast enough to reside in the range of con

mplexity of O(T^ 2.4 ) +O(T n^2 ), a significant improvement over the naive application of DDP, which leads to a complexity of O

nning efficiency in numerous areas of robotics, including mapping and search tasks using uncrewed aerial systems, environme
e solution. To this end, we propose an anytime refinement approach that works in combination with CTMP algorithms. Our pro
nstrate our method's capabilities on a Spot robot in complex real-world outdoor scenes, including dense vegetation. We observ
ematically valid trajectories. Like classical planning, for any new scene at the time of inference, we compute scene-specific cost

dition, we describe a method that can improve upon solutions found during the formative stages of the main algorithm, using a
put forward to improve the path-shortening performance and avoid unnecessary heading changes.Second, an iterative path per
Infeasibility proofs in 5D complete within minutes.

mming algorithm enables the planner to generate a path visiting more than ten targets within seconds. Moreover, the proposed
mputing power hardware. The experiments yielded promising results with a 87% success rate.
 effectiveness of the proposed methodology, showing that the lap time can be further reduced by taking into account the gate's
sly. We perform experiments in simulations and a real platform. We find, somewhat conveniently, that geometric front-ends are
trained environments, and tracking aggressive flight trajectories accurately, even under significant uncertainty. We plan to relea
optimization method is proposed, which can jointly optimize the hitting terminal states and time cost, considering dynamic feas
opening of a gripper, located at the lower edge of the cable. The performance of the proposed control scheme is experimental
g horizon and fully exploiting the utilized vehicles' motion capabilities. Based on the evaluation results, the approach indicates im
Successful outdoor experiments with five novice subjects are presented that demonstrate the feasibility and potential applicatio
bility of this pioneering concept is still in its very early stage of development, which yet needs to be experimentally verified. Tow
ask of the DMJ. In the presence of unknown backlash and disturbance, the proposed strategy can improve both transient and s

lidate the proposed uniform passive FTC method, outdoor experiments are performed for the first time, which have demonstrat
olved Acceleration Control (RAC), are translated into a feasible joint configuration. The validity of the proposed approach is dem
proposed JetRacer-4WS can be used as a low-cost platform for both research and education on motion control, path following,
ith the bilinear matrix inequality constraints in the nonlinear problem, we develop a parabolic relaxation and a sequential algorith
stitute a stepping stone towards a quantitative metric for robots adapting their behaviour to human motor fluctuation.
y of ILC in mitigating tracking errors, even in scenarios where the dynamic model of the D-CDPRs is unknown.
indicate that the proposed FFC architecture enables the manipulator not only to interact smoothly with external forces, but also
on cone condition at the end of each stance leg are generated from the low-level controller. The effectiveness of the approach i
duced-order and full-body models to ensure safety, integrating safety-critical control and footstep re-planning approaches. We v
d method are employed to address noises effects, and related theoretical analyses are given. To validate the proposed ARC$^
t-in auto-tuning routine.
step phases to generate new gaits of spatial locomotion. We evaluate the scheme in real-world experiments, including steering
or with seven degrees of freedom (DoF). Experimental results verify the effectiveness and improved performance of the propos
ome such a burden, in this study, the design procedure of a model-free control and synchronization technique of piezocermam

elaxes the requirement for precise tracking of hand position and enables precise robot assistance guided by the velocity of hum
ies and confirms the theoretical claims of the approach.
s successfully validated through simulations and experiments with a collaborative manipulator in a surface polishing task.
ce to reduce the impact of cable elasticity deformation and friction on joint motion accuracy. Finally, servo control is applied at t
em to learn a reduced-order model (ROM) within a model predictive control (MPC). Results show a 49% improvement in viable
ximated by a quadratic programming problem. Further, the estimation is updated at each iteration until the convergence is reac
on model does not require raw shape information, making the model's dimensions smaller and easier to learn than a naive sha
on hardware to solve contact-rich manipulation tasks, where a robot uses its end-effector to roll different unknown objects on a t
ol converges with 46 s of training data, almost four times less than the shortest human dataset gathering protocol. We show tha
rol problems, allowing a global view. A condition for the existence of orthogonal foliations is derived. If the condition is not satis
ment in final action cost.
acteristics and warrants further exploration.

city and acceleration limitations. Finally, simulation and practical experiments demonstrate the feasibility of the proposed metho

rajectory and utilizes a dynamic programming algorithm to find the optimal motion by linking the IK solutions. We demonstrate t
data. Our results show that the proposed approach efficiently solves various CMP problems in both simulation and real-world, in

ain a drastic increase in failure rates and collision rates - leading the planner to fail 95% of the time and collide 90% of the time
optimization using clarity, an information measure bounded between [0,1]. The clarity dynamics allow us to capture information
third, we present a rigorous analysis supporting the consistency of the algorithm. In the Supplementary Material we provide the
T) problem of a tethered robot into a multi-objective optimisation framework. Explicit consideration of the constraint of the relati
artially observed environmental information. Compared to manually crafted CBF which suffers from over-approximating robot g
omputed subsequently based on the obtained base trajectory. To validate our method, we propose a discrete optimal base traj
uning the model. We also tested our method on a physical Franka Panda robot with real-world sensor data, demonstrating the g
robots: a quadruped, a wheeled quadruped, and a legged excavator. We validate our hybrid planner in the real world and simu
into the process of travel point selection. Finally, the reference path is gradually generated and optimized, so that the UGV can

waiter robot.
e goal of this project is to verify the effect of non-verbal cues in group persuasion by a robot and comparing it to single person p
clusion problem that arises in a complex environment, we develop an onboard multi-sensor perception system for the bimanual
asks with the received information. Crucially, LIM2N creates smooth cooperation among the reasoning of multimodal input, mult
he dual-modal e-skin is capable of transmitting tactile information bidirectionally, enabling object recognition and fine-weighing o

pyramid features and global vertex features from multi-scale feature maps of multi-view images, enabling the exploration of rich
re 2D skeleton+RGB fusion for action segmentation.
gnition is significantly more accurate than monomodal analyses. Furthermore, our method continuously monitors interactions th
efully selected 150 challenging scenes as well as real-robot platforms. Results show that our method demonstrates superior gen
n leverage user feedback exhibits significant improvement. However, obtaining a good balance between task completion and th
ules for visual grounding, question asking, object grasping, and most importantly, answer interpretation for pragmatic inference.
w that MM-CNN outperforms traditional EIT algorithms like NOSER and TV, reducing location dependency and improving shape
areness, consisting of pairs of high-level commands, scene descriptions, and labels of command type (i.e., clear, ambiguous, o
ts of user study with 40 subjects show the significant effectiveness of our system in assisting group discussion.
rison of the presented device with previous adaptable interfaces for the exoskeleton: the user command interface and the mon
ated. The effectiveness of our tracking method is validated through extensive experiments.
g a collected dataset, resulting in a notable 28.7$%$ reduction in drift compared to the baseline.

ation solely through images. We also present a real-time dense reconstruction framework for multi-UAV with autonomous visua
luations of our system using both public and our custom datasets showcase notable improvements in the accuracy of collabora

atasets and is compatible with various LiDAR manufacturers and scanning patterns. We open-source the code for public acces
data associations, and it uses incremental merging to assemble large maps from individual trajectory segments given in random
eld experiments in indoor and outdoor environments, involving push disturbances, varying and unknown payloads, uneven terra
aph attention networks and transformers. By relaxing the constraint of one-to-one correspondence in bipartite matching, our ap
y, of which their future control decisions are all unknown. It includes: (i) the assignment of monitoring tasks, modeled as a flexib
 actually live on borrowed time as it is on a transient to breakdown. We provide two modeling options (based on queueing theo

se, where performance does not degrade. Furthermore, we built an aerial melee scenario agent.
proach also helps minimise the sim-to-real application gap.
ates across all tasks in RoCoBench and adapts to variations in task semantics. Our dialog setup offers high interpretability and
n of this mechanism on policies for swarm behavior deployed on severely compute-constrained hardware. Our work is the first w
, and takes advantage of meta learning mechanism to conclude the general meta policy. When some robots crash and get repl
ction space that grows linearly with the number of agents rather than exponentially. In this paper, we outline the algorithm, sket

eras to efficiently comprehend and imagine their surrounding environment through cooperative communication. Experimental re
verage mission, analyzing how the number of deployed robots affects performances and making a comparison with the ideal ca
ng. Our simulations show that using reputation to weigh the size of the Voronoi cells outperforms the case where we neglect rep
munication. Via the use of robust factors in GBP, our method is tolerant to a high percentage of faulty sensor measurements or
ve information gathered from both expansions results in more effective and accurate predictions. We develop an expert algorith
ication range of time offset estimation, we further propose an iterative method to recover the time offset from coarse to fine. Co
evaluating the performance of our approach as the number of robots increase and communication ranges vary. Finally, we dem
h a greedy approach our method reduces rendezvous overhead on average by 11.33%. We demonstrate proof-of-concept with
cuss the dis/advantages of the different designs we tested. We also introduce a strategy for genderless passive docking of cell

ystem leverages utility values and pre-conditions to ensure that the allocation improves the overall mission execution quality wh

viability and relevance of our algorithms.

he number of hypergraph vertices scales linearly with the number of either robots or objects, while the number of hyperarcs sc
namics and sensor blinding due to sun position. Results revealed that the iSAM2-based method is capable of handling sensor b
eedy algorithm and a depth-first search algorithm. To evaluate these algorithms, we test their performance across simulated run
ed point clouds, and the bandwidth transmission advantages, derived from sharing only the latent vectors. Our proposed appro
their routes overlap. The paper provides simulations of various representative situations, showing that the approach of runtime
ach robot to improve the exploration and surface inspection capabilities. The experimental results demonstrate our method can
, the proposed method is validated in real experiments.
improved plan convergence and coordination performance over standard Dec-MCTS enabled by the intent representation’

a sequential monte carlo implementation of the Probability Hypothesis Density filter for posterior inference combined with Thom

rce code, represents a first step to formalize and analyze a complex problem that has not been solved in the state of the art.
t of tasks to the set of robots. The original capsule attention network architecture is fundamentally modified by adding encoding
blasting and excavation, transportation of ore and waste, crush and screen, stockpile and load-out, rail network distribution, and
pproach based on time-expanded graphs. Moreover, the presented control architecture aims at grounding theoretical methods

etting factor as a measure of the cumulative acquired information, and (ii) the use of High Order Control Barrier Functions (HOC
wn that arbitrarily large formations may be designed for which all singularities lie within a known set of geometric conditions. An
relationship between each robot's patrolling strategy and the individual steady state distribution. Then, we derive a multi-agent
 ed strategy and validate its effectiveness through numerical simulations involving a high density of unicycles.
. We use our fitness measure definition to asses the optimized parameters through 100 randomized simulations that vary surfa
anned for independently, but the computational challenge remains. We present a hierarchical approach that first 1) uses A* to d
mulated and a real-world environment. Experiment results show that the proposed FFCGS achieves a significantly improved pe
elemanipulation tasks. The designs, the electronics, and the code are open-sourced to allow replication by others.
ff-the-shelf sensors (air-pressure and capacitance). The proposed robot hand could effectively sense interaction forces and pro
ation strategy to enhance the target pose for more feasible trajectories. Extensive validations on two public datasets show that
lementation of dexterous operations. This HIRO Hand empowers the operator to utilize their own tactile feedback to determine
d that our proposed hand is useful for interaction with humans including receiving and handing over things.

of expert data, which should enable the training of a proficient BC agent. However, the expert data is not labelled in the datase

on one. Therein, slack variables are introduced to fulfill the assignment and facilitate the searching of a balanced solution betw
y while optimizing the performance of the desired task. Numerical simulations validate its effectiveness.
ally asymptotically stable controllers and describe simulation experiments that demonstrate the effectiveness of these controlle
telligent collaborative behaviour rather than just guiding robots to individual, selfish goals. We encourage the reader to view ou
autonomously managed by the 5G-enabled UAV. This problem is addressed in a novel feedback loop fashion utilizing a proba
ns executed by the proposed robot arm design. A local network API throughput and latency analysis is also provided. CloudGrip
ions to balance cost and latency.
awareness setting and enabling foraging robots to effectively coordinate their actions during multi-robot exploration.
substantial enhancement in tracking performance due to the proposed modifications, particularly in scenarios with high commun
tic and real evaluation settings, we show that our method can improve the angular and translation error accuracy by 80.1% and
se dataset and synthesized dataset built under Carla simulation. Code will be released at https://github.com/PJLab-ADG/VeloV
n accuracy. In detail, the proposed method achieves an satisfactory speed of ˜16.7 ms per frame in experiments on the pub
d by using both the absolute and relative depth optimization loss to promote the precision of depth prediction, while the capabili
ting on style-related information of high-level features using an attention module. Using self-supervised contrastive learning, ou
ms the state-of-the-art in both single-view and multi-view 6D pose estimation. We furthermore show the effectiveness of our sym
ur simulation results show an inverse correlation between inherent predictive uncertainty and NMI throughout the model's trainin
ucted on both synthetic and real datasets, demonstrating the superior performance of our method over previous state-of-the-ar
nd 11.4% on IoU50 on the real-scene REAL275 dataset) over the state-of-the-art and real-time inference (42 FPS).
cting unknown objects by incorporating the object-detection fashions into the pixel-wise anomaly detection methods. To achieve
, outperforming the prior art of SWFormer by +1.7 mAPH L2.
allows to significantly improve the performance of the event-based student object detector, nearly halving the performance gap
for localized feature aggregation within bounding box proposals. Both attention operations are model-agnostic network module
ance of various 3D detectors by about 2.0 vehicle level 2 APH with negligible latency and slightly enhances the performance of
ble to other representations. To achieve this, we introduce fine-grained pillar feature encoding (FG-PFE), which utilizes spatio-te
mes and non-keyframes of compressed video. Through comprehensive experiments, we show that our model can achieve the s
on-local regions and thus enlarges the receptive field of clusters. We conduct extensive experiments on the KITTI, Waymo, and
f-the-art performance in an efficient CNN-based paradigm, attaining a balance between accuracy and efficiency.
te-of-the-art (SOTA) methods for both radar and LiDAR object detection while maintaining competitive efficiency in runtime.
for low quality pseudo-labels early in self-training, we propose an approach to dynamically adjust the number of pseudo-labels
 , we further introduce (1) cross-attention within the associated subsets to transfer image features into the BEV, and (2)a multi-s
s fused into branch (ii) to transfer predictive knowledge. We conduct extensive experiments with the large-scale nuScenes data
red with previous works but also emphasizing the important information in both channel and spatial domain. Finally, it is proved
d finer structures. Furthermore, we propose a new pipeline to integrate CLIPUNetr into UIBVS and apply it to control robots in r
We assess the effectiveness of our approach through a series of comprehensive experiments conducted on three challenging
ynchronous LiDAR traversals of the same location for monocular 3D detectors. We show consistent and significant performanc
ur resulting approach can reliably detect OOD instances in a scene, as shown by state-of-the-art performance on OOD detectio
cing motion to regions of segmentation uncertainty, we are able to drastically improve segmentation accuracy in an uncertainty-
n cause potential collisions due to frame misalignment. Addressing this, we introduce a novel frame alignment pipeline that recti
objects. Notably, our method achieves the improvement of 7.96% mIOU score compared to the baselines. Furthermore, it offer
ainties. However, applying EDL to 3D object detection presents three key challenges: (1) lower pseudo label quality in comparis
e that HandyPriors attains comparable or superior results in the pose estimation task, and that the differentiable physics modul

quently, based on ensemble learning, our method performs feature-level fusion on multiple networks to exploit their different ex
al-world data sets, achieving 86% IoU at 17 FPS in typical robotic settings. The method outperforms a recent appearance-base
ing OD models on the diverse data in our VIP500 dataset, revealing the promise of OD models trained on VI-SLAM data; we ex
int features in 3D. We evaluated the proposed method on large-scale autonomous driving datasets i.e. SemanticKITTI and nuS
nce on the moving instance tracking benchmark of the RadarScenes dataset compared to the current state of the art.
quivariant object detection network framework that is equivariant with respect to multiple levels of symmetry groups based on m

ance without affecting the inference speed. Extensive experiments on Cityscapes-DPS and SemKITTI-DPS show that EfficientD
mation loss with temporal association excavation of historical features. Based on the predicted features generated by the HFP m
scene representations in robotics, and it outperforms the baseline method in terms of grouping quality. Our project website is a
k dataset at 20 fps and provide comprehensive insights through detailed analysis of our tracking model.
ss on the target domain. Extensive experiments demonstrate the effectiveness, robustness, and detector-agnosticism of our WL

ublic nuScenes dataset demonstrate that our proposed RCM-Fusion achieves state-of-the-art performances among single fram
semi-supervised learning method for post-processing, referred to as OVAAL+. Our results demonstrate that both OVAAL and O

o location changes or ego-vehicle motion. Next we solve a mixed-integer linear program that provides a general solution to the
size-balanced dataset that is also more class-balanced. Furthermore, in contrast to existing information measures like entropy o
Time-Varying Model Predictive Control (LTV-MPC) is utilized in the UGV controller for dynamic tracking. Experimental results d
blem for the 4WS vehicle. Practical constraints, such as limits on wheel velocities, steering angles and their rate-of-changes ar
nvergence. Lastly, we evaluate the performance of trajectories generated with and without jerk limits in terms of peak power, to
blems reveal that as compared to the best alternative parallel preconditioner from the literature, our symmetric stair precondition
s. This is driven by a custom PCG solver which outperforms state-of-the-art, CPU-based, linear system solvers by at least 10x f
adjusted on-the-fly, therefore enhancing the tracking accuracy and robustness. We evaluate our approach through simulations
ng process for the optimal control problems is shown to be fast and intuitive. Similar to motions in free space, the proposed inv
ends prior works, which find only solutions on the boundary of the feasible region, while maintaining the same linear time comp
 r sharp transitions. The approach is tested during three task transitions, for different tasks such as constrained trajectory trackin

ze both control input and clutch sequence simultaneously. A penalty in the objective function ensures the prevention of unneces
stiffness actuators and underactuated compliant robots. We present a state-feedback controller that uses locally optimal feedb
preserving the benefits of model-predictive control, thereby outperforming traditional force control techniques. This work paves
0 and resulted in the publicly available library gafro. The benchmark shows faster computation of the kinematics than state-of-t
of a planar multirotor attempting a safe landing in an unknown wind field.
ms ranging from toy linear systems to a jet aircraft with a 16-dimensional state space. Finally, we validate our approach on a two

tabilizing controllers for continuous-time systems including the inverted pendulum, the cart-pole, and the quadrotor as well as a
he effectiveness of the proposed method in various scenarios.
ccurate. The control commands are also applied to the ground truth model, i.e., physics-based simulator, to further validate the
d on several reaching problems featuring non-convex obstacle avoidance with different dynamical systems, including a car mod
ling and iterative learning of the vehicle dynamics in high-speed domains is observed both in simulations and experiments.
choosing the noise's realization. The controller aims to minimize its cumulative tracking error despite being unable to know the
esired bound is characterized by some LMI conditions. The effectiveness of the overall arguments is validated through some co
mplete state, the implementation of the proposed method only needs to access the input and partial state information. Based on
ysis provides a formal guarantee of stability for a set of off-nominal states. The proposed algorithm has been tested on two diffe
ile skin) and use all sensors to optimally explore the given region. We show that our approach increases the exploration perform
or small-to-medium-scale problems and offers order-of-magnitude speed improvements for larger-scale problems.
ation and with real robots to demonstrate CODAK’s ability to enable safe navigation in different environments. We also exp
the effectiveness of our approach in generating learning models that accurately predict the follower's multi-step behavior when
paths so the convoy reaches its destination while minimizing the total time (cost) the two vehicles are active, including any time
s weight distributions as parameterized curves over coefficients. Our numerical results demonstrate that distributing and assign
i-robot systems. Our extensive experiments reveal that SAMARL outperforms existing baseline and ablation models in our desi
control horizon. Through ablations and dissections, we offer insight on the contributing factors to OBS-KATS's performance. O
esults confirm the effectiveness of our framework in assisting the robot reach its target by reducing transmitted information, on
collision. Our experimental results show that trajectory prediction can help make better decisions in task assignments for minim
approach is fast and accurate on real-world data.
use different types of robots, environments and human motion models. We show a possible application of the benchmark fram

result, our approach is scalable to arbitrary complex environments. Through extensive comparisons in simulation, we demonstr
cross a variety of environments.
babilistically complete, asymptotically optimal, and finds near-optimal solutions quickly. Experimental results with robot dynamic
opicity. We further allow agents to make short-term predictions about each others' paths, as a means to reason about each othe
forms a state-of-the-art online centralized multi-robot coverage path planning algorithm in terms of the time required to achieve
ging knowledge of the topology of the environment, we also find higher-quality solutions than other methods.
startup to further improve the efficiency of both the original MIP model and the reduced-size MIP models. We validate the effec
ode is released open-source along with the training images.
idated in an indoor environment, where ground truth is available from a motion capture system, and is subsequently deployed i
 n addition, the online path replanning ensures safe navigation through the map created during motion. The performance of the
s large-scale objects and the other optimized for regular convolutions. Their outputs are fused to effectively detect objects acro
formance evaluation, our pre-processing step improves other literature methods running time by a factor of 10 while maintainin
AM library achieves resource efficiency by having a data-sharing mechanism. Our experiments on three challenging datasets:
perimental results show that the drone can rapidly detect, estimate, and recover from collisions with obstacles in 3D space and
erical analysis using multiple simulated mission profiles. Our algorithm achieves a success rate of 88-96%, constituting a 16-29
nd redundant data in a point cloud. These contributions increase computational speed by an order of magnitude compared to s
ts show that our approach surpasses state-of-the-art methods by 50-60% in efficiency, which we measure by the fraction of the
global and local planning strategies are then in corporated to guarantee both high quality and efficiency in path generation. We
ing a team of up to 3 aerial robots, and show that our method outperforms the baseline by 33.4% in simulation and 26.4% in re
s multiple UAVs to quickly explore complex scenes in a decentralized fashion. By providing the decision-making capabilities to
deployments.
erred into different domains with fine-tuning while maintaining the success rate. Our proposed framework, which integrates onbo
en the CBF and state feedback from the VI-SLAM module, a safety filter computes a certified reference that best matches the te
f whiskers responds asymmetrically when multiple flow sources from different headings are applied to the array. This asymmet
cle detection. Besides, we introduce a new feature-based data association and tracking method to prevent mismatches utilizing
for efficient optimization resolution. Through validations conducted in both a photorealistic simulator and real-world experiment
ed to program different plant treatments. Finally we benchmark the proposed setup with a classical mobile manipulation approa
wind velocity. Nevertheless, our method can additionally make use of measurements of the roll moment induced by updrafts, w
ging datasets. Apart from its superior performance, our approach enjoys a real-time detection speed of 14.8 frames per second
robots, achieving a 12.5x reduction in bandwidth and a 20.5x speedup. We leverage our method to correspond 700k plants and
ng our approach to transform the reconstructed object point clouds back into the scene. Unlike conventional methodologies, wh
e-skins are generated for use in simulations. Finally, we demonstrate the benefits of 3D e-skins in a tool manipulation task.
controller, that constantly adapts the orientation of the suction cup so as to enhance the quality of attachment. Throughout all e
he proposed method substantially reduces the domain shift and thus leads to consistent performance improvements. Extensive
e manipulated, we reduce the ambiguity faced by learning the changes with the new object. We propose training two additional
of each candidate is updated in a Bayesian fashion using a likelihood evaluated on the image observation. Finally, we take the
manipulation. More importantly, our dish garnishing robot has a wide range of applications and holds significant commercializatio
nstruction from HandNeRF ensures more accurate execution of downstream tasks, such as grasping for robotic hand-over.
hen tasked to pick all trusses from a pile. 93% of the trusses were successfully grasped on the first try, while the remaining 7%

as well as improve the algorithmic robustness to low-light conditions. Extensive experiments are finally conducted on two newly
proposed approach and demonstrate satisfactory reconstructions of complex contact surface shapes.
interaction. In this paper, we implement a full system using shared autonomy for robotic opening of articulated objects, in partic
in of the grasp. This safety margin represents the extra lateral force that maintains the contact away from the frictional limit. To
ns these set correspondences. We compare SCAlign against other cross- pose estimation baselines on a synthetically generate
t closely rivals that of supervised methods.
ng its potential in handling DLOs under various, possibly uncertain, constraints.
e achieve a 93.3% success rate when clearing bins with over 300 unseen objects, which is on par with human subjects under c
esting success rate of 86.67%, with an average successful harvest time of 32.46 s, showcasing its continuous and robust harve
 ontrol to this grasp. In experiments on the DLR EDAN platform, we report a grasp success rate of 87% for 10 unknown objects
quid manipulation with a swept frequency sensing mode and a frequency-specific impedance measuring mode. Our developed
bot gathers experience in the real world, without constantly requiring human intervention. Through experiments on a Stretch RE
r improves recognition. Additionally, a trained classifier is also shown to be successful in distinguishing between shape categori

iple viewpoints before manipulating objects, our approach can estimate a more accurate pose without redundant manipulation
e demonstrate smooth, online object pose tracking with a live camera feed at >300 Hz.
fast training of the real-world manipulation neural policy while not requiring scene reconstruction at test time, but solely a point
n multiple robotic platforms where accurate state estimation is vital, and compare it to alternative approaches such as finite diff
earning methods. Dataset and codes can be found in https://sites.google.com/view/ad4rl.
evaluation metric that allows for a fair comparison in highly cluttered scenes taking into account the difficulty of the clutter. A we
ng to faster progress in developing robot manipulation methods. Appendix, code and videos for the project are available at http
he structure of the roads, models' performance, and data properties for fine-grained ranking of prediction models; 2) A set of ne
h error tolerance, suggesting that existing camera-based detectors already have the potential to surpass LiDAR-based detector

ative, RobotPerf remains committed to evolving with community input to advance the future of hardware-accelerated robotics.
escription of the garments used in the evaluation. This paper proposes a set of these measures.
sented framework.
ained on expert and beginner data, and achieved accuracies ranging between 94.8% and 99.97%. We assessed the system's
ne task execution. This generator determines the desired positions in the 2D plane according to the contact point displacement
anging objects.
dditionally, the Z-width of 7-DoF haptic exoskeleton is illustrated using experimental data and improved by exploiting varying vi
ing an overall success rate of 97.5%, even though the dynamics and the shapes vastly differ from the simulation setup.
uate the significant reduction in muscle activity during conventional and exergame-based squat sessions. In addition, we asses
s that impose varied dynamics at human anatomical joints is studied. Twenty-four healthy participants, eight for each CDLE co
ng the NASA Task Load Index (NASA-TLX) as a template, RS and PC questionnaires are designed and used in a series of pHR

data is competitive with state-of-the-art methods that rely on real human motion data both in simulation and on a real system. I
chitecture is loosely based on the human strategy observed in the human-human experiments, and the key component is a rea
robots can utilize to improve human performance. On the other hand, robots are becoming instrumental in quantifying the perfo

ator’s physical effort, task velocity, and task error. From the experimental results, it is discovered that base compliance has
with pilot human testing. Simulation results are compared with experimental estimation of interface power, showing agreement
ference-Based Optimization (PBO) guides the choice of the admittance parameters (i.e., virtual mass and damping) to better fit
balance the robustness of sub-map merging and exploration efficiency, we develop a new approach for lidar-based multi-agent
sumptions on the underlying trajectory-generating distribution. To the best of our knowledge, these are the first results that prov
of uncertainty. Moreover, the proposed approach is validated through two robotic applications, a micro Unmanned Aerial Vehic
gorithm is tested on an aggressive autonomous racing platform both in simulation and in hardware. The results show that the p
MPC)-like safety filter to apply corrections to the reference ego trajectory thereby promoting safer planning. The efficacy and com
ch in highly aliased simulated environments where unresolved data association leads to multi-modal belief hypotheses.
mplete set of hypotheses and the value function based on a pruned subset of the hypotheses, enabling us to establish a trade-o
 We propose four distinct shielding methods, differing in how the shields are computed and integrated, including factored varian
onstrating that our uncertainty-aware roadmaps effectively trade off between plan quality and planning efficiency for uncertainty
safe motions and accurate risk bounds and is safer than baseline approaches.
monitoring can be used to achieve runtime safety. We demonstrate COBeTS in several safety-critical, constrained partially obse
field. Our experiments demonstrate that our approach is the only one that can explicitly reason about high uncertainty both on 3
an obstacle) so that future trajectories will avoid unreliable motions. To infer the properties of a new transition, we introduce th
al dynamic environments, including large-scale outdoor environments. DAEP outperforms state-of-the-art planners in dynamic
actice, our approach results in a 6% improvement in coverage of traversable areas and an 89% reduction in trajectory portions r
proved plans. IBBT is efficient by reusing results between sequential iterations. The belief tree search is an ordered search gui
constraints, we consider conservative approximations of chance constraints in the form of distributionally robust Conditional Va
planner success is likely, all while producing agile motion.
ng on a large real-world dataset of challenging long-range planning problems from the DARPA RACER program. Our method, D
points. This information is seamlessly incorporated into the planning and fusion processes by adjusting the uncertainty associat
osed framework and algorithms using simulations and hardware, with robots exploring a partially observable, time-varying envir
res the unique challenges in the SLAM setting, and then, based on this formalism, we propose modeling the SLAM ASP as a p
s and physical experiments demonstrate that ASPIRe achieves real-time computation and outperforms benchmark methods in
ing this database, we propose a general experience-based POMDP solver, E-RTDP-Bel, that uses the solutions of similar plan
is applied to the path to ensure a safe motion despite tracking inaccuracies. We demonstrate the effectiveness of the massive d
BS scales up to 30 robots.
pling-based search algorithm, enabling the trajectory planner to avoid the unexpected risk of vehicle rollover and extremely high
ce the concept of the trajectory braking zone and adopt the chance-constrained formulation for robust collision avoidance in the
e it in two paradigms: First, we use RL for a team of agents to learn how to coordinate and reach the goal with minimal cost on a

o simultaneously leverage topology and geometry. The tree grows from the starting location and explores all topological routes,
and near-optimal LWCS algorithms, with the near-optimal algorithm targeting large maps. A complete prioritized planning meth
methods struggle with more complex instances involving large task-to-agent ratios, where frequent coalition changes are needed
e, given the known outcomes. We validate our results in simulation for a manufacturing domain and demonstrate that our metho
-level coordination, we implement the conflict-based search algorithm within each sector to ensure conflict-free local paths. In a

ntal results shows that smaller torque is required for the motor with the spring mechanism. Therefore, less electrical power is co
ound truth bone locations and those locations of US were recorded simultaneously. Therefore, we could label bone peaks in the
nstrates robust performance across different speeds and is capable of detecting variations in payload and electrode placement

uracy over classic linear interpolation methods. In the proposed controller, a weighted sum of Fourier series is used to model th
he AFP mechanism has the advantages of low power consumption, human-like joint moment profile. In particular, it is shown th
ocuses both on spatial and temporal changes in coordination using specific metrics. The results reveal that even brief exposure
of the EMG decomposition algorithm on our developed prosthesis hand control system. The control scheme provides an accura
istive strategy that automatically provides adequate support for lifting tasks and grants autonomy to users during walking. In the
ension, 40° for radial deviation, and 42° for ulnar deviation. The use of SMA spring actuators confers restoring force and flex
n in plantarflexion is evaluated in a stationary torque experiment demonstrating the ability of the frame to transfer large torque l
 ystems. Moreover, experiments with healthy participants were conducted on an active pelvis orthosis to validate the robustness
nder unforeseen forces thereby improving the overall user stability and safety. To validate the effectiveness of our approach, sim
walking with the other profiles. Interestingly, within-subject performance was more consistent than across subjects (P1: 52% ra
rques were commanded based on reference ground reaction force (GRF) and CoM trajectories. The subject could walk comfor
t stress during interactions with objects of various natures and weights. We conducted a study to assess the effectiveness of th
rated versions with a new dedicated large-scale dataset. The project page, code, and dataset can be found at https://theairlab.o
ous Localization And Mapping benchmark. The source code of the camera model conversions studied in this paper is shared w
ed approach that makes obtained point clouds resilient to rapid movements (e.g., aggressive turns or spins), that commonly oc
h representation is sufficient for metric localisation by registering point clouds taken under different viewpoints on the KITTI data
d achieves higher accuracy than the other existing approaches. Applying our proposed method is beneficial for improving the r
of-the-art approaches by achieving, on average, a 19% reduction in runtime on real-world open datasets.
y, the models exploit multiple point cloud measurements of the same scene from different view-points in order to learn to reduc
h an average IoU of 94.78% while maintaining a high run-time performance of 171Hz. The source code is available at https://git
nthetic ShapeNet and real-world Pascal 3D+ and Pix3D datasets. The proposed work achieves state-of-the-art results compare
bstacle avoidance with additional objectives, such as goal-seeking, to obtain a fully-fledged navigation system that is versatile a
ing navigation with wall-following, the third, hybrid algorithm enhanced the map traversal speed by roughly 26−43% compared
oes not work in a case where there are optimal and near optimal solutions in one scene, e.g., there are several paths to avoid o
anipulating a common tool in an unorganized table-top layout reinforces our claim of robustness in stationary and non-stationary
cenarios with cluttered obstacles, slopes, and external disturbances.
ractions. To assess the feasibility of the framework, two medical applications are presented in this work: 1) coil placement durin
n needed for the task execution. Our approach is capable of extracting such task representations from a single demonstration v
sful participation in the RoboCup 2022 and 2023.
ncluding the supplementary video showing the experimental and real-world results, is available at https://rb.gy/3ov2r8.
n by increasing the number of variables in the design, which was impossible with naive methods. The results show significant im
evealed inefficiencies in power transfer at the interface level, likely resulting from the compliance in the structural components
ational dynamics, and subsequently determined the necessary actuator output. The system characteristics were evaluated base
ersal is based on imitation learning by taking the analogous healthy limb of the patient as a reference and encoding the forces u
rn. To validate the effectiveness of the proposed method, the new ankle trajectories were tested on a group of healthy voluntee
y using 90 seconds of real interaction data, our RL-based control can perform the reaching with the average error over the wor
ss. The performance of this hybrid adaptive controller is tested in simulation and on one healthy subject. Our results indicate an
isually estimated kinematic model, a haptically estimated range of motion, and the person's exercise performance. The system
ate an offline dataset. We compare RbSL with three offline GCRL algorithms and one offline safe RL algorithm. As a result, our
robot task of box-pushing around obstacles. Further insights are available from the videos on our website: https://sites.google.
he efficacy of prior DRL algorithms but also presents formidable competition to contemporary top-tier RL algorithms, signifying a
urs and maintains the cost violations below a desired limit. We evaluate ASRL in a multi-robot system and competitive multi-ag
es higher task performance than the nominal baseline while also speeding up training convergence.
our theoretical analysis of SORL, we propose a condition for SORL's converged policy to guarantee safety and then use it to in
ur approach with three state-of-the-art DRL algorithms and find that differential image encoding reduces the memory footprint b

ggregation module (BLFA) to fuse them in both point-level and line-level manners, enabling a comprehensive feature representa
ving reflection and refraction through an intermediate medium. This paper aims to address the newly emerging problem of verti
dynamically improve the sampled actions to be closer to nearby optima before applying them to the environment. Our methods
g techniques solely utilize gradients computed based on private user data to optimize models, without storing or transmitting the
modate various types of robots. Extensive experiments are conducted to establish the first benchmark featuring several high-pe
m the distribution provided by the generator and reweight the simulated data using the discriminator. Our experimental results i
ocity of the drone is immediately affected by disturbances. Finally, we demonstrate the effectiveness of our method through ext
s comparable to state-of-the-art deep learning while substantially improving the performance and energy consumption efficiency
to estimate flexibility by analyzing their visual configurations. Secondly, we train a policy conditioned on the estimated flexibility
y learns 1) an embedding of belief LTL, which is multiple branching possibilities on LTL instructions using a graph neural networ
from multimodal foundation models and aligning these representations not only across vision but also haptic and auditory sens
ing path planner in quickly finding robot motion sequences. Furthermore, we demonstrate our framework in challenging scenar
amera distributions and view directions, which are not only limited to simple forward-facing scenes. Our key idea is to encode th
eriments, and demonstrate that it is able to speed up a commercial bin picking application.
se our approach on a new neural field dataset for evaluating registration problems. We provide an exhaustive set of experimen
e occluded patches using the remaining ones. Specifically, we design an asymmetrical encoder-decoder architecture based on
rious simulated and real robotic tasks, we show that imitation policies for robotic manipulators trained on POCR achieve better
ucture features of the supporting items. We conduct extensive experiments and demonstrate substantial improvements in our f
n. This representation motivates further research on data-efficient and transferable learning methods for online manipulation ta
multi-object scenes. SlotGNN learns a graph representation of the scene using the discovered slots from SlotTransport and pe

’s perspective instead of solely relying on the image attributes. We evaluate our approach on real-world datasets such as K
uds can improve performance, reduce learning time, and increase robustness for robotic learners.
hierarchical approach which employs a coarse voxel grid that captures the observed parts of the environment together with high
rained on SOFT(PVT) far outstrip standard PVT representations for manipulation tasks in simulated and real settings, approach

nodes to finish multi-resolution plane segmentation. The efficiency and robustness of the proposed approach are verified via sy
anges in depth distributions and scale. With rigorous quantitative evaluation, we show that our method outperforms multiple ba
w that the proposed method achieves better performance than other state-of-the-art methods under the same evaluation criteri

ncing credibility and enabling confident refinement of algorithms and models. Most existing works do not disclose their datasets,
clustering fashion. SupeRGB-D outperforms existing baselines on unseen objects while achieving similar performance on seen
segmentation. Our experiments show that MSMFormer achieves competitive performance compared to state-of-the-art method
proach on the SemanticKITTI dataset and compare it with leading SSC approaches. The SLCF-Net excels in all SSC metrics an
ourse, our strategy establishes a depth consistency constraint between different courses toward robust depth estimation in dive
patial and temporal connections of objects in a unified way for collaborative decision-making. We evaluate our approach using a
atches between the two point clouds and consider the practical case where these correspondences are plagued with outliers. W
ption pipeline solution with better performance on 2D DLO segmentation and 3D DLO reconstruction compared to State-of-the-
f scene geometry to match observations. To validate our methods, we capture 3,800 measurements of eight planar surfaces fro
ge-text pairs. By transferring the knowledge of the VLM to the 4D encoder and combining the VLM, our VG4D achieves improv
rge 3D datasets or finetune models. We demonstrate the utility of this representation through a number of downstream plannin
trated experimentally. The source code and demonstration data are publicly released.
 D domain. To address these challenges, our work proposes adaptive regularization for frequent classes and sampling-based re
weight, and effectively manipulate target objects. This demonstration shows the real-world application of the proposed joint desi
substantially shading each other. In order to address these issues, we introduce OriTrack, a small three-degree-of-freedom (3 D

two working modes. The gripper exhibits coiling diameters ranging between 40 mm and 60 mm and is successfully attached to
6 collaborative robot by Kinova. In this paper, we study the complex singularities of Link~6 by investigating all possibilities for sc

e its motion and performance.

ces and the gripper mechanism. A significant advantage of the proposed robot over serial architectures is that the robot is back

om 22 different robots collected through a collaboration between 21 institutions, demonstrating 527 skills (160266 tasks). We s
eeper insight into the phenomenon of SARA.

multisensory pre-training for robotic manipulation.

asks in a zero-shot manner with no deployment-time training. Importantly, while the plan predictor can leverage large-scale hum

essary and sufficient conditions for FT-NCBFs to guarantee safety, and develop a data-driven method to learn FT-NCBFs by m
sed approaches while allowing control frequencies of up to 1kHz.
to move from any state within this set to any other state. To illustrate the use of our algorithm, we provide the numerical examp
first using a 2D mobile robot example, then on the Franka-Emika Research~3 (FR3) robot manipulator both in simulation and e
s and elastic potential energy. In addition, safety constraints, such as avoiding obstacles, can be specified using signed distanc
ble sets that bound the future states of the system, as well as numerical results demonstrating how it can be used for safety ve
al one, using a walking-in-place task against frontal impacts. The statistically best controller was able to withstand the impact m
ety of the overall system. We also design a fallback controller that robustly handles these detected anomalies to preserve syste
d the environment or which is carrying an unmodeled payload. We use a conditional variational autoencoder (CVAE) to learn a
dcopter despite the error from the perception module, while the baseline algorithm without using the learned IPC failed to do so
cles. We test our approach on simulation experiments and a real-world implementation to validate it.
han classic methods. In particular, recursive feasibility is guaranteed by making the safe-set constraint move backward over the
mpared to the state of the art on systems up to dimension 6, our algorithm resulted to be more than 2 times as accurate for the
onstrate the impact of semantic information and scene change handling on robot behavior, validating the practicality of our appr
tuator architectures for MR actuators with no other changes. The paper also presents a simple, reliable, and fast collision detec
iments involving high-dimensional belief spaces.
ound humans and robot manufacturing.
, we develop a broadcast scheme that estimates driver intention and integrates BCBFs with multiple backup strategies for huma
By repeating the feature extraction and node pooling, we obtain a graph hierarchy. Finally, we determine rotation and translatio
on to the training data density by starting from a fitted Gaussian mixture model instead of the conventional standard normal dis
ucture of the square-root covariance matrix. Furthermore, we choose a special ordering of the state variables which is amenab
scenes with sparse points. Extensive experiments conducted on large-scale datasets show that SeqTrack3D achieves new stat
rld datasets to evaluate the effectiveness of the proposed approach. The results demonstrate the achieved tracking and mappi
extract visual subgoals embedded in the videos, while incurring zero additional training cost on top of standard visuomotor polic
he apple 80% of the time, regardless of branch stiffness. Moreover, the gripper withstands ±20mm offset when it approaches
 method based on plane-constrained binary division point cloud completion. This method separates the plane point cloud and s
to classify the kernels as normal or defective. We also built and released a large-scale dataset, named GrainDet, that includes
hing transformer. The original LoFTR network was trained using imagery of outdoor urban areas. One of our contributions is to
attempted with computer vision-based methods).
o achieve this we propose to use an additional neural network to predict the quality of each segmented leaf and discard inaccu
ect website: https://kantor-lab.github.io/tree_gnn
negative, meaning that traffic congestion in the multi-robot system used in the experiment was alleviated without the need for in
propose a novel self-supervised task-specific pre-training approach to initialize the backbone of a network for leaf instance seg
ffector; and 3) learning-based machine vision for automated visual identification of leaf xylem wetness during SWP analysis. Ta
eriments were conducted in the greenhouse and tested under various conditions. The time cost is formulated as a function and
learning. By leveraging pre-trained neural networks, accurate semantic segmentation of fruit in the field is achieved with only a
points in the form of a graph, and demonstrate that different types of plant phenotype parameters can be obtained in a fully au

e proposed control law is proven, and the overshoot reduction after the object extraction is analyzed. To show the improved pe
warm-start our optimizer. We improve the state-of-the-art in the following respects. First, we present a detailed analysis of the r
a UAV that navigates from the base to the target using an RGB camera.
or agricultural environments consisting of rowed crops and through our proposed system emph{Osiris} incrementally construct a
r data rate which can be stored to disk in real-time is increased for large robotic data types like images and point clouds in com
ich can control the end-effector to grasp along the direction of the fruit peduncle and perpendicular to the tangent direction of th
effectiveness of these proposed methods and problems we have to tackle on. Basically, it was possible to perform automated t
e the best performance, we train the pH-estimator with different input modalities and validate each result using Mean Squared E
eld trials in a sweet cherry orchard, ultimately achieving a cutting success rate of 58% across ten trees. Though not fully robust
ulated tests are used to evaluate our proposed algorithm's behavior in a variety of settings. Physical experiments in agricultural
s images. Through simulated experiments, we demonstrate that our viewpoint planner improves sizing accuracy compared to s
rajectories due to inefficient sampling. We propose a gradient-based NBV planner using differential ray sampling, which directly
sing RGBD camera images (obtained from an eye-in-hand configuration) and one demonstration for each task, it is feasible to p
dataset. This allows the network to reason about higher-level semantics of the entailed objects, thus improving its performance
labels corresponding to the generated images. We show that our approach can improve the performance of the crop-weed seg
all logs within its reach on a trailer. Notable features of our implementation are a generalizable perception stack, a grasp plann
true waypoints of the mowing robot. Three cruising demonstrations are presented to show the mapping and optimizing results.
ntation and tree trait estimation performance, underlining the potential for automating forestry services.
hetic data and is shown to evaluate multiple hypotheses correctly. Its effectiveness is also demonstrated on real datasets. The
round and herd the animals. Moreover, the animal-robot behavior model resembles interaction of livestock under the presence

ed distance.
ndamental assembly tasks: the robot assembling another robot, and the two robots collaborating to assemble a small structure.
rial facility containing static, semi-static, and dynamic objects. Our method demonstrates superior performance in detecting cha
o large areas.
ard function for the material loading task. Finally, we examined how the sampling time of the reinforcement learning algorithm a
olicy that can be generalized to various domain parameters in excavation tasks. We propose two techniques for improving the p
orm automatic NDE data collection; 2) an IE data processing module that utilizes both learning-based and classical methods to
 reasoning. Such coarse-to-fine inference allows the robot to decompose and achieve long-horizon tasks interactively. In exten
h text descriptions and more than 3M objects, surpassing prior datasets. Empirically, we show that Grasp-Anything successfull
ate a 31% reduction in execution time compared with a system that does not consider upcoming tasks.
generalization with as little as a single demonstration. We have made Language-World available as open-source software at htt
ia self-instruct and compare its performance against a pre-trained LLM like GPT-4. Results are demonstrated on several tasks
ts the potentially inadmissible LLM guidance in multi-heuristic planning. We demonstrate efficient planning of complex natural l
t of skills (specified in language) and associated preconditions, where CAPE improves correctness by 76.49% with higher exec
cepts out of the training set. The effectiveness of GraspGPT is further validated in real-robot experiments. Our code, data, appe
ty of the deployed robotic systems. To illustrate this point, we demonstrate performance of DOS by training and deploying deep
d performance of such a controller in a high-fidelity scenario using the Robot Operating System and Gazebo frameworks.
uniformly ultimately bounded. Simulation and experimental results, including comparison data, are provided to validate and as

effects. Additionally, we establish an upper bound for linearization error, effectively linearizing the nonlinear system, reducing O

d data but also adapts the model to environmental changes, even when the model parameters exhibit substantial deviations. W
k-sensitive behavior of the proposed filter when used in a nonlinear MPC loop on a planar drone and industrial manipulator in s

h through experiments on a fixed-base robot performing stable fast motions, and a floating-base humanoid robot robustly react
approach to limit the velocities of both the robot's base and arms, adhering to injury biomechanics-based safety thresholds. Th
varying conic constraints from the variability observed in expert demonstrations, and bounding the evolution of the DS with Con
method can check distances for 500 waypoints along a trajectory within less than 1 millisecond using a GPU at runtime, making
sk. Thirdly, we compute two intrinsic object properties. The concatenated outputs from these branches pass through extra dens
itable collision avoidance responsibility. We evaluate our approach against well-established multi-agent planning algorithms and
with the human-driven vehicle while actively inferring human internal states. The simulation results, demonstrating the lane chan
milar level of accuracy. Experimental findings demonstrate that integrating this strategy into an object detector results in a notab
he robot is passive when feasible. Our approach is validated in simulation and real robot experiments on a 7DoF Franka Resea
, and compared to other data selection methods, has 10% higher accuracy with 10% training data selected, and 7% higher robu
or localizing human joint estimates. We then establish a convention for a coordinate system that encodes the direction the hum
ch module.
e identified from tensile tests and dynamic excitation experiments. Dynamic simulations confirmed that the controller reduced th
ation rate required and target response time, finds the minimum number of robots able to meet the SAR mission goals and the
ons to construct the global navigation layer. Another merit of the proposed scheme is the ability to handle scene changes (e.g.

d on the reported results, it is found superior compared to the existing stochastic randomized dual sampling schema.
ng-range and low-power communication technology from the internet-of-things domain (i.e., LoRa) and a customized communic
ios, but escape the burden of data-intensive learning and operate safely (transparency and causality — we fit sensor data into
le autonomous map generation. The proposed method significantly outperforms a variety of baselines, including an object-base
on the ScanRefer benchmark and demonstrate state-of-the-art zero-shot grounding accuracy. Our findings indicate that LLMs s
of self- supervised labels by incorporating methods of self-supervised learning of visual representations. To conduct a compre
which has a fixed volume of voxels. Thus, we propose a new LiDAR 3D semantic segmentation network (Hi-VoxelNet) with Hyb
sting panoptic segmentation method. Notably, we demonstrate that SPINO achieves competitive results compared to fully supe
s demonstrate that the proposed network surpasses existing methods on the Cityscapes dataset and notably outperforms state
odal fusion with higher flexibility. To enhance the 2D representation and 3D representation, the Dynamic Representation Enhan
Radiance Field (LERF) for semantic monitoring. Experiments with 3-5 objects arranged on a tabletop and a Turtlebot with a Rea

ovel neural network design for lightweight semantic map post-processing. Our system achieves state-of-the-art semantic mapp
y efficient. Moreover, the proposed framework can be applied to both structured and unstructured indoor and outdoor environm
bulary query without the need for additional 3D training. Benchmark tests on the ScanNet dataset against leading zero-shot me
s critical to promote spatially compact instance predictions as spatio-temporal instance queries tend to merge multiple semantic
g like clustering or associations between predictions. We extend a mask-based 3D panoptic segmentation model to 4D by reus
ed Global Scene Priors (GSP). Then, efficient semantic navigation is achieved by employing an immediate LSP to retrieve cond
y for generating a 3D scene graph dataset based on semantically annotated real-life 3D spaces. The generated dataset is then
ntitatively and qualitatively by improving alignment between estimated topics and scientific interests.
at are co-learned through both semantic segmentation and unsupervised stereo matching tasks, enabling the features extracte
mpactness of the graph. We eveluate our method on two widely used 3D point cloud scene segmentation benchmarks within sp
lso enables reliable, class-agnostic instance assignments. In sum, our approach shows superior performance on the new movi
e the importance of map calibration on a downstream task by showing that incorporating proper semantic fusion to an indoor ob
composed of a mask generator and a masked cross-attention mechanism, is used to refine the features from onboard sensors
tion loss encourages the network to extract complementary and meaningful representations from a single modality or complem
we segregate the environment into intersecting and non-intersecting road areas. Subsequently, we construct a multi-layered sc
tion weights of LiDAR and camera BEV features to each other. Our extensive experiments demonstrate that the suggested me
inally, based on the updated map, the hierarchical path planner efficiently searches for energy-saving paths for navigation. We
ns of dynamic features. Then, we design a neural network that densifies radar data and eliminates spurious radar points. We bu
urrogate metrics, We conduct experiments using several camera-LiDAR placements and parameters inspired by self-driving co
poral information from events while preserving crucial edge information. % for fusion with RGB modality. Secondly, as different
era branch) to a 3D network (2D knowledge branch) so that the 3D network can generate 2D information even for those points
valuable semantic information from optical images while preserving spectral knowledge from the IR modality within the feature
the plug-and-play characteristics of our module, we were able to push the performance of existing state-of-the-art model to a ne
estimation and use image features to guide long-range LiDAR feature prediction. We benchmark our SuperFusion on the nuSce
ental data at https://github.com/KIT-ISAS/SFUISE.
sion framework is established, which combines almost all GNSS measurement error sources by fully considering temporal and
s into the same image coordinate, discretize the timestamp of events to obtain a time image that can reflect the motion confiden
omprehensive spatial data fusion. AYDIV's performance on the Waymo Open Dataset (WOD) with an improvement of 1.24% in
ur proposed method significantly outperforms the state-of-the-art Radar-Camera depth estimation methods by reducing the me
annotations, and depth information, HabitatDyn 2.0 also features top-down view labels for dynamic objects, which is required fo
gs3D, as well as generalization on the real-world dataset KITTI. We illustrate that our approach exhibits enhanced robustness i

e-grained structure at both daytime and nighttime. Our method consists of two key technical components. Firstly, we propose a

angular velocity using accelerometers during extreme rotations caused by tumbling. We show that our method reduces the me
s, a lateral Human-ware Proposal Network is designed together with an Anti-shake Proposal Compensation module. It alleviate
ata, to be beneficial to network performance and generalization capabilities.
eration in the contact force during the transition. Furthermore, although the preemptive impact reduction uses a crude optical pr

k with respect to these modeling parameters.

VO tasks in complex illumination scenes.
and optical flow. Results indicated that MagicTac could capture fine textures and is sensitive to dynamic contact information. T
ncluding those involving simultaneous sources and various sound events. The second training step enables our approach to es
that the speech recognizer and sound source localizer work well in the reverberant indoor environment with a small single-boa
brush is fine-tuned to generate stylized trajectories for robotic arm control. Evaluation in simulation and physical robot scenario
pproach yields accurate estimation at a competitively low inference time in challenging greenhouse environments.

sent a series of experimental evaluations in both simulation and real-world settings, demonstrating the effectiveness of our appr
ous failure cases.
e one. Both simulation and real-world experiments validate the efficacy of our method, in spreading and recover a quilt over an
eneration, feature clustering and distance decoupling are proposed to further improve efficiency, precision and generalization.
Supplementary material in: https://sites.google.com/view/neural-uncalibrated-vs
action and IoT-based vital sign extraction modules. The ROS-based semi-autonomous navigation facilitates the robot to be an
manipulator for nursing robots has low requirements for workspace, suitable for complex spaces and can have a large carrying c

ion methods to satisfy the requirements of each perception task and scales to different robot platforms. The generality and effe
eneralized pattern search, particle swarm optimization, and surrogate optimization are independently considered optimization te

vior. We also propose a set of metrics to evaluate the faithfulness of crowd simulation. Experimental results demonstrate that So
We apply MuRoSim for training navigation policies for omnidirectional mobile robots equipped with lidar sensors using DRL. Fin
state-of-the-art deformation and frictional contact models. Our system uses the Finite Element Method (FEM) to model deforma
undulations tend to apply excessive force, displacing real fish more than expected. Additionally, the vortices and correspondin
ned on our synthetic data, a yolov8 model achieves performance close to a model trained on real-world data, with 5% performa

available on our project webpage.

ure system to compare the behavior of a real robot with its simulated counterpart across various motions. By leveraging expert
OAM), whose real-time odometry relies on keypoint descriptor matching-based RANSAC. The sparse keypoints are further use
at includes GPS data and 64-beam, 125k point LiDAR scans totalling 8M points (raw packets also provided). The dataset consis
ccuracy and achieves state-of-the-art sonar odometry performance. The source code is available at https://github.com/SenseR
he-art 3D mapping results and competitive LiDAR odometry accuracy. We will release the CURL-MAP codes for the community
method is evaluated in real-world tests on a flying robot experiencing degraded conditions including geometric self-similarity as
challenging real-world scenarios such as narrow corridors, featureless spaces, and dynamic objects. We show that our system
on outperforms existing low-cost SLAM systems in corner cases. We release the code and datasets at href{https://github.com/S
sting implicit field-based methods in challenging scenarios. HERO-SLAM provides a new pathway to enhance the stability, perfo
evaluations, conducted on three public datasets that include both dynamic and variable lighting sequences, demonstrate that ou
straints without affecting any landmarks. Our method decouples the management of landmarks from pose-only measurements,
cribing these maps, as well as the size of the database required to store these features over a long sequence. This yields a rea
ining its consistency with the normal estimation. We further fix the distribution after its convergence to balance the time consum
illustrate state-of-the-art performance of our odometry techniques in terms of accuracy when compared to existing alternatives
ed to a subset of the global problem, where all previously integrated inlier measurements have veto power. We evaluated IPC o
h extends the applicability of DFM to accommodate a wide range of backbone networks pre-trained across diverse computer vi
drics, employed to represent physical objects. It facilitates the creation of a 3D map that closely reflects reality. Next, we use ob
novation lies in the multi-feature sharing architecture, which fuses the semantic and depth maps as pose features, thus significa

motion, we introduce as a prior the gyroscope measurements via pre-integration. Experiments on publicly available datasets ju
environment. To extract this high-level information, we present novel room and floor segmentation algorithms utilizing the map
d the evaluation of VPR algorithms, e.g., the evaluation of a VPR system that has to find all matching database images per que
ddress Doppler velocity measurement uncertainties. We present a method to optimize body frame velocity while managing Dop

y, when dynamically fusing the three factors together, the absolute trajectory error for both monocular and stereo configurations
ame optimization to increase accuracy. We provide the source code and some experimental data on https://github.com/davidsk
prove the odometry accuracy, we propose a residual adaptive weighting method and a probabilistic point cloud map based on
ons used in depth-supervised neural implicit frameworks. Finally, this paper shows that LONER estimates trajectories competit
erimental evaluation suggests that the proposed system performs on par with the state-of-the-art LiDAR-inertial odometry pipeli
Multi-LIO emerges as a robust, scalable, and extensible solution, adaptable to various LiDAR configurations.
he problem. To this end, we introduce a front-end agnostic framework using GTSAM that can be used to evaluate various Dyna
ane and the uncertainty will decrease significantly which can further improve the performance of LiDAR(-inertial) odometry. Exp
, up to the second order, of the BA optimization based on the point cluster coordinates and show their theoretical properties suc
iguous static points, guided by the inherent characteristic of dynamic objects in urban environments which necessitates their int
cets in each voxel. To the best of our knowledge, this is the first work in literature that can reconstruct online the triangle mesh o
y sensor selection and fast convex relaxation-based post-hoc verification enables the efficient recovery of emph{certifiably optim
BoW) model to deal with hybrid point features in matching and loop-closure-detection. By integrating the proposed framework i
tended Kalman filter (EKF)-based object-visual simultaneous localization and mapping (SLAM) with a camera-object relative po
m. Second, the edges of the measurement graph can be distributed evenly. The algorithm considers the minimization of the larg

mapping in the mentioned environment like smoke. Therefore, the presented approach in this letter solely depends on the UWB

xtured environments, using the linear Kalman filter (KF) framework. Our experiments in various indoor MW spaces, including op
ognition methods using diverse datasets with varying qualities. Experimental results demonstrate that our method outperforms
nt places. Our proposed system is evaluated using both public, simulated data and real-world data, covering office and home e
rticularly in adverse environments, but also leverages the inclusion of numerous object poses to ameliorate and integrate errors
se accuracy and furthermore produces globally consistent volumetric occupancy submaps which can be directly used in downs
ompared to existing methods, both in controlled simulation experiments and real-world deployment.
hod exhibits the highest reconstruction quality among all planning methods and superior planning efficiency in methods involvin

ation to maintain the temporal fidelity and asynchronous nature of event-based cameras through Gaussian process regression w
ases from significant errors without the need for error uncertainty adaptation, observability constraint (OC), or other consistenc
pose of vehicle. Through localization experiments on the KAIST dataset, it has been demonstrated that the proposed method
ation policy, which identifies useful vulnerabilities of both navigation policies for further improvements. Moreover, these non-coo
ctively perceive its environment to perform grounded commonsense reasoning. To evaluate our framework at scale, we release

irections and the associated constraints. Our work advocates for fostering socially adept and safe autonomous robot navigation
tions of each agent. The spatial-temporal graph reasoning ability allows the robot to better understand and interpret the relation

signing social robots that are time-sensitive in interacting with humans. Future explorations should focus on the cognitive and em
discussing cultural themes with their remote peers through the social robot.
a case study for context-aware conversation adaptation. The results are very promising. Our proposed method can achieve a m
rnet of Things, and intelligent manufacturing. Future research will focus on expanding the Rotenna concept to different types of
ovide a robust performance evaluation based on the generated postures. Finally, preliminary tests were conducted on a two-se
e gripper is capable of capturing a flying table tennis ball with a high speed of 15 m/s in only 6 ms. When the proposed gripper
the flow of the beads, an initial noncontact, frame-type strain constraint structure was designed for the soft actuator. Experime
mbination, is modular, scalable,lightweight, and low cost in comparison to existing solutions.

w that our proposed agent can be comparable to or even outperform a state-based agent. In particular, the sample efficiency al
sults, it is concluded that force sensing is replaceable for the environment interaction.
rs. Our results with ten experimental varieties showed a picking success rate of 95.4% demonstrating the capability of a closed
itions between walls with different pinch angles within 180°. Additionally, the soft robot can carry a miniature camera on vertic
he use of path planning to attain a desired position. The size of rolling step can be changed at any time by dynamically varying
a simplified way for position control. The entire workflow from mechanical design to actuation and control is presented. A proto
ed object recognition at an accuracy of 93.78% for standard geometric objects in 9 categories based on a machine learning mod
g the aerospace connector for communication is designed based on the drive device away from itself to reduce the inertia of the
safety of the endoscopic intervention.
ed spaces, inclined surfaces, gaps, and stairs, or to avoid obstacles by circumventing them. Our study substantially improves t
or harvesting experiment, the gripper attempted picking 26 berries without ripeness sensing, with 65.4% (17) being successfully
nly distributed on the double-wing structure of the robotic fish to realize the integration of the energy control system. This circuit

on of many antennas. We control the array with a single focused IR light that rasters across the actuators to precisely control a
nts one of the first soft robotic hands with actuated fingers and a passively compliant palm, all of which have high-resolution tac

h an adhesive force of up to 27.7 N in a 40 × 40 mm contacting area. We introduce the design, working principle, fabrication,
n origami crawling robot and an origami robotic arm using 3D-printed origami structures driven by TSAs.
tools. Furthermore, we suggest an automatic transportation method to enhance manipulation precision. A series of experimen
ed as an encoder to extract the latent representation of the geometry in a low dimensional space, which is further used to cons
al design method paves the way for designing soft robots with complex and dexterous deformations.
ulator is experimentally verified. The robot can achieve an elongation of 198% and a maximum bending angle of 240°. The ro

ness. Furthermore, the glove's effectiveness was confirmed through testing on a stroke patient with MAS = 2, validating its on-b
ming through the alteration of folding patterns of the flat tube and the material structure of the elastomer laminates. Finally, an e
ficing their inherent advantages or interfering with their inner chan- nels during retraction. The proposed mechanism’s straig
nt in orthogonal directions. The film is used as the vine robot body, oriented such that it can stretch over 120% axially, but only
ependent variable stiffness joints, which are the key feature for guaranteeing different bending angles of each segment. Results
Operation is validated when gravity is both tangential and perpendicular to the manipulator backbone, a feature uncommon in p
an azimuth prediction loss function which reduces the RMSE by 3.24â—¦ compared to L1 loss. Finally, we conduct experimen
e friction coefficients and grasp force.
of generating geometrically complex designs such as a double-helix network for a twisting actuator is outlined. Its possible exte
n process of the sensor. Finally, a complete test platform was built to test the performance of the rigid-soft hybrid robotic glove w
ment methods models, to create a scalable framework for selecting the best performing design parameters for each application
At the most rigid stage (0% alignment), the stiffness exhibited a significant increase of 111.1% compared to the most flexible s
errors of 0.06 N and the proposed operation during neurosurgery.

The R-Square for prosthetic grip force and hand grip force were 0.867 and 0.9724 in the periodic and sustaining grip force exp
pproximately 3% average relative error and taking into account internal fluidic actuation, our estimate of external force disturban
tactile sensing as a means of inferring the required robot motion paves the way for imitation of more contact-rich tasks, and ena
e experimental results obtained for different DEAs (four configurations, two materials and multi-DOFs) demonstrate that our dy
work-based tip force estimation method that utilizes arbitrarily placed camera images and actuation inputs to predict applied tip
position sensing shows accuracies with errors of 3.1%, 3.0%, and 6.7% for the x, y, and z axes, respectively, using feature extra
foundation for future “growing” robots that are capable of seeking light or heat, yet are extremely simple and low-cost. P
e surface quality during the morphological evolution, by imposing the geometric constraints in terms of the principal curvatures a
ponse time of the hydrogel valves with three different crosslinker concentrations. Among the tested concentrations, we found th
Furthermore, the sensor is demonstrated in two applications to detect foot placement and grip location. Overall, the sensor repr
additional computational potential for advanced high-frequency control algorithms. Control structures can be switched at runtim
ROS applications, a collection of useful machine learning-enabled base container images, as well as a CLI tool for simplified in

ed with the introduction of AutoGDM+, capable of procedural environment generation, CFD and particle-based gas dispersion s
caching scheme that dynamically builds a queryable action library at runtime. We instantiate Cook2LTL in a realistic simulation
ne the architecture of this system and strategies for task decomposition and code generation. Finally, we demonstrate how our

ing Nav2, we developed a system that using a depth camera performs pixel level classification on the images, and in real time
estimate of the uncertainty associated with the enhancement process, vital to ensure mission planning integrity. This work demo
n the Zero-G Lab validate this approach, showcasing the adaptability of the PPO agent. This research opens new possibilities fo
ris at the launch of the net, computes the final aiming positions of each MU, from which their trajectory can be derived. Each M
OST is performed on images of the target viewed from a diverse set of directions during RPO.
and subsequently position itself at the desired pre-grasp location. We assessed our approach through a series of experiments
mpared to RGB sensors. To facilitate further training and evaluation of DL-based models, we introduce a new dataset, SPADES
stereo vision systems, making it particularly suitable for embedded systems with computational limitations. Functionality includ
ng significantly faster and more precise localization without any reliance on ground infrastructure.
trol performance enhancement.
closedform Minkowski sum method is employed to solve the robot’s free space, and a concise locomotion method is elucid

ling the pendulum and understanding the shell dynamics. The paper concludes with experimental validations of these models a
MI is 2000 N. In further experiments, new contact blocks responsible for the transfer of electrical power and data were tested fo
d with controllable stiffness joints and a linear spring-damper system. A free-flying robot on an air-bearing platform can success
ependently change diameter while rolling. The main frame of the VWDER can keep the top of the frame facing up under the act
lus Hinf controller, in the presence of parametric uncertainty, noisy measurements and disturbances, using a planar example. T

ch, we conduct comparative simulations with existing methods and perform experiments using phantom and cadaveric soft tissu
nical systems, without hardware changes, and is available to the dVRK community through a software update.
nd input window by relying on a computationally efficient model.

alyze and discuss the success rate, trajectories, and recovery behaviours for various models that are either using the recurrent
ons. To address these limitations, we propose an improved simulation environment for surgical robotics that enables the automa
us on making the simulator more efficient, making training data much more accessible than previously possible. We introduce
ding of dynamics inherent in manipulations, thus making the model's reasoning ability to be task-agnostic, we also design a cros
e shortened to increase its stiffness when it reaches the target position, creating a stable surgical platform for procedures like bi

modeling by capturing local deformations and long-term dynamics modeling by introducing global temporal compensation. We
a residual mapping and (2) optimizes its stiffness parameters online. Our method ensures a small residual gap between the sim
erent methods in several anatomical environments including a real-life CT scan dataset. The experimental results show that our

s to assess their ability to generalize to unseen users and tasks respectively. Gesture recognition models achieve higher accura

Project website: orbit-surgical.github.io

e proposed algorithm processes registration over 50 Hz (rigid) and 20 Hz (nonrigid) and obtains competing registration accurac
e of a simulation policy with no additionally acquired physical training data. The differences in simulation and hardware forward
rototype CPPRs with two tubes and three tubes and non-constant laser-cut patterns that create variable curvature and stiffness
ues, establishing ground truth using 3D stereo vision and cone beam computed tomography (CBCT) acquisitions, respectively.
ovel nested calibration algorithm for sensor positions and magnet dipole strength is introduced. This letter further proposes a no
nimal misalignment of 3.4±0.9°. The hybrid Force/Shape Observer (FSO) algorithm was able to predict distributed external f
a method based on an improved N-wire phantom is proposed. Twenty puncture experiments are performed in different breath-h
action. We validate our approach by using DDC to control a collection of Crazyflie drones to achieve formation flight and reach
ble option for UAVs with limited capabilities. Our method is adaptable to other swarm platforms with only minor parameter adjus
s, and (ii) to enable a gradual transition from simulation to reality. Moreover, M-ARK enables the interaction between swarms d
eighbors. To demonstrate its advantages in the morphogenesis of robot swarms, we designed two morphogenetic engineering

mation. This work addresses the interests of the active matter, swarm robotics, and control barrier functions communities and d
evolution to produce the control software for the swarm of shepherds to coordinate the sheep. We show that automatic design
e plugin simulates a radio communication delay between the drones and the ground station to test offboard control algorithms a
and prior works. Further, we analyze the agent's behavior for the usage of the container and demonstrate meaningful generaliza
d more supplementary materials on our project website at url{https://branvu.github.io/coast.github.io}.
nstrate how to model different objectives and constraints using the CGA-CDTS. Using a setup of two Franka Emika robots we

and object proposals. The next target is selected by prioritizing object search while avoiding excessive robot backtracking. The
euristic to balance branching width and compute depth using polynomial level sets. We show completeness of the resulting sol
ndly, we demonstrate that adopting a continuous sampling strategy, which has been generally overlooked in prior research, ma
path trajectory optimization. Our framework proves to be time-efficient in computing optimal and collision-free solutions, while o

sing modalities via our Mutually Attentive Fusion method. Through experiments with diverse comparisons, we show that an age

orld problems. Our benchmark results show that our MILP method outperforms state-of-the-art open-source TSP solvers OR-To

bility of the previous one is unlikely. The performance of PROTAMP-RRT was assessed in simulated pick-and-place tasks, and
n with four different baseline methods. Our experiments show that our solution outperforms the baselines when the budget is tig

P to form the substrate for a community-driven endeavour that evolves as capability matures.
ates for the human hand, namely pig claws and chicken drumsticks. We then demonstrate the ease and efficiency of robot use
s and actions to solve long-horizon planning problems. We evaluate our planner in an object rearrangement domain, where we
o never-before-seen scenarios.

ynomial level. Moreover, the effectiveness of the proposed method is demonstrated by solving some benchmark problems and
reedy algorithm causes each robot to consider only the exploration area with the largest gain for itself, which easily leads to rep
e awareness about the robot's response and interactions with the environment is greatly improved. We evaluated the proposed
ollected from 18 participants. The results show that weighting improves task performance in reproduction by up to 42% in the fo
orque. The mechanical joint compliance of the FDA allows the passive accommodation of the arms to the physical interactions w
tems. User intention recognition is done at the operator side. Based on user intention recognition, a deep reinforcement learnin
ile rendering interface, called the Soft Pneumatic Tactile Array (SPTA). We carried out two user studies to assess the system's
of force feedback for complex environments. An experiment without human-in-the-loop was conducted, showing that for an onl
ng performance.
hich is identical to actual tool already used to accelerate training. POstick-KF has large workspace and high kinesthetic feedba
robot EDAN.
different control modalities. Our experimental results show that most users were able to succeed by building at least a tower o

d in two simulation and one real world environments and we show that our method can greatly increase the physical plausibility
ase without the optimization action.
ere, at any time. The code is available at www.github.com/wearable-motion-capture

dentification. We implement the algorithm on a virtual reality (VR) setup to teleoperate robotic hands in a simulation with variou
orms of DMM: grasping a target moving at an average speed of 0.4 m/s, and pushing boxes weighing up to 105% of the robot's
 tended motion autonomously. Also, we introduce a Mixed Reality (MR) user interface for better human-robot interaction. Finally
mIoU, we evaluate the number of tunable parameters and additional FLOPs for a fair comparison of our method to different fine-
istency loss leverages the 2D prior semantic masks from SAM as pixel-wise supervision signals to encourage consistent predic
hese mapping functions. Once the mapping functions are learned, we can seamlessly transfer the policy from the source doma
c datasets and multiple 3D tasks, i.e., point cloud recognition, few-shot learning, and part segmentation. The implementation is
etric-space Aligned Student (GAS) model. DAT tactfully combines target lidar and reliable depth prediction to construct depth-a

e higher frequency levels require much less RL trials, thus significantly accelerating the learning process. To implement our ide
a. We refine the source domain controller using target domain data to enhance the controller’s adaptability to the target dom
oving devices to perform dance. This method is applied to real dance scenarios, including an inclusive dance show where danc
hen placing the obstacle on the front and side of the foot, respectively, compared to the previous foot. In addition, the proposed
ding wires and delivering objects during the welding process was given to confirm the auxiliary capability of this robot.
force can be kept over 1000 N on the surface with curvature radii ranging from 1.1 to 14.6 m. Equipped with a small manipulat
thod accomplished the insertion tasks with the position-controlled manipulator. We considered that some hardware problems ca
chen setup. Our goal was to develop a mechanically intelligent platform with proprioceptive actuators, which is highly capable e
ause Wi-Fi can penetrate most residential building materials, a single sensing device covers multiple rooms and through non-m
r smartphones. This synergy of AI and human knowledge allows for handling not only routine tasks but also critical situations. T
e across various industrial sectors.
man meshes from a monocular camera is intrinsically challenging due to texture and scale ambiguities, especially in real-time a
a complete integrated system. The complete system showcases its ability to interact with several users. Implementing Deep Lea

. In addition, the architecture enables the agent to provide relational descriptions of its decisions and beliefs as on-demand exp
or environments due to the advantage of IMU sensor; while the proposed method will automatically analyze the behaviour of the

been validated through tank experiments with a hovering-type AUV, Tri-TON, and a sea turtle replica, successfully tracking the

robot guides users through the music-based task, in which they can sing/dance/tap along. The system is optimized for persons
ure using autonomous robotic catheters. The project's core concept involves two robots: one for handling the clip delivery syste

obot compliance. We achieve this by optimizing similarity in the parameter space while maintaining human-likeness. Our metho
ground and maintaining a bent knee, the foot naturally falls back down while the other foot acts as a supporting pivot, resulting
ctions. Furthermore, we implement and evaluate a prototype system within SnT's LunaLab, an environment that replicates luna
The shared controller features seamless switching between EMG-based control and automated grasping control by detection o
ects motion on a specified warning and detection areas. The dynamic safety configuration of the scanner is parsed in Unity and
ze of obstacles it can traverse, and it can bend its limb to place grippers around obstacles or above ledges. It also has wheels f
fields. By doing so, we aim to bring these robotic systems to life and demonstrate their performance in real-world scenarios wit
as the potential to make the collection of deep-sea creatures and cultural relics safer and more efficient. In the second part of th
switch between rolling and crawling, and execute precise clockwise and counter-clockwise turns. Remarkably, it survives a 20
ated many researchers around the world to study even more advanced tip growing devices. The second half, Slip-in Manipulato
We also demonstrated that the regulator can change phase and volume as normal under conditions of 126Mpa, and remained u
and elbow flexion/extension part adopt the method of coupled tendon-driven transmission by changing the winding method of te
ipants were instructed to lift a load equivalent to 15% of their maximal voluntary contracted force, in collaboration with the robo

e light source is approximately localised, which is widely applicable in robotics applications such as endoscopy. Our second tec
ress the problem of tracking ambulatory mobile robots from a single camera. The method combines differentiable rendering wit

l dependencies, and ease of deployment on robotic platforms. We test the proposed pipeline on a novel MOT benchmark speci
s. Based on the above two ideas, we implement a tightly coupled estimator that dynamically adjusts the number of UWB anchor
to enhance translation estimation via 3-DoF bundle adjustment. Moreover, we introduce a novel approach for determining the

k from 70% to 73%. Along with better accuracy, through targeted analysis, we show that XGX is also more efficient at goal-con
tdoor environments covering different terrains, outperforming previous state-of-the-art methods that had a higher mean averag
on success rate, respectively.
d semantic keypoints and depth information to estimate poses of the furniture. Tracking is applied to achieve smooth estimation
del and states. The model-based RL facet produces steering angles and quantifies inherent uncertainty. At the same time, the n
ectly on coarse maps using the predicted local occupancy grids as observations. Using the latest belief, CMN extracts a global
challenges, we introduce ‘Bicode,’ an indoor visual marker designed for event cameras. Bicode seamlessly integrates 2
mid length at inference time and make it more robust to changes in input size. Also, it enables our model to offer multiple trade
orms previous methods in all evaluation metrics under both seen and unseen scenarios on two existing benchmark datasets, To
poral consistency of subgoals. Our experimental results verify the design and the new method obtains a 76% higher success ra
f a continuous knowledge graph architecture and multimodal feature alignment empowers the navigator with a remarkable zero
ding end-to-end RL techniques and prior map-based strategies. To provide a comprehensive evaluation, we introduce new perf
assistance robots. Experiments demonstrate the robustness of the PixNav which achieves 80+% success rate in the local path-
performance across moderate to severe domain shifts, and is competitive with fully supervised test-time adaptation approache
semantic costmap, which enables the planner to distinguish between the traversability of different terrains and accurately iden
oying the same navigation policy to survey two distinct Objects of Interest (OOIs): oyster and rock reefs. We compared our me
mework for extrapolating operator terrain preferences for visual navigation. PATERN learns to map inertial-proprioceptive-tactil
rior-guided reconstruction block based on the parallel attention mechanism is present to adaptively correct global image with sta
ct images collected with the cameras of a robot arm, a mobile robot, and a surgical robot, which not only demonstrates the cros
verages sequential frames to construct comprehensive 2D rendering for each object. To our best knowledge, RenderOcc is the
hod outperforms previous methods and is more efficient in computation and memory consumption than cost-volume-based meth
upervised correspondence learning baselines. We also apply Doduo to articulation estimation and zero-shot goal-conditioned m
ch challenge—a large-scale robotic manipulation dataset about real-world warehouse scenarios. RoboLLM not only outperform
he effectiveness of our proposed designs.
ated camera can collect sequences of training image frames and camera poses, and infer accurate 3D depths of the environme
ware timestamp inference strategy to accurately recover event timestamps from event voxels in a continuous fashion and elimin
cts in one forward pass. It learns to aggregate both object embeddings and object parameters over multiple time steps using cr
mation models, opening new possibilities for cross-modal applications in computer vision and robotics research.
an open-set surgical activity benchmark utilizing the public JIGSAWS dataset. Besides, we also collect a novel dataset on endo
 losses during the synthetic domain training, followed by the incorporation of 2D supervised and 3D self-supervised losses on r
we use a spatio-temporal nearest-neighbor search strategy for robust keypoint tracking. Extensive experiments are conducted
bility of the proposed method to generate highly controllable and visually consistent videos and its capability of modeling dynam
ensure that the fused images are salient both in a holistic sense and at a local scale. One discriminator encourages the fused im
rties from Lidar boxes. We call our method “you only label once”, which means labeling on the point cloud once and aut

es from a BEV layout.Our approach comprises two main components: the Neural View Transformation and the Street Image G
n novel view synthesis quality but also achieves a five-fold increase in training speed and a ten-fold improvement in rendering s
ometries, along with the groundtruth of images and depth maps for all camera viewpoints, which significantly facilitates other re
mance through the strategic fusion of camera and 2D LiDAR data. Our approach employs a novel self-supervised network alon
benchmark datasets showcase the robustness and potential of our method in addressing this challenge. We believe our work l
er supervise the proposed lightweight pose network. Extensive experiments on the camera re-localization task on the Cambridg
he state decoder reconstructs raw image pixels and other states from the intermediate representations of the model. Experime
We evaluate our approach in various real-world applications, showcasing its ability to learn fine-grained bimanual manipulation t
unar environments.
onstruction. With this contribution, we also developed the View Planning Toolbox (VPT), a lightweight package for training and
zed for both encoding and decoding the information content. This algorithm possesses the ability to recover corrupted message
or each agent based on its detected boxes and uses a graph neural network to identify common subgraphs between agents, lea
image adhesive waterdrop removal methods.

construted by NeRF. Moreover, current NeRF-based methods naively optimizing pixel loss fail at occluded images. Thus, we fu
d ones. Finally, to enhance interaction within the selected groups, we apply an interpart association module, considering the fu
m a teacher network. MAL is a novel, plug-and-play module designed for seamless integration into multi-frame self-supervised m
ell-detailed data patches. A decoder component, enhanced with a normalized module, effectively utilizes these high-quality feat
on which reasonably augments the dataset while maintaining the geometric relationship between the FV images and the BEV s
ti-modal 3D auto-labeling. We conduct comprehensive evaluations on both public and in-house real-world datasets, which dem
st generalizability. To achieve this, we introduce a lightweight SAM-Event-Adapter (SE-Adapter) module, which incorporates ev

fically, we design a Coarse-to-Fine Knowledge Distillation (CFKD) approach to facilitate comprehensive cross-modality knowled
ar dataset, focusing on enhancing the density and quality of short-range radar data using the long-range radar data as supervis
metry camera position for the dynamic layout design based on digital twin. The optimisation framework follows an effective coa
e original X-ray image with the resulting segmentation maps as input for the regression module to perform precise Cobb angle m
with different imaging simulations and with/without masked license plates, leading to a total of 259,200 images) from 162 differe
ur decoder is adopted to implicitly constrain the keypoints prediction in the former stage, which improves the accuracy of the ke
modulate the location of a standing wave field. Our approach is simple, flexible, and controllable. We envision that the soft aco
ect which are random in shape, dimension, pose, and material. Video of the real robotic experiments is available at https://www
enerates annotations for both segmentation and pose estimation. Our experiments, conducted on synthetic datasets, highlight

opic images, RIDE requires no manual labeling of training data. We test RIDE in the context of surgical tissue tracking on the Su
al procedures. This paper proposes to address these two problems with a multimodal large language model (LLM) and an adap
rgeries was established. Experimental results underscore the efficacy of this approach, with segmentation accuracies exceedin
 sping practical objects in 3D space using only 2D image features for feedback. Our second platform, the Atlas humanoid robot
void damaging the object, the following information is extracted from camera images and observed: the centroid of both the han
t extensive simulation and real-world experimental results that demonstrate the efficacy of our approach.
trol signal from the traditional velocity-based controller. To achieve this, we redefine the camera orientation using quaternions,

by a 3D camera in each frame. Overall, our approach is more broadly applicable than existing approaches. We validate the ef
echanism enables adaptive fusion between prompts and between prompts and the base model. Extensive experiments show s
nd outdoor environments, we demonstrate our system's effectiveness and adaptability. In conclusion, our research represents a
small-sized, blurred, and occluded objects. Then, an innovative Motion-Driven Association (MDA) scheme is introduced, conside
in the presence of prolonged and complete occlusion. Project webpage: https://opfilter.github.io/
or static image detection and segmentation: we detect open-vocabulary object instances and propagate their boxes from frame
he individual characteristics of the template intact. To enable the network to easily preserve the information learned at different
n the training signal. Focusing on obtaining a "clean" training signal from real-world unlabelled video, we propose to separate la
evaluate our SWTrack implementation on the NuScenes autonomous driving dataset to demonstrate improved tracking perform
aintyTrack, a collection of extensions that can be applied to multiple TBD trackers to account for localization uncertainty estimat
rials (cables, papers, rubbers, etc.). The same experiments are made using two different humanoid robots of the same family.
peed. As an example, our fastest variant, LiteTrack-B4, achieves 65.2% AO on the GOT-10k benchmark, surpassing all preced
performance over long distances in experiments on the Rosario dataset, verifying its effectiveness in reducing cumulative posit
ferent terrain types and wheels. In experiments, we demonstrate that ST-VIO can not only adapt to wheel or ground changes a
cale drift or tracking failure. We present a dynamic inverse distance filter that can adjust the convergence range to update cand
rame transformation till converges, rather than collecting measurements for batch optimization. We also perform online calibrat
ditional optimizer to achieve higher global consistency in depth. We conducted experiments on both simulated and real-world d

achieve superior performance compared with the state-of-the-art image-based or event-based VIO. Furthermore, we demonstra

nts on public datasets. Our MAVIS won the first place in all the vision-IMU tracks (single and multi-session SLAM) on Hilti SLAM
s. Meanwhile, we utilized the positions of the track lines and obstacles as constraints to guide people with impaired vision safel
hat the grounding point of the VGFL accurately tracked the forward shift of the CoG during the STS assistance. Moreover, expe

nerate torque. All electronic and pneumatic components are integrated into the system, and the exoskeleton weighs 3.9 kg with

g the training process to further shape the trajectory inside a moving safe region. The interactive data is fed back into the ProM
s stress on the lumbar spine, and mitigates back muscle stress. This pneumatic back exoskeleton offers a promising solution to
he results confirm the effectiveness of the leaf-inspired FSR array, giving the possibilities for portable and cost-effective motion
the HELIE acquired from knowledgeable wearers. In addition, based on the proposed LIET framework, we establish the mathe
is study introduces a knee-joint exoskeleton capable of conforming to the natural movement of the knee in both the Sagittal and

e the step obstacles.

nd contact state and velocity estimation.
 applied this framework to dynamically balance an underactuated 2 degree-of-freedom leg on parabolic terrain and achieve pres
essions of the dynamics equations, including the kinematic loops. To track the trajectories, we propose a simple LQR-based con
rformed for uneven terrain negotiation and obstacle climbing. Results demonstrate the effectiveness of CPG control in transfor
methods, together upper and lower bounding the exact slip prediction. The proposed PTCM classifies the planned trajectory as s
odel is developed for planar motions of the two-wheeled robot, and kinematics of the three modes are special cases with particu
ering angle of the robot are regarded as control inputs, which are designed by making the derivative of a positive definite Lyapu
method is demonstrated to dramatically enlarge the region of horizontal velocities that can be dealt with by a naive approach th
prediction model. Our approach has been validated in simulation and hardware experiments, highlighting the necessity of whol
nce of these policies on a set of benchmark contexts in simulation, and provide baseline validation on a physical robot.

xperiments on the humanoid TORO.

n both simulations and with 7-axis Franka Emika robots, comparing against baseline methods, and demonstrating its effectiven

rehensive human experiments were conducted and the results concluded that bilateral feedback enhances the whole-body tele
 gular and arbitrary-shaped scene text spotting benchmarks in terms of both accuracy and model efficiency. The dataset, code
hich benefits the decision boundary learning in target domain. We conduct experiments on PointDA-10 and PointSegDA for 3D
Multi-Object Tracking (MOT) method, and the results showed that our module, shortly named CARPE-ID, accurately tracked ea
eved by global similarity search. SpectralGV is deterministic, robust to outlier correspondences, and can be computed in paralle
he strength of the pre-trained VL model and the SG comprehension. We establish a new state-of-the-art in several zero-shot ac
sults demonstrate that LPS-Net achieves state-of-the-art performance at the task of point cloud-based place recognition while m
nsive experiments on four UAV benchmarks and one generic benchmark have indicated the superiority of our trackers compare
g geometric correspondences and are used for estimating the object's 6D pose with RANSAC-based Ptextit{n}P. This approach
information. The method employs spatial interest clustering to convert input data into tokens, improving the preservation of spa
se output torque performance of VTSEA are verified by experiments.

educed control difficulty, and linear stiffness profile. Two experiments of step response and stiffness tracking have proved the hi
ol signal obtained by trajectory optimization. Here, we compare the performance of the prototype with a comparable SEA imple

ry, the proposed actuator is powerful and energy-efficient.

ergetically conservative despite having a customizable force-deflection behavior. We present the theory of these springs and de
f the motion imposed by limits on the velocities and accelerations of the robots, or raising the tether's clearance. Simulated and
study considers different types of UAVs; for the fixed-wing UAV, the minimum turning radius constraints are taken into accoun
ed planner can address the trade-off between energy efficiency and perception quality under heteroscedastic uncertainty. The
ated multi-rotors that allow decoupling translational and attitude dynamics.
upervision strategy. Our approach demonstrates an improvement in mean absolute error up to 30% compared to the pre-traine
aerobatic sequence for three tailsitter aircraft.
rodynamic drag, 2) agile flight of a quadrotor subject to wind gust, and 3) high-precision end-effector control of a manipulator su

sing power (210 μs execution time) with a frame rate of 15 fps. The presented autonomous nano-size drone reaches 100% re
on spaces, enabling the seamless transfer of brush manipulation techniques from simulation to practical robotic sketching. Furt

existing hardware for a new course. In this paper, we describe the design and evolution of the MBot platform, and keys to succ

HMAC, we develop two techniques: 1) Hierarchical Iterative Learning, which jointly trains representations to caption the various
hing module that uses the PPM-generated prior information to dynamically adjust the sampling range and refine target coordina
ks collaboratively with distillation to enable lightweight inference to deploy the neural network on edge devices such as NVIDIA
mework is evaluated with four new evaluation metrics. Experimental results show that ViLaIn can generate syntactically correct
efficient training in under 5 min using only 100 data points. It outperforms a state-of-the-art neural estimator by up to 68.1% in
-level AUC score of 96.6%. In other experiments collected from semi-manufactured inkjet printing OLED panels, MCDEN demo
matic structure of different objects, we design a unified kinematic knowledge parser, which represents various articulated objec
es an initial plan, which is then assessed and refined in the iterative self-refinement step by a validator. We examine the perform
cal dataset and operating conditions. Our approach demonstrates superior efficiency alongside comparable accuracy outcomes
tensive experiments on both simulated and real-world datasets. The results show superior performance compared to existing m
ttps://github.com/KTH-RPL/deflow.

ual commands from LLMs. Furthermore, we design general reward functions that consider the distinct motion features of human
mance on unseen terrains, highlighting the adaptability offered by reward learning.
d hardware for a visual servoing task, and show that our method indeed issues an alert before a failure occurs.
ction sequence, given a set of candidate action sequences, with substantial acceleration. On top of that, we introduce an adapt
to use, flexible, and with formal guarantees, our work demonstrates how conformal prediction can be an effective tool for senso
eater degrees of freedom, particle filters tailored for high-dimensional state spaces will be indispensable.
rsity and naturalness. This framework promotes reliable and secure deployment of deep learning models, fostering interdiscipli
b-modules in the system, and from the fast capability to frequently co-adapt the interaction.
ork consists of collecting a few real food slicing data, calibrating our dual simulation environment (a high-fidelity cutting simulato
ering the shape information from the initial scene and object semantics. Finally, for execution, the initial and envisioned goal sc
D task scenarios for increasing numbers of agents. We find that a hybrid framework achieves better task success rates across a
, we demonstrate that robotic manipulator imitation policies trained with our enhanced instructions outperform those relying sole
translating these accomplishments to the real system remains challenging, often accompanied by a performance decline. This
ories, capturing actions ranging from spontaneous impulses to tasks requiring deliberate contemplation. Our results, both in sim
ombine different LLM prompting styles to write code for new manipulation skills. Firstly, we evaluate different prompting styles a
ound robot.
ment learning, which can learn from experience to decide appropriate replanning timings in the given environment and planning
ugh experiments conducted in both simulated and real-world environments.

ask descriptions, temporal decompositions, and corresponding python reward functions for reinforcement learning. We show G
complex long-horizon tasks.
anguage with visual perception and achieves significantly greater robustness on out-of-distribution situations.
orithm to jointly solve the task and motion plan. To improve translation, we automatically detect and correct both syntactic and s
ronments, ASC achieves near-perfect performance (59/60 episodes, or 98%), while sequentially executing skills succeeds in o
ncy. The efficacy of forgetting techniques is evaluated by comparing their performance in a task where two ARMAR robots sear
ompatibility with the environment map. We validate the proposed approach by conducting several collaborative planning and ex

efficiency and dynamic consistency of the proposed method are verified through both simulations and experiments, including c

s to its base. The feasibility and practicability of this method are demonstrated by real flight experiments.
ed LiDAR point clouds into the camera image due to the viewpoint variation of both sensors. We release about 26K images and

ent levels of the occupancy map. Moreover, our dynamic module ensures that the planner dynamically updates roadmap inform

luding a 67-dim humanoid robot and a 96-dim 3-fingered manipulator. It first projects the dynamics of the controlled system into
he closed-loop system is achieved under a partial IE condition that is much weaker than IE. Experiments on a collaborative rob

Multi-Layer Perceptron, which regresses the control commands for the ego-vehicle, namely throttle and steering angle. We exte

cenarios that are well-aligned with human preferences through comprehensive evaluations on the nuScenes dataset.
aneously enables natural language (NL) to temporal constraints encoding, safety violation reasoning and explaining, and unsafe
ving decision-making. The proposed method is validated through extensive experiments conducted on the CARLA simulator. R

ormance on the Argoverse motion forecasting benchmark.

amework in achieving homogeneity of high knowledge across the entire group compared to a heterogeneous system without th
HBA with the consensus-based bundle algorithm, the distributed recursive Hungarian-based algorithms, and the cluster-based H

or performance of LGX in detecting and navigating to visually unique objects.

etherlands. Additionally, the framework was employed on data from an analog mission in the Tabernas desert. Through the fram
ulation in different practical scenarios.
ults show that with DyHGDAT, the CVAE-based prediction system outperforms state-of-the-art methods by 12.5%/5.3% in ADE
nchmark, datasets, and model available for further exploration.
nteractions among heterogeneous traffic agents (e.g., vehicles, pedestrians, and cyclists). Besides, We integrate a scene conte
rements handling non-linearities, an Unscented Kalman Filter is used. The developed framework is evaluated with reference da
gles needed to actualize those trajectories. We train our TrajNet on a comprehensive video dataset incorporating visible and the
or generalization capabilities in long-tail cases. Our models and benchmarks are publicly available. Project website https://jche
comparable to the traditional HD maps.
e common observability assumptions in both domains and their performance impact. Our approach outperforms existing method
vailable KAIST Urban dataset and in scenarios recorded by ourselves. The experimental results demonstrate that our method is
de measures. These objectives respectively optimise probabilistic accuracy and the ability to capture distinct behaviours, and th
risk tolerance levels. Owing to the more unified combination of behavior and motion planning layers, our framework achieves e
imulations on the well-known NGSIM dataset and experiments on two typical lane-changing scenarios on the autonomous drivi
e future research on wheeled mobility.
arge number of social situations. We, therefore, ask if we can rethink the social robot navigation problem by leveraging the adva

V-Bench, a benchmark encompassing 1000 human-annotated BEV scenarios, with more than 20,000 questions and ground-tru
combines the temporal relationships between waypoints with the spatial features extracted from sensor inputs. We evaluate ou
ed cost function searches for the candidates with the least novelty that can lead the robot to the familiar places. We collect offlin

, we aggregate them into one target point via the attention mechanism. IPF enables us to learn continuous representation of 3D

entary nature of 3D skeletal poses in reducing prediction error in such challenging scenarios.
mental architecture and optimization objectives to facilitate the proposed uncertainty-aware reinforcement learning (UnaRL) algo
nes at both training and new scenarios. Additional visualizations and experiments demonstrate the interpretability and transfera
bility classification with the UniTraT. Compared to UniTraT, the accuracy of the MultiTCNN distilled via UMTraDistill only decrea
an be estimated in real-time. The experiments are conducted in a typical scenario that requires adaptive modeling: Autonomous
on), and directly generates navigation actions by implementing the reinforcement learning with a novel reward function based o
over existing methods in complex environments with odometry drift.

nsisting of well-synchronized and calibrated high quality LiDAR point clouds captured across a range of urban and suburban ge

deos can be found on our website: https://xingyu-lin.github.io/spawnnet/.

12 unique skills, and demonstrate its generalization over 38 tasks spread across common daily activities in diverse kitchen scen

ion language models (VLMs) to build a robust reward classifier for autonomously providing reward signals during the online fine
aderboards. From these ensembles, we train distilled student models which have high performance at a fraction of the compute
hat developers often make unforced and paradigmatic physical unit mismatches through differential drives, small angle approx
s smaller than 1. This is significantly different from conventional rotary-wing and vertical take-off and landing (VTOL) UAVs. Thi
hich shows the rationality of the structural design and the feasibility of its application in underwater robots.
eater than the wingspan at a constant height. Results show a very good agreement between the theoretical and experimental tu
y, one can easily study the principles of different swimming motions and design the multi-motions controller for a robotic fish thr
se. A simple model successfully captured the observed robot performance, and informed adaptation strategies that increased r
cost, thrust, and detailed flow fields, all of which is extremely challenging for real animals. We mutually verified our platform wit
g methods in straight-line swimming experiments. The experiment of S-curve swimming also demonstrates the effectiveness o
of tunable stiffness into the total cost of transport compared to previous robots for the first time. The cost of maintaining medium
dden environmental changes. Several comparison studies demonstrated that the proposed approach can significantly improve t
ust gaits? To answer these questions, we train our perceptive locomotion policies in simulation and perform sim-to-real transfer

pright orientation from all starting faces in the passive righting test and righted itself and continued walking in all falling trials. Th
cture is capable of diverse and omnidirectional jumps including forward, lateral, and twist (turning) jumps, even on uneven terrai
ther than by directly contributing to lateral acceleration.
in the future.
precision and stability. Extensive validation through physical platform experiments highlights the effectiveness and feasibility of
exploration, and monitoring.

ping-wing MAV with insect-scale weight.

spectively. Meanwhile, our robot has a high degree of stability, which allows it to maintain a proper aerial orientation without flip
in multi-modal microrobots.
In addition, the robot can carry an additional 419 mg, enabling future sensing, control, and power autonomous operation.
y-optimized propulsors, and achieves turning rates as high as 0.28 rad · s−1, when tracking path references. The measured
mages contain sufficient texture and illumination. Moreover, we show that the trajectory followed during the initial learning flight
lustrated through both theoretical method and experimental method. Moreover, the feasibility of the design is further verified thr
d speed, thrust quantification, and airflow visualization, directed design refinements and validated performance. Through structu
 l the motion state of the mobile robot in real-time. The sensory signals of the robot are encoded and transmitted to the sensory
11.1 percent, even when the robot's length was varied between the training and test data. The unrestricted extension capability
e robot. Experimental results show that the proposed multimodal robot can realize the function of a quadrotor, respond quickly w
erimental results highlight the emergence of mechanical intelligence when the robot is equipped with an appropriate level of bod
e a reaction wheel and generate pitch rotation; however, the robot is underactuated on the yaw rotation. To achieve its real-tim
we developed a mechanism to couple different modules corresponding to three different environments (on a flat surface, on low

of eye movements: visual fixation, saccadic pursuit, and smooth pursuit. The experimental results are presented which succes
y providing precise camera poses for locating and interacting with objects. The code is available at the project page: https://oot

n, applicable to both single-pair and multi-pair cases. Extensive experiments conducted across indoor and outdoor environment
TO monitoring reacts faster on abrupt decalibrations. The tracking variants of both methods follow random calibration drift with

owing to the efficient feature extraction from both lidar and camera sensors. This novel, efficient, and accurate method can prov
ity and accuracy of our method. The codes are open-sourced on the Github website. The code link is https://github.com/OpenC
the initial value from the first stage. Extensive experiments on solid-state LiDAR, conventional spinning LiDAR and KITTI data
ompensate for its distance to the camera. This makes the extrinsic calibration time-consuming and cumbersome. To overcome
data with varying degrees of error, LCCRAFT demonstrates adaptability to diverse mis- calibration scenarios. This includes cas
rical simulations and experimental testing, respectively. The obtained results demonstrate a high level of agreement with the gr
ch was successfully evaluated on a set of scenes with a wide range of object interactions, and we showed that our method can
on robustness. Extensive experiments on the KITTI-odometry dataset demonstrate the superior performance of our HIFMNet c
n optical tracker's rigid body marker attached to the TMS coil or femoroplasty drill guide. Previous research documents low repe
the weights used in least squares, which indicate the relative importance for each joint. Our framework requires only encoder m
elation module is utilized to identify correlations between features more accurately. Lastly, we employ transformer architectures
edges extracted from events and point clouds and correlates them to estimate extrinsic parameters. Experimental results demo
phere. The benefit is the target center is a viewpoint-invariant feature. Thus, it is ideal for large viewpoint-difference calibration.
ng terrains.
l quadruped robot ANYmal running in real-time (
ss weighted by path length (SPL) for the Object Goal Navigation task. Furthermore, we show that VLFM's zero-shot nature ena
ransformer- based policy trained on data from multiple ground robots, with a diffusion model decoder to flexibly handle both goa
he significant potential of the proposed geometric regularity in enhancing robot learning capabilities.
d more sparse.
ce online. We demonstrate our approach on a mobile robot navigating around pedestrians.

g and normalize the geometry of real-world obstacles to match that of simulated objects, thereby bridging the sim-to-real gap. O
enerates a more realistic and continuous path. Moreover, this approach considers collision avoidance and mitigation in an integ

iency of our method. The robot can travel through the unknown uneven region safely and reach the target rapidly without a pre
which requires the path to predicate the interactions of the robot w.r.t. the dynamic environment in multiple time steps. This, in
etization and a planning-friendly zonotope uncertainty representation; the method also enables differentiating through a zonoto

outperforming smoothness, comparable path length, and safe distance from obstacles.
obot in low survivability environments (e.g., tight corridors), but also allows it to move away from obstacles safely when needed
-world narrow, dense environments using a real Turtlebot and Boston Dynamics Spot. We observe significant increases in succ
our proposed approach through robust simulation testing, highlighting its effectiveness in achieving safe and efficient navigatio

tate-of-the-art baseline methods in terms of accuracy and generalizability. We also show that the proposed rewiring improves t
advantages of JOB, we compare its performance against traditional LOS checks using a widely recognized public dataset. The
in a collision avoidance gym environment and show that our approach minimizes the number of collisions without making overl
erated by a local planner to avoid potential collisions and failures. Our algorithm demonstrates robust predictions when one or m

rounded VLM in an interactive framework with a large language model-based robotic planner, and show improved planning per
t our PRV-based view planning method outperforms baselines, achieving good reconstruction quality while significantly reducin
fined by implicit representations and set covering optimization problems. Simulated experiments show that our method achieve
d showcase the benefits of input-dependent multimodal query initialization, while being more efficient than the available alternat
view and multi-modal data of articulated objects spanning diverse categories. Evaluations demonstrate that UM3 surpasses exi
r MF-MOS achieved a leading IoU of 76.7% on the MOS leaderboard of the SemanticKITTI dataset upon submission, demons
ability of reducing the interference from prior estimation errors and achieving high-quality scene reconstruction with rich geomet
ormal and density values in the grid cells instead of latent features. Our modification allows direct access to the geometric value

grounding. We demonstrate the tractability of amodal optical flow in extensive experiments and show its utility for downstream

chieving at most 0.1094m/0.089m/0.091m in EPE3D, 54.23%/53.67%/74.1% in AS, 32.75%/21.87%/40.25% in AR, and 70.981
emantic consistency are also conducted for supervising the network, which enhances its performance of understanding the sce
p-performing model (ID) sets a new state of the art on DSEC benchmark. Meanwhile, the base model (TID) is competitive with p
ameras for information aggregation based on pose information, instead of interacting with all cameras indiscriminately. Experim
R completion pipeline based on the powerful capabilities of DDPMs. We demonstrate that our method surpasses existing metho
ncurrently handle multiple tasks, notably 3D object detection, while maintaining real-time processing speeds. Our rigorously op
n layers to enhance and integrate the object-level and class-level features. LiDARFormer is evaluated on the large-scale nuSce
dback control can be accomplished to perform interactive manipulation, in which the contact force between the finger and objec
ment error of 1.34°±0.48°. Besides, the proposed method requires no training data and can achieve real-time sensing durin
ace. Experimental validation demonstrates that this approach outperforms an off-the-shelf vision system in estimation accuracy
ased MPC. The proposed approach is validated on two robots, showing the efficacy of sensorless contact estimation and the ef
oarse but fast reconstruction for the contact objects. Two experiments are performed. The first verifies TacShade’s ability t

parameters and decision variables than the Linear Time-Varying (LTV) MPC approach, making it easier to implement. Real-wor
 ing friction as a function of Fn and v, we demonstrate that friction parameters can be estimated using regression models from a
lity and the set of tasks that it can solve. We evaluate LEAGUE on four challenging simulated task domains and show that LEA
domain-similarity between the source and the current target domain. With the synergy of the proposed framework and regulariz
demonstrate our method's efficacy and superior performance over existing methods.
ularization constraint. In order to improve experience utilization efficiency, the policy is then trained using both the distilled data

rs knowledge from the memory to the library, building composable, general and interpretable policies, while bypassing gradient

ut also achieves at least a 50% reduction in model parameters on the robotic tactile CL benchmark.
to prior methods. More experimental videos and results can be found on the project website: https://ut-austin-rpl.github.io/Lotus
hand, the performance of single bi-copter including trajectory tracking and passing through narrow gaps was demonstrated thro
o further illustrate the distinctive merits of our proposed system.
e and is able to synthesize modular robots for industrial tasks in cluttered environments. ​
Bot 1.0 was developed, and several experiments were conducted to evaluate the feasibility of the proposed system focusing on
distance from obstacles and preventing overstretching of the deformable object. The implementation is performed in ROS, allo
The effectiveness of the proposed algorithms is demonstrated in synthetic robot navigation experiments in a realistic simulator.
Vehicle Routing Problem with Time Windows to optimize the UAV’s trajectories and minimize time when UGVs have to wa
emonstrate the efficiency of the proposed design.
ed from a sophisticated setup involving robotic assistance and OCT integrated microscopes. Furthermore, we present a neural n
a baseline deep neural network. Results clearly show the advantage of using our method, with a 15-fold speed up in labeling tim
pper position help improve failure detection and action segmentation accuracy.
st.github.io/CEA-IMSOLD/.
oud completion for real-world navigation. The code and datasets are available at https://github.com/yyxssm/PointSSC.

aims at identifying interaction categories between two given objects in a video clip. Further, we benchmark our dataset for the
over a period of 17 months, and encompasses over 49 kilometers of trajectories, making it the most extensive dataset of RTS-b

iments reveal serious challenges for some robot scene understanding tasks in dynamic scenes. By sharing this dataset, we aim
ow that our conditional model improves over various marginal baselines. We also introduce new techniques to tele-operate a 7-
simulated in closed-loop to account for interactions with other traffic participants. We present a detailed analysis of numerous ba
ll three components. We further introduce our ever-expanding dataset from the ongoing data collection effort -- the TBD Pedes
ments with an intervention mechanism enables performing tasks to completion, making it deployable with human oversight even
ut scene graphs, achieving 89.9% prediction accuracy on our testing dataset. We further validate the generated scenarios on b
collected in the real world. Each sequence in the dataset includes visual, force, audio, and action information. Moreover, we al
n turn requires data that allows us to model the behavior of humans in the presence of robots. Therefore, our approach is based
ks, (2) anticipate risks, and (3) facilitate decision-making. We conduct extensive experiments and summarize future research on

ortant object localization and per-object explanation prediction. Overall, our dataset and introduced prediction models form the f
nding scenarios, captured from hybrid realistic LiDAR simulator. To our knowledge, LiDAR-CS Dataset is the first dataset that a
methods are evaluated on the constructed dataset. The results show that methods based on dense coordinates currently perform
ous robots operating in unstructured environments. This framework also makes it possible to query data for specific weather co
between the gripper and the bin or object are a concern, and distinguishing the background is crucial. The results show that our
chmark evaluation of several state-of-the-art VPR algorithms using our annotated dataset, revealing its challenge and thus valu
-source diameter estimation and off-the-shelf semantic segmentation algorithms and share our baseline results.
erimental results show that the proposed method achieves the better performance qualitatively and quantitatively compared with
eometric curvature properties of primitives and propose a geometric saliency guided dynamic prototype matching and label grap
will be available at https://github.com/wangmohan17/FFLOGO.
er-based architectures in articulated object analysis.
the research field of LiDAR perception in adverse weather, we release the SemanticSpray dataset, which contains labeled vehi

s evaluated only within the specified high-relevance regions, corresponding to the child points of these selected S points. Addit
baseline algorithm on a real-world robotic platform to illustrate its practical utility in straightforward HRI tasks. The demonstratio
of trainable parameters based on data distribution during the continual adaptation process, including domain-specific parameter
sm (denoted as STNet and TemNet, respectively) are trained based on the same dataset. The experiments show the validity of
nd past human joint poses. The surrounding environment context and past human joint poses are two heterogeneous inputs an
ce of 2% of the robot's workspace dimension. We demonstrate the capabilities of this model on motion planning tasks as an exe
strate that our proposed method works effectively on a wide range of open-vocabulary affordances and outperforms other base
lts from this real-world scenario showed its remarkable performance surpassing the baseline methods, highlighting its potential
thms without the need for camera calibration or robot models. Using this approach, we trained a model for realtime keypoint de

ms all baselines.
al network (CNN) by applying the previously collected optical flow distortions to a dataset containing a mixture of real and synth
We compare our FCNN against three State-of-the-Art (SoA) systems. Considering the best-performing SoA approach, our mod
ng (PCAT) mechanism to fully exploit cluster-level pseudo labels. By combining and strengthening both point-level and cluster-le
o improve inference time by more than a factor of 2 while being memory efficient with approximately 12 times fewer parameters
orms existing methods and achieves state-of-the-art performance, signifying a meaningful step forward in indoor depth estimati
ors or self-occlusions. Leveraging the robust modeling potential of diffusion models, we reframe this issue as a conditional 3D k
we propose a novel LiDAR-based 3D lane detection model, LiLaDet, incorporating the spatial geometry learning of the LiDAR
rt PCP methods on the prediction results and semantic similarity, and has a good real-time performance. Our open-source cod
The resulting approach can perform unseen, long-horizon tasks in unexplored environments while instantly reacting to dynamic
simulation, validate it on a real robot, study SOGM predictions on real data in various circumstances, and provide a novel indoo
ning techniques. Furthermore, we show outstanding robustness towards disturbances such as clutter, occlusions, and noisy de
and report results on the Argoverse dataset. Our experiments show that PBP achieves competitive performance on the standar
compliance and task compliance experiment.
olicies. Our implementation is available at https://github.com/daniellawson9999/merging-decision-transf ormers.
n motion datasets.

arison study, while humans achieve 66.9% recognition accuracy, our best approach reaches 90.0% in under 5 touches, highlig
ce of noise. The experiments highlight the substantial influence of the architectural and synaptic-model choices, respectively. O
ucted on two widely-used benchmarks, Sceneflow and KITTI, we perform ablation studies to analyze the effectiveness of each m
Experimental results demonstrate that our incremental method achieves a level of accuracy on par with approaches employing g
roach using the classifier on a 6D object pose estimation task from tactile data.
d performing in real-time. Additionally, our qualitative results showcase the effectiveness of our approach in generating human-r
m. The base layer bitstream contains geometry data, while the enhancement layer bitstream utilizes semantic-guided residual d
ms of the time to find a valid solution and planning success rate, and performs comparably in terms of trajectory length over time
and extra perturbations. Comprehensive experiments on (1) classic control problems (Inverted Pendulum), (2) simulation bench

ing our approach suitable for dynamic environments. We validate our approach on~two~simulated tasks and in a demanding re
cess rate compared to baseline approaches.

that can be characterized by semi- or total transparency, and irregular or curved surfaces. The results of our experiments demo

accurate pose estimation. Experimental evaluations encompass real-world and synthetic datasets, demonstrating competitive
e results compared to baseline approaches and is capable of correctly identifying off-nominal behavior with 91.65% accuracy in
ly, the paper presents suitable candidates for the framework’s neural components, covering essential adjustments for their

antially improve policy learning. We present a novel task scenario designed to evaluate the progress in visuomotor policy learni
distributive method that can accommodate an arbitrary number of robots or human partners without reconfiguration or retraining
ated both in simulation and real-world environments. The results demonstrate that the single-view visual student policy can suc

pulling from visual input. To efficiently collect data for training PickNet and PullNet, we embrace the self-supervised learning par
ideos, code, dataset and trained models are available on the project website: https://sites.google.com/view/foldingbyhand
us grasp control. Real robot experiments show that our method has 74.2% success rate for grasping 7 unseen objects. Compa
n grasping success rate and action efficiency. Finally, the framework's generalization and adaptability are demonstrated by con
ment in enabling shape servoing methods to bring deformable object manipulation closer to practical, real-world applications.
l robot setup. Videos and code are available at: https://sites.google.com/stanford.edu/gengap-icra
dustrial needs.

data, without the need for object detection, pose estimation, or manual design of controllers. We evaluate our approach on pick
ation images as a visual observation space, we successfully transferred the agent to the real world without supplementary train
hance manipulation performance. We train GAMMA with the PartNet-Mobility dataset and evaluate with comprehensive experim
n generate diverse grasps with multiple collision-free grasp orientations. Experimental evaluations in simulation and on a real ro
ur proposed approach on various sparse reward robotic tasks and compare it with the state-of-the-art methods. We show that o
hieve over 90% success rate for generalizable transparent object grasping in both simulation and the real via seamless sim-to-r
d method using RLBench and demonstrate that the task performance is significantly improved by the correction stage.
he estimation of 6D object poses, which can be utilized for manipulation. While, obtaining images from more and diverse viewp
our framework in both simulation and real-world settings.
 osed method shows significant improvements in grasping performance compared to existing methods for both generated datas
ons, our framework can perform high-quality grasp detection in real-time and achieve state-of-the-art results. In addition, real ro
ntly reduced.
s of previously unseen adversarial objects, and a 74.3% mean zero-shot success rate on a class of previously unseen, challeng
posed architecture and provide an application study that highlights the potential of our model for tasks like odometry estimation
se class-aware feature consistency distillation (CFCD) that first leverages the outputs of each student as the pseudo labels and
from the support set. Then, a prototype fusion module is designed to integrate the features from both prototypes. CrackNex ou
distribution that occurs after the softmax operation. The primary focus of this paper is on addressing the performance degrada
super-resolution tasks. Furthermore, we demonstrate that our enhanced radar point cloud is capable of downstream radar poin
es. Interestingly, our method exhibits a language characteristic related to context sensitivity.
nly methods, and approach performance of supervised methods, both with respect to keypoint tracking as well as for a robot g

idly close the sim-to-real gap across a variety of perceptual modalities.

ess rate of 73% using our four-fingered Allegro robot hand. The increase in performance is 108% higher than policies using tac
rious scenarios, significantly increasing the success rate of shape control processes.
d real-world scenarios, outperforming recent state-of-the-art methods. To ensure reproducibility and further the academic disco

obot. Notably, this work represents a pioneering effort to learn the synergy between pushing, grasping, and throwing actions. E
e underlying inverse Jacobian closely, which can be used to perform manipulation tasks across a range of objects successfully
f visual and tactile tokens for visual-tactile representation learning. We utilize our dataset to pre-train the fusion model and emb

ce feedback, is mapped by the grasp matrix to the desired fingertip forces, which are then used for computing torques to be ex

work contributes to enhanced transferability across distinct operational contexts and tasks.
mprehensive understanding of the framework.

stability can be designed to deform the object while online estimating the modal parameters. Simulations, experiments, and co
fective in global planning. Applying our method on challenging contact-rich manipulation tasks, we show that model-based moti
ics and models. It requires minor modifications to accommodate different scenarios and robots. Hence, it provides a flexible an

ision-free and efficient trajectory. We validate our strategy in both simulation and practical scenarios, and experimental results
om consistent distance estimation and collision checking on Cloud and Edge. FogROS2-LS reduces collision failures by up to 8
k will investigate the ability to use tethers and strain for reactive behaviors and more diverse tasks.
quadratic programming methods is also illustrated. Finally, hardware experiments on a multi-robot platform verify the applicabil
select more appropriate parameters and promising initial solutions during local optimization. Compared with manual heuristics o
e control policy, is constructed to improve efficiency by avoiding exhaustive searches and local sensing pitfalls. The effectivenes
We validate LC-ADMM and iDFGO in simulations with examples from multi-agent Simultaneous Localization and Mapping (SLA
ediction and optimization of the robot's state for control purposes. Finally, numerical simulations and experiments of the Point-to
computational burden. In the comparative studies, several scenarios are considered to demonstrate the effectiveness of the pr
directing the sampling process toward more optimal directions. Through numerical experiments, we com- pare our approach w
are also derived. We present case studies to demonstrate the feasibility of our approach.

can be achieved for high dimensional robots like quadrupeds and humanoids using our algorithms as they scale better than the
heme, where a non-smooth subsystem with contact interactions is present. The advantages of constructing RIM using flexible
t its rewards and generate navigational behaviors, even when real observations have not been updated. Accordingly, RIDER e
s during task execution, we can modulate contact forces in the new environment, thereby generating trajectories similar to thos

echnique focuses on refining the precision of the estimations. Experimental results demonstrate the effectiveness of our metho
roject website: asap.csail.mit.edu

on of multiple single-session maps and to build the globally consistent maps in ambiguous and perceptually-degraded environm
patroller can effectively provide protection against multiple, diverse attacker behaviors.
maximum weighted idleness of the graph and is shown to be NP-hard. Then, we describe an exact solution for the problem and
experiments.

ing in surface-to-surface impacts, and allows for tuning of the sensitivity between expected and unexpected post-impact behav
eline method for anomaly detection: It relies on deep-learning-based density estimation with normalizing flows, tailored to the da
on phase failures using the same exteroception setup. We cover a diverse set of failure types for primary tabletop manipulation
even be executed in various dynamic tasks and terrains.

eless losses control through an aggressive yawing manoeuvre. As the metric depends only on system and actuator models, it i
b-uva/SGSM

an the state-of-the-art algorithm in a large-scale simulation environment and demonstrated a 62% improvement in explored vol
e, providing a close-up platform for inspection equipment. The proposed system was tested in an outdoor testing environment a
ontribution, which consists of GPU-accelerated functions to learn GMMs 10-100x faster compared to existing CPU implementat
les are integrated into the autonomous terrain traversal system. Our approach is validated in simulation and real-world scenario
ethod is thoroughly evaluated and compared to state-of-the-art methods in simulation and during real-world experiments, under
h four field trials, including missions in both shallow water and open sea environments. The trials involve picking and placing va
. We demonstrate the suitability of our approach in a series of numerical experiments covering complex source location scenar

w-level local planner to avoid dynamic obstacles and navigate to the target point. Besides, our framework contains a novel dyn
nce different loads. The result is a device that provides solar panel users with an accessible, affordable, and convenient means
e can automatically detect leaks and anomalous angular geometry with a low failure rate compared to human labeling.
of energy consumption using four different manipulators, namely, Universal Robot's UR5e, UR10e, Franka Emika's FR3, and K
ditions. Secondly, the robotic arm can transport the UAV inside the hub, perform maintenance tasks like battery replacements,
potentially high reward locations, while ensuring that the robot can travel to and sample the chosen locations within a user spe
 ts by leveraging Gaussian Process model estimates of future trends based on collected remote sensing and point measuremen
te the vertical translation of the calibration. Challenges arise due to non-static parameters while the robot moves, indicated by s

eness of its surroundings, enabling it to predict the traversability of terrains that are not immediately visible. This enhanced awa
roposed sensor across three distinct tasks: i) grating identification, ii) pose regression, iii) contact localization and force estimat
nsor performs well in detecting the position and force of contact in a large area. The sensor's performance shows promise in em
rmore our system is capable of identifying textures at human-level performance in as little as 84ms. Additionally, benchmarking
chnique, we demonstrate its ability to provide accurate torque readings over two different tactile sensors and three object geom
show that with significantly more training samples, a model-free policy can outperform a model-based planner, generating sho

force sensing modes can operate independently, but the entire L3 F-TOUCH remains Light-weight and Low-cost while facilitati
ustness when performing grasping tasks. The integration of vision-based tactile sensors can significantly enhance the capabiliti
mented in previous camera-based tactile sensors. With these advancements, we make the integration of tactile sensors more a
mics prediction. Through experiments and ablation studies, we demonstrate the effectiveness of our shared tactile latent space
an fabricate the open-source design and have a ready-to-use state estimation model. A set of experiments demonstrates the a
proximately 100,000 image-force pairs from 175 complex objects and train a neural network to regress the 6D force, which can
es. Comparative experimental results show that the proposed sensor can effectively detect cracks and has the potential to ass
ormance for both tactile perception and 3D vision. Our results show that the proposed sensor can reconstruct a 3D view of a sc

d images of markers' movements during the interaction with silicone specimens with different compliance at different indenting
challenges met during the sim-to-real application, e.g. the learned policy tended to squeeze an object in the bi-reorienting task

ate with the actual softness of the terrain. Furthermore, these attributes played an important role in achieving high accuracy in t
s state-of-the-art generalizable 6D pose trackers in synthetic and real-world scenarios. Our code and dataset are available at ht

mpared to commercial load cells, our sensor achieves a small measurement RMSE of 0.41 N and a percentage error of 4.15%.

for newly fabricated sensors. Consequently, the resulted sim-to-real generator could be built on top of the tactile simulator to p
, which eliminates pre-calibration requirements and enables the derivation of surface normals from a single image without relyi
in the proposed GAN allows the network to integrate the history of deformations from the depth image sequences to generate r
regrasping, insertion, and tool use. Please see https://sites.google.com/view/texterity for videos of real-world demonstrations.

o move towards the corner, while the tactile cost defined by garment edge poses prevents the garment from falling from the grip
existing state-of-the-art self-supervised and supervised techniques. In evaluating our methodology, we focus on two distinct tas

he hand-eye relationship. Our algorithm requires neither external calibration sensors/objects nor large-scale data acquisition pr
 e identified motions, show the effect of degenerate motion on state estimation, and demonstrate the generalizability of our ana
with the deformable object. Several experiments are conducted to verify the effectiveness of the proposed framework. The exp
rmance is comparable to that of conventional force sensors. Because the prototype has a difference in resolution between the a
in contact location with computational time less than 1 ms. The system has been evaluated on a phantom with sinusoidal moti

ntal results demonstrate that the proposed scheme outperforms the state-of-art methods.
manipulation disassembly skills, e.g., levering and unscrew-driving regarding their object requirements. Supported by realistic

orms the state of the art.

d dataset, which highlights the advantage of our approach compared to traditional approaches in terms of expressiveness. By i
ied. This paper provides a theoretical elucidation of various misconceptions surrounding this approach and offers experimental
iciently solving the problem. The case studies highlight the efficiency of the proposed approach
ions and success probabilities) that consider agents' capabilities to perform tasks in the environment. The proposed iterative m
s to satisfy key safety constraints, expressed as temporal logic specifications, with desired probability. We demonstrate our app
n, but the time-dependent distribution over possible delays is known. Finally, we develop a receding horizon planner for Markov
Markovian ones, but also improves the sampling efficiency by simultaneous learning of multiple sub-tasks. An environment-agno
ction. We introduce the semantics of this abstraction and show how existing tools can be utilized to synthesize strategies to ach
ously generate an LTL formula to realize the local temporary task. A joint sampling algorithm is then developed, in which the aut
almost quadratically with the number of automaton states. Extensive simulation and experiment demonstrate the validity, the e
planning level of a quadruped mobile manipulator to demonstrate its zero-shot transfer ability for fetch-and-deliver and navigatio
results confirmed that the predicted values agreed well with the measured values in the L-shaped surface painting test, validat

a illustrating how plan quality (in terms of cost to execute) can be improved through an optimization operation that is robot agno

Position Exploration (TPE) module are added to solve the problem of grasping transparent objects in irregular and visually unde
he geometry and kinematics of the fingers, including their number of degrees of freedom, allows to greatly alter the intrinsic stif
n during training. Simulation results show that our method acquires better spatial precision and grasp accuracy compared to th
g domain. Finally, we deploy our method on a Panda robot to grasp heavy objects.
tion performs satisfactorily in real-world scenarios for unseen objects.
and reliability for generated grasps.
grasping robots. Source code can be found at https://github.com/Naif-W-Alharthi/Physical-and-Digital-Attacks-on-Grasping-Netw
We perform experiments over a recent benchmark with over 190K+ unique objects, extend the dataset with the additional refe
mpling Aggregation method effectively integrates global and local information, relieving imbalanced distribution of features. Fina
T Skin technology for a real-time gripping force control application.
atters on which researchers on grasping should focus in the future. A QD approach has finally been proposed for making grasp
imitives allows for a straightforward adaptation of grasp characteristics while preserving the essential grasping motion learned
w that it can enable grasping of slow moving objects, such as those encountered during human to robot handover.

object in the scene. This representation is further used for object reconstruction and grasp detection in cluttered table-top scene
aset includes ranges of grasps conducted across four objects and a variety of orientations. Manipulator states, object pose, vide
 nces for training, MoMo learns a latent-space dynamics model and a failure classifier that combined enable MoMo to simulate f
hat our proposed representation outperforms the baselines in terms of both training speed and generalization. For the generaliz
rcement learning (MBRL), demo-bootstrapping, and effective exploration, MoDem-V2 can acquire contact-rich dexterous manip
on, enabling smoother grasping execution. Experimental validation confirms the effectiveness of our approach in generating fea

can generate precise and determinable strategies for human-like object grasping. Experimental results confirm the efficiency of
lly driven jumping – using the Smart Suction Cup – is especially useful for unprocessed waste PCBs with large surface mo
monstrates our approach’s consistent ability to predict and execute grasping postures based on the completed object shap
o repeat the data collection process. We demonstrate that differences between the algorithms, which would not be apparent fro

er performance is also evaluated through the grasping tests.

s demonstrate the proposed gripper’s merits. We expect this new gripper to provide a more competitive solution for robots t
en the fingers and the grasped object was used as a measure of the grippers’ adaptability. The test results demonstrate tha

nfirmed the design guidelines.

ending angles. Comparative analyses demonstrate that our design exhibits superior stiffness and enables five-mode grasping i
netic fields, and demonstrated its potential for future applications in MIS.

sassembling a model of an electric motor using a robot manipulator with a multitool end effector.

od ingredients within a pick-and-place setup to evaluate both the grasping and the releasing capability of the gripper. Obtained
l constraints are an effective representation for co-designing diverse manipulation systems. Our methodology scales well with i

nd simulated food items with varying geometry, weight, and firmness.

terials required to recreate the setup can be found on https://sites.google.com/view/dmanus.

and. Our model also demonstrates efficient salient feature learning capabilities on ALU-Textures, which is further augmented b
stimate the length of human's phalanxes, by using the fingertip's circular motions. With knowing the aforementioned information
ensory perception in prosthetic applications and offer valuable insights for future prosthetic design.
with five actuators in a sparse array. The vibration patterns from all actuators are superimposed linearly to synthetically generat
sfer of the acquired skills to an unassisted surgical scenario, like the clinical one.
ectal cancer screening task. Users correctly identify the position and development stages of cancerous tissues in 25 out of 27 t

ic sensor, exploring "tactile scenes" composed of shapes, textures, and objects. Each learned representation provides a perce
al range for shapes we could stably simulate. We then demonstrate potential scalability by simulating a multi- segment version o
ent results on real-world scenes show that VIDAR efficiently delivers high-quality visual guidance and outperforms the non-inter
nature of human gait and is modified with a force sensor for increased reactiveness to human gait variations. This approach is e
ntegration of humans and robots in unstructured environments, this research contributes to the broader exploration of robotics i
at the proposed self-supervised regression approach overcame traditional unsupervised methods, even considering different ro
 -robot coefficiency score as reward to adapt and learn online the combination of robot interaction parameters that maximises su
perators. After their evaluation, the data is categorized into False Negatives(FN) and True Negatives(TN). FN are then subcate
rst instance of using the variance of the return computed in an RL framework as a guidance measure for a Human-in-the-Loop
me necessary to observe performance gains and mitigate damage. Results indicate that task completions noticeably improved
e object Cartesian pose needs to be optimized. We compare the policy generated with our approach with a random, a cyclic an
oach to recent state-of-the-art preference-based RL semi-supervised learning techniques. Our experimental findings reveal tha
stimate of uncertainty is learned via evaluating the feature extraction capabilities of the robot's visual navigation policy. By incor
spectively, and are more robust in rare vehicle-pedestrian scenarios.
ntify human preference uncertainty and fuse it with the reinforcement learning framework. We evaluate our framework using tw
st suitable view angle is employed to further alleviate the operation difficulty. Simulations and real world experiments are implem
ing load-compensation and load-releasing are carried out on eight human subjects with 0.5 and 1 kg loads to evaluate the relea
ct response in humanrobot co-packing. The results show that our method enables the robot to remain compliance in the face o
bot. The results highlight the potential of virtual workspace restrictions in human-robot collaborative disassembly tasks.

Y, and RGB-D pedestrian datasets. Our study demonstrates that FlowMNO outperforms some of the state-of-the-art deep learn
plementing a reward function to learn an effective response to pedestrian unpredictability. Compared to the original SARL policy
sive and inclusive. We also find that people's self assessments of risk and time-urgency do not correlate with their path prefere
human behaviors.
c and diverse social environments with pedestrians modeled via MAC-ID. We have trained and compared various navigation me
ges MPC with learned prediction models in a computationally scalable manner to provide us with the best of both worlds. In pa
igning edges with asymmetric Gaussian model, a trajectory evaluation strategy is designed to reflect motion tendency and moti
t the cost of a slightly lower success rate. We demonstrate the capabilities of the methods through concrete examples and qua

hieving an accuracy of 74.99% in slice-wise cross-validation and 62.00% in trial-wise cross-validation. This outperforms baselin
we achieve 97.2% accuracy on a standard multi-user dataset with 108 surface classes, outperforming the state-of-the-art mach
eference profiles.

sable objects such as curtains and grasses for verification by instructing the robot to traverse them. Besides, traversing curtains
urorobots. The proposed 3D Dilated Efficient CapsNet model trains on an augmented input space to computationally `force' the

beliefs to guide the robot policy learning, which aims to balance between avoiding human trust collapse due to robot reverse ps

ons on a tablet while running on a treadmill at 6.5 km/h and 7.5 km/h. The results show subjects could recognize the cued direc

nments featuring multiple specialized cutting boards. Videos and supplementary materials are available at https://emprise.cs.co
tion performance compared to the alternatives, capable of closely predicting positions during fast motion along non-constant cu
sed of nodes and edges is constructed and the actions at each node are stored. This makes it possible to estimate the path and
n operator-constrained policy. Our results, simulated based on input from five professional rescuers, display effective task ment
on of the real-time application of eye gaze in civil infrastructure inspections.

n usability, expected increase of productivity, as well as high trust and low anxiety, but both work cells scored low on expectanc
effect of the direction of pHRI interaction on the EoP of the wrist, with increased energetic passivity in the abduction-adduction d
tion of the region of interest. We evaluate our method against different baseline algorithms in simulation and also demonstrate i
d and 5% decrease in percent stance compared to walking with a passive exoskeleton. This controller also decreases the overa

stacle avoidance component constructed using behavioral dynamics is integrated into the impedance controller. Experimental r

allenge, we present an algorithm for solving the optimal policy synthesis problem. First, we enhance the belief MDP (derived by
ctive controller in a collaborative task with real users to show that it can improve users' interactions with a robot collaborator qu
m performance.

ns resulting from problematic objects such as transparent or reflective surfaces, the fused map is overlayed onto the environme
during the network training process increases the accuracy and generality of motion classification compared to using only real

nt scenarios: one composed by a mobile manipulator, i.e. a UR10e mounted on a Neobotix MPO-500, and one composed by o
train an intelligent agent that can skillfully collaborate with the human operator during the surgical task. The proposed shared co
nces but fail at further distances, while Tonal-Sonemes remain recognizable as the distance to the robot increases.
ramework's efficacy is validated through simulations and an experimental study employing a Franka Emika Research 3 robot.
he proposed method using tasks in an Assistive Gym. We demonstrate that policies trained with a popular deep RL method are
he task, the intelligent agent helps to reduce surgery time by 31.7% and the human input by 60.5% compared to direct teleope
nd low-level skill primitives. We demonstrate DROC effectively distills the relevant information from the sequence of online corre

fferent experimental conditions. This is followed by a comparative statistical analysis. The experimental results show that the le
Compared to direct control and simple shared control, the proposed shared control scheme exhibits superior tracking accuracy
of our method and the results confirm the advantages of the FIC approach over traditional impedance controllers, showcasing it
is evaluated for a number of goals and agent conditions, and the results show that correct and executable plans are found in m
ollaborate more efficiently with new robot partners.

ferent robot tasks for different HRC modes in a unified manner. The unified task control set enables the robot to switch smooth
h validations confirm the effectiveness of the proposed planner, which returns shorter completion times while ensuring safe inte
discussions with domain experts can effectively facilitate navigation by perceiving instruction-relevant information, correcting ina
capable of manifesting perceivable and distinguishable personality traits.
trajectories, in order to reduce the entropy of the credit assignment. We show the effectiveness of our work in both simulation a
ith robots to enable co-painting in the physical world. Our open-source approach, CoFRIDA, creates paintings and drawings th
g material property through visual cues and offers a promising step towards improving the functionality of wearable robots for PV
eal-time through a impedance tuning unit. To validate the framework we present a set of experiments both in laboratory settings
earn a VMP library from sequentially-provided motion capture data.
ound that without assistance, astronauts would impulsively exert themselves to perform a post-fall recovery, which resulted in h
onfiguration to uncover target body parts without exposing the rest of the body. Source code and supplementary materials are a
1s48). Participants viewed videos of a robot performing the same tasks at the same level of competence, with task order differin
minar and turbulent flow components and uses predominantly the laminar components to make flow predictions. Based on the h
arning and memory for efficient and real-time learning of new knowledge from humans. We integrate our architecture with a phy
th low-cost hardware, user-friendly interface, and interesting creation experience. Experiments show that our robot can stably c
training task, and found a significant improvement in the classification accuracies of the participants who trained using the VR-2

g paintings from 2D images, our framework has the advantage of rendering stylized ink paintings from 3D models by using a co
a user study and real-world deployments. We make the mixed-reality application publicly available at github.com/cvg/hololens_
successful on-site walking acquisition with a compact nine-dimension dynamics model, 40Hz real-time planning, and on-site lea

omotion patterns, including standing, push-recovery, squat walking, human-like straight-leg walking, and dynamic running. Nota
rios, the NC-MPC is proposed to regenerate the reactive motion online to adapt it to the real local environment. Finally, we valid
mulation. We also verify that the method takes advantage of the intrinsic characteristics of its neuromorphic encoding scheme to
s study hold significant implications for enhancing the safety and reliability of bipedal robotic systems.
versely drives the bending deformation of patches, which can be controlled by layer-jamming bars. We design the three-layered
the existing state-of-the-art approach based on the recursive Newton-Euler algorithm. Results demonstrate that our method ac
cooperative perception datasets OPV2V and V2V4Real demonstrate that the proposed S2R-ViT can effectively bridge the gap
method maximizes the effective information utilization of each modality and leverages inter-modal complementarity. Our metho
nchrony. We also introduce a Feature Compression (FC) module with channel and spatial compression blocks for transmission
dings, and daytime conditions.

collected dataset annotated with nine road defect classes. The experiments show that our proposed method outperforms exist
and dynamic scenarios. It uses a graph-attention Transformer to extract domain-specific features for each agent, coupled with
nsuming 3D convolution network is replaced with a novel residual-like architecture, where features are mainly digested by a ligh
y the network. Experimental results on the Waymo Open Dataset demonstrate the state-of-the-art performance, and improvem
mploy a distillation module that transfers temporal information and richer knowledge from a larger image backbone to the mono
uential track confidence refinement. We offer the first self-contained framework that addresses all aspects of the 3D MOT probl
ntermediate features to minimize the distribution gap among multi-agent features. Intensive experiments on the public OPV2V a
y sensitive natural discrepancies: appearance-based discrepancy, distribution-based discrepancy, and task-aware discrepancy
closer to real-world deployment. First, we propose a unified evaluation pipeline for forecasting methods with real-world percept
ur work can further facilitate the cross-agent representation interaction for better cooperative perception in practice.
for defogging. Experiments demonstrate that our method can accurately mitigate the impact of fog and meanwhile enhance th
of transmitted data, we incorporate communication-aware regularization which enforces the network to extract complementary
a novel 3D multi-object cooperative tracking algorithm for autonomous driving via a differentiable multi-sensor Kalman Filter. Ou
architectures perform indistinguishably well on in-distribution deployment, resolving the causality gap. In situation 2, We obser
eriments in a simulated kitchen environment demonstrate that H2RIL significantly outperforms other prior baselines and is capa
s motivates our looking ahead approach. Besides, hidden confounding factors may affect the decision making process. To that

"progress" through a task by learning how to temporally rank the video frames in a demonstration. By inferring an appropriate r
with domain transfer, and a natural language processing problem with real datasets demonstrate the superior performance of o
fabrics of different colors and sizes, and its performance can be comparable to human-level operation. Furthermore, the frame
oth on the quality of the control decisions and also on their frequency. The studied optimized models significantly increase infer
h to various complex quasi-static manipulation scenarios. Quantitative results indicate the effectiveness of our approach in prov

he parameters of the spherical mechanism are optimized. The optimization results show that the MAW of the spherical mechan
with random signals from a Deep Random Generator to remove the dependency between the output and the removed features
an appropriate control law to form textit{behavioral dynamics}. Behavioral dynamics unlock a new action space and safe, guidin

tracking accuracy while adapting to dynamic changes. We validate the approach in multiple experiments on a 16 degrees-of-fr
of the proposed climbing robot using active suction cups with an optimized design. Finally, a painting application was presented
he-art approaches. Supplementary materials can be found at https://leizhang-public.github.io/cg-cnn/.

ment Learning (HARL) framework we previously proposed, which can effectively perform cooperative navigation tasks among h
visually pleasing results. Extensive experimental evaluations on various low-light image datasets demonstrate that our propose

sliding on the surface geometry to process a specific area. Alternatively, he can let the controller execute a parametric trajecto

rs such as collision avoidance, comfort, and progress for safety and interpretability. Our approach achieves better highway driv

5% more fuel saving in simulation and the open-road highway testing, respectively.

ensional quadruped robot dynamics and a real-world differential drive robot (TurtleBot3) under different types of task specificatio

ur real-world trajectories when using our proposed techniques. More material can be found on our project website: url{https://sit
uture expected risk. We assess the predictive performance of our risk model through a comparison with real vehicle driving logs

ng experiences.
oving performance in planning tasks through reactive and nonreactive closed-loop simulations on the nuPlan dataset.
sses point clouds into a latent vector by encoding networks that are suitable for processing point clouds. We then design multim
n challenging parts of the task. We evaluate our method on two robot manipulation tasks and find it improves the planning perfo

in simulation and on a real collaborative robot, showing great potential for easier and less error-prone fixture calibration.
s physical measurements with the purpose to identify accurate degradation dynamics. Once the degradation model is identified
d 2022 to benchmark localization, insertion, and disassembly tasks. We present data from 10 DR.J task boards, describe a me
model uncertainty due to its adaptive design. The precision and efficiency of calculating distributed states using the proposed m
are interconnected using the Robot Operating System (ROS), enabling seamless communication and integration. The propose
LR SARA robot. The results show high coverage for complex freeform geometry and that adaptive spacing can optimize the ov

oach is able to ensure safety while allowing the robot to continue its mission during an interrupted connection and thus avoiding
cles. Finally, we introduce grouping vehicles to extend the proposed method for solving extremely dense and large instances at

the same ODD query, the number of final matched scenarios increased by 36 times, diversity increased by 99 times, and scen
ative, reachability-based approach. A key takeaway is the potential value of cooperation in enabling the deployment of autonom
proach in modeling human behavior. The results reveal that our model effectively captures human behavior, showcasing its vers
se benchmark dataset, we show that our method achieves state-of-the-art performance on various metrics, and further investiga
nhanced and the computation burden is considerably alleviated with the proposed innovation development. We also examine th
eal-world driving data. The proposed courtesy measure allows us to auto-label the courtesy levels of trajectories from real-world
epancy between the predicted driving strategy based on partial observation and that based on full observation. This approach c
m the motion domain along with distance information to account for the worst-case threat. Our evaluation results using RetinaNe
performance and reduce the model's variance under different sampling methods. Extensive experiments confirm that our appro
nd MOTPS that address this limitation. STT achieves competitive real-time performance on the Waymo Open Dataset.
er mechanism to filter the detrimental image data reconstructed by adaptive decoders. We evaluate the effectiveness of our pro

ile guaranteeing asymptotic stability when disturbances are zero. Finally, we present comprehensive simulation and experimen
ars. We extensively tested our solution and created the NexusStreet dataset, by means of the CARLA simulator, employing an

RL, as the alternative: 1) images do not require a complete re-imagination of the observation space from environment to enviro
a hybrid setting of letting LLMs help the driving model correct mistakes and complicated scenarios. The results of our experime
efficiently in terms of runtime. We show that joint training delivers significantly better performance than separate training of the
ct global encodings. These encodings along with agents’ local information, are passed through an encoder to obtain time-de
del robustness, including increasing the training dataset size and using targeted data augmentations that randomly drop non-ca

00 traffic scenarios, we demonstrate that it emerges as a promising approach to synergize the strengths of DRL and conventio
h partitions the graph into sectors based on the predicted demand and the maximum number of taxis that can run sequentially g
on, and AV Training. CLIC frames AV Evaluation as a collision prediction task, where it estimates the chance of AV failures in th

oach on Argoverse 1, Argoverse 2, and nuScenes real-world benchmarks, where it all achieved state-of-the-art performance w
ate that our decision module is real-time capable and yields significant speed-ups compared to executing a motion planner stan
pes. By eMem-NDT, we infer each instance in behavior prediction of vehicles by bottom-up Memory Prototype Matching (MPM)
a long horizon. The warning generation problem is formulated as a partially observed Markov decision process (POMDP). An o
fidelity simulation scenarios focusing on autonomous on-ramp merging.

a high success rate of 96%. Finally, our framework is validated in the high-fidelity simulator under dense traffic, demonstrating
ely. Our findings underscore the potential of leveraging human cognition principles to enhance the proficiency and adaptability o
oach and its state-of-the-art performance. In all tested datasets, ContextVAE models are fast to train and provide high-quality m
ectiveness of the CNN-LSTM model in achieving subject-independent estimation, the offline estimation under the backpropagat
will make the code and data supporting our experiments publicly available at https://gitlab.kuleuven.be/detry-lab/public/robot-tra
task. We provide a comprehensive evaluation of the proposed classifier.
rrent two-frames-input based networks. The results show that our method performs better, even in wild images. We fed our ma
g approaches, with an accuracy of 87.5% and an average prediction time of 4.35 seconds before the maneuver takes place. W
of the objects in the scenario.

evaluation on common benchmark datasets and show that our model not only achieves state-of-the-art performance, but also s
Empirical experiments substantiate that the proposed framework robustly predicts intended endpoints of hand motion before at
ontroller or actuation energy. An electro-adhesive locking mechanism attached to the tendon locks the claws within 20 milliseco
on data provided by a motion capture system. The study demonstrates closed-loop controllable robotic manipulators based on a
showed that the VS-SCA provides a range of stiffness from 20 to 75 Nm/rad for dorsiflexion, necessary for providing ankle assi
nts, and other settings where cleanliness of surfaces is important.

endon-constrained synchronous rotation mechanism for the manipulator is carried out, and the linkage performance is greatly im
oretically decrease the stiffness to zero while maintaining the same robot shape. Experimentally, the robot is made 4 times mor

nctional: it streamlines the modeling and automation of kinematics and enables a nuanced evaluation of diverse modeling techn
95 mm respectively. The maximum water jumping height and distance are 241 and 965 mm.
h iteration. In our evaluation, we thoroughly analyze the performance of our method using seven-dimensional reference paths w

osen: (i) Serial robot - Universal Robot- UR5 (6-dof); Rethink Robotics- Sawyer Robot (7-dof) and (ii) Parallel manipulator - Cab
st cases of common manufacturing tasks are presented yielding optimal designs and reduced computational effort by up to 92%
ures and a more consistent segmentation compared to conventional representations. We believe that other existing segmentati
-kinematic scenario with gravity loading. We show that this method is at least 50 times faster than the ABAQUS implementation
d. The system is tested for different scenarios in simulation showing the repeatability of the procedure and in real scenarios wit

e the Genetic Algorithm is implemented to search for the time optimal trajectory. The simulation results indicate that the optimal
ss-validating them by using a lumped model, for which the stability can be assessed by analyzing the spectrum of a reduced H

ed based on the metric to optimize the total MP traveling time. This study shows the potential of robot wrench-twist capability a
d software implementation are publicly available on GitHub at https://github.com/dfki-ric-underactuated-lab/ricmonk and the vide

periments using small quadrotors. Simulation results show that the proposed algorithm does reduce the conservativeness of the
ystems in various scenarios, where our approach outperforms the conventional CBF both qualitatively and quantitatively.
ents show that the proposed method achieves the given goal while ensuring safety by preventing failures. We also show that wi
ions to specific locations in a semantic map of the environment. Our method leverages the fact that LLMs, trained on a vast cor
te drops significantly under disturbance, the proposed model enhances robustness against disturbances to modalities it does n
the current sensor information. We apply this approach to the whole-body tool-use on KXR, a low-rigidity plastic-made humano
o different tasks, which is particularly challenging. We show that the priors lead to better distribution estimates, particularly unde
 mika Panda Robot Arms and validated its effectiveness.
it using datasets and a real robot. In the former case, motions are 2 and 4 dimensional, modeled as first- and second-order dyn
s and obstacles. After movement learning, EditMP can controllably and interpretably edit the learned movements for new task s
positions, height distributions, and even configurations of the objects/goals. Thanks to the generalization, only a single image of

ic nature can facilitate repetitive learning. The study employs periodic Dynamic Movement Primitives and Locally Weight Regre
tions for unseen sponges, outperforming the LfD method without pre-training. We validate the motion generated by our semi-su
he start and goal bidirectionally. The collision risk of the intermediate connections is verified based on sum-of-squares optimiza
ion to find actions that ensure future states lie within the learned safe set. We apply our approach to different simulated control
e labels and generate goal-reaching plans in complex Sokoban styled domains that scale with number of objects. RLAP furnish
thod is evaluated on a collaborative robot with a clutch in series for each actuator. Real-world experiments with payloads from

elliptical excision skill and apply a Bayesian optimisation scheme to optimise the excision behaviour with respect to expert rating

nd robust policy across the various stages of the task, enhancing the success rates in task completion even with the presence o

regions in 3D Cartesian space and fits a probabilistic model of motion trajectories to these regions. New motion trajectories, wh

get domain that is distinct from training. Our experiments in both simulation and real-world settings show that PromptAdapt is a

reinforced by empirical validation through a robotic experiment.

proposed approach is validated on the LASA data set and two representative robotic experiments.
monstrate the robustness of TCL on two real-world robotic tasks.
ion experiments revealed that the ERTC-HD could identify valid motion in cluttered environments within shorter time periods tha
gmenting state representations with intent embeddings generated by an IQL-TD-MPC manager significantly improves off-the-sh
g possible interference occurring among rewards which hinders convergence to useful skills. Furthermore, in the context of table
ns to a level similar to offline optimal planning methods, while the average time to complete a cutting task is within 25% of the o

ration (RLTSC). RLTSC incorporates recent advancements in contrastive unsupervised learning, endowing it with potent repres

en supervises a student world model that takes the domain-randomised image observations as input. By distilling the learned la
nticipate in order to intercept.
ose a novel training objective with a weighting function, based on ODT, to improve sample efficiency. Experimental results on m
cooperative behavior between agents. A high-level imagination policy is trained to provide additional rewards to guide the low-

uadrotor. Project website: https://sites.google.com/usc.edu/hyperppo

 the learned policy is robust against local motor malfunctions and can be transferred to new tasks.
pervised learning, multi-modal perception, inverse reinforcement learning, and representation learning. The dataset is available
ze a lightweight design. Furthermore, to ensure balanced training for both tasks, we present a multi-stage training strategy aim
its remarkable effectiveness.
method outperforms traditional local mapping and inpainting baselines by up to 40%, and demonstrate its efficacy on multiple leg
deliver more accurate localization results. We conduct extensive experiments on the KITTI dataset to verify the robustness and
bots navigating an alpine environment. Comparative evaluations against state-of-the-art traversability estimation methods demo
e real world demonstrate that our method can improve the success rate by 28% and the trajectory smoothness by 27% compar
heterogeneous behaviors. We carried out extensive comparison and ablation studies to illustrate the effectiveness of our appro

an) to solve temporally extended tasks. In REPlan, a Disentangled Representation Module (DRM) is proposed to learn compac
tiveness of our model in accurately performing forward prediction and learning plannable object-centric representations for dow
d (iv) maintain robustness against increased levels of uncertainty and horizons.
related latent representations, enhancing the pursuit of the optimal policy. Simulations and real robot experiments demonstrate
onstrate that our method performs well in real-world experiments. Supplementary videos can be found at https://sites.google.co
vailable at url{https://maltemosbach.github.io/grasp_anything}.

patory action, especially with targets that seek to evade the robot following a simple policy. Experiments show that our method
sly unknown objects. Our method significantly enhances task success rate by up to 15% in manipulating diverse objects compa
hms in terms of learning efficiency and effectiveness. Depending on experimental results, our method has more efficiency, high
nsitions. Our evaluations highlight the proficiency of TEST on the in-distribution furniture assemblies, its adaptability to unseen f

ach outperforms existing methods in terms of both performance and planning efficiency.
on tasks involving deformable and articulated objects that go beyond typical rigid object manipulation tasks that prior work cons
aselines on standard DRL benchmarks (Cartpole, Pendulum), Meta-World tasks (door manipulation) and a vision-based grasp
also show that MultiAC6 has a 66% higher success rate than a single-agent approach. Further experimental studies demonstr
creasing complexity of long-term place recognition. We demonstrate BioSLAM in three incremental SLAM scenarios: 1) a 120k
dy the effects of different candidate position layouts and demonstrate that our layout design facilitates higher performances in lo

rain-adaptive locomotion of snake robots.

h the LiDAR map. We extensively evaluate our approach on multiple real-world driving datasets and show that RaLF achieves
repeatedly optimizing keyframes that were initially under-optimized through a compensation strategy. The consistency of the lo
ard shaping. The key intuition is that wireless environments vary, but physics laws persist. After learning to utilize the physics in

https://raaslab.org/projects/MIM4Robots.

reducing potential collisions without additional time or travel.

The results indicate that the proposed approach can compete with the modular setup while yielding greater training efficiency. T
ments demonstrate that FSPO outperforms existing methods in terms of safety guarantees and task completion rate.
selves. In particular, we investigate two approaches to incorporate symmetry invariance into DRL – data augmentation and m
between the input states during the whole task execution process to help find the most reasonable object-to-arm corresponden
fline-to-online safe RL framework. GOLD distills an offline DT policy into a lightweight policy network through guided online safe

ation efficiency and generalization for multi-agent exploration tasks. MANTM mainly comprises a Topological Mapper and a nov
manner by utilizing the future footholds of the robot to estimate the support surface based on a Gaussian process. Our method
gest that our policy learns to use the arm as a “tail” to initiate robot turning at high speeds and to stabilize the quadruped
also introduce unique control challenges.
bot dynamics.
he real world, where we achieve an average transition success rate of 97.22% (N=360) in our experiments.
/sites.google.com/view/bipedal-motions-quadruped.
rupedal robot that can take diverse behaviors in the real world under natural language commands.

xpected scenarios where a single leg experiences concurrent joint faults in two joints. Furthermore, the policy retains the robot's
kills ranging from basic locomotion to complex interactive navigation, including navigating around obstacles, pushing objects, c
y-efficient, and safe interactions with the environment.
can efficiently produce more symmetric, periodic, and energy-efficient gaits compared to Vanilla RL with a minimal burden of re
g foot reflection. Moreover, we intentionally designed the learning module as a general plugin for different robot platforms. We tr
We show our robot can perform a high jump on obstacles 2x its height, long jump across gaps 2x its length, do a handstand and
need for additional reward function tuning to achieve risk sensitivity. We show emergent risk sensitive locomotion behavior in s
the resulting policy exhibits improved robustness in harsh and uncertain situations in legged locomotion. Videos are available a
eight adjustments for both objectives. Our experiments evaluate EIPO on various quadruped locomotion tasks, revealing that p
vigation goals, we introduce a hierarchical locomotion controller that combines a classical planner tasked with planning waypoin
sks such as locomotion on stepping stones, where the robot must select from a set of disjointed footholds to maneuver succes

ur baseline walking controller over three steady-state walking speeds. Lastly, we demonstrate some of the benefits of utilizing r
c form. The feasibility of the generated reference trajectories is demonstrated by extensive whole-body simulations with the hum
o unseen scenarios.
ms. The dataset and source code will be available at https://github.com/Mr-Zqr/LIKO.

ntroller for spinal flexion, designed to enhance balance without altering the leg locomotion. The effectiveness of our proposed c
approach by transferring the motions of cats, dogs, and humans to example robots including a quadruped (on hardware) and a
thors, this is the first time a mechanism like the C-Claw is proposed for improving quadruped robot locomotion over high obstac
y environments.
g from 18 cm to 100 cm. Our policy modulates a representation of the Central Pattern Generator (CPG) in the spinal cord, effec

ces at the foot, disturbances at the base, and slippery contact surfaces. Videos of the experiments are available at https://sites.
ficiency increase compared to reinforcement learning methods such as PPO. Transitioning to real-world testing, our policy achi
sparity between the model and the real-world robot, thereby enhancing the locomotion performance of quadruped robots. We v
mulations and experiments for the robust and cooperative locomotion of two holonomically constrained A1 robots with different p
allenges for the simplified model, thanks to RL’s flexibility. Additionally, our framework effortlessly supports the training of c
e linearized model is discussed.

gh terrain utilizing the generated terrain map.

is incorporated in an iLQR framework to optimize a trajectory while taking into account the closed-loop convergence of the traje
e quadrupedal robot successfully performed acceleration from a static state to various velocities and demonstrated the ability to
mpared to regular experience accumulation. Finally, our planning framework is validated in both dynamics simulations and real h
s and expand the use of quadruped robots in various applications.
observation within the controller, enabling the activation or deactivation of a stair-climbing mode. We present our results on dif
mpact on walking. These findings contribute to the design of spine structures in quadruped robots.

divided in training and validation are accessible at www.rvp-group.net/datasets/slam.

ed, achieving state-of-the-art performance on the KITTI benchmark.
work shows how VDNAs can be used for VPR by learning a very lightweight and simple encoder to generate task-specific desc

botics research.
ive in appearance changes but also competitive in opposing viewpoints.
er feature fusion. Moreover, MBFusion is simple, and can be plug-and-played into exist- ing pipelines. We evaluate MBFusion o
erform maximum-a-posterior (MAP) estimation. Thanks to these quantizers, the proposed QVIO estimators significantly reduce
phasize our ability to construct a comprehensive map. Our implementation is open-sourced and available at: https://github.com/
se graph based representation, only updating the origin of neural volumes is required without remapping. Finally, these function
sets including Oxford RobotCar and Multi- Spectral Stereo, demonstrate the robust improvements realized by our approach.
ox.ac.uk/labs/drs/nerf-mapping/
-art algorithms. The proposed method achieves an average accuracy improvement of 5cm during the registration process. The
KITTI dataset and author-collected dataset with highly dynamic objects. Compared with other state-of-the-art methods, our meth
t: https://github.com/XiaoHan-Git/RO-MAP
uilds maps up to 41.2 times faster than OctoMap and 9.3 times faster than the recent SuperRay CPU implementation. Moreove
substantiate the superior accuracy and generalization capabilities of our proposed method. Codes, models, and datasets will b
onstrate the compatibility of resulting costmap predictions with a model-predictive controller. Finally, we evaluate our approach
ving, also known as a footprint. This enables the prediction of traversability that reflects the characteristics of the robot platform.
ne graph prediction models on the nuScenes dataset. We propose our SSL method as a scalable and interpretable continual le
e models in terms of recognition accuracy and model parameter efficiency. Moreover, our ablation studies show that the propo

ca and Matterport3D) demonstrate that our method achieves better generalization performance when zero-shot transferred to s
hich interpolates regions missed by InstanceMap and resolves ambiguities. Additionally, we introduce InstanceLoc, which enab
cenarios. Project page:https://sites.google.com/view/salient-ground- factor/home.
aintains the accuracy and efficiency of pose tracking. Comparison experiments are performed on publicly available large-scale
results show that our algorithm improves accuracy and robustness with real-time performance for robot localization tasks. Spec
o our knowledge, our method is the first to register images onto the point cloud map without requiring synchronous capture of c
additionally publish a new dataset, that captures five real-world environments in challenging, geometrically degenerate scenes.
npoint sensor localization without any prior information.
ng the map boundary or into unmapped regions.
n all baselines tested and running faster than all but one. Moreover, the proposed method assumes no a priori knowledge of wh

Based on that, we develop a SLAM system that capitalizes on satellite imagery for global localization. This strategy effectively
videos and achieves state-of-the-art performance. Moreover, the proposed method achieves competitive performance on KITT

bal position information, improving the robustness of global representations to environmental changes. Experimental results dem
he state-of-the-art systems. Besides, the LiDAR-IMU extrinsic and time-delay parameters are estimated effectively, and the on

orld datasets, which are collected from various large-scale crowded indoor environments. The experimental results show that, i
nts. We conduct extensive experiments on various indoor and outdoor datasets and show that our method achieves state-of-the
that enables precise image retrieval. Through extensive experiments on the standard benchmark tests, we demonstrate that th
Based on the three arts of uncertainties, a novel view selection policy is formed that significantly improves data efficiency. Expe
ng a lighter architecture and allowing to process sequences of various lengths. The code is available at https://github.com/ga1i1
data. We thoroughly evaluate our approach on Oxford, MulRan, and PNV datasets using three different LPR methods. The exp
uated in hardware using a quadruped robot on various terrains, yielding a 65% improvement on the Root Mean Squared Error c
ments in performance gains. Code is available at https://github.com/csiro-robotics/P-GAT.
thin the same colonoscopy or within different colonoscopies of the same patient. Code will be available upon acceptance.
estimation with mean errors below 2 deg at 10 Hz in a workspace of 80 x 80 x 60 mm.

https://github.com/IRMVLab/LHMap-loc.
ed on matching LiDAR scans to the neural distance field.
s the performance of various baselines in different scenarios.
n terms of vehicle localization.

sensor frame rate.

and Cambridge Landmarks datasets while maintaining the SOTA accuracy in single-image APRs.
frame and the other in global coordinate frame. Given a single image, it estimates these two 3D scene representations, which a
d determines the most reliable pose. Our method has been validated in two real indoor parking lots, achieving new state-of-the
g Point Cloud} (QPC). Then the QPC is compressed by the same GPC, and is aggregated into a global descriptor by an attentio
hat each output spiking neuron responds to just a single place. We evaluate our system on the Nordland and Oxford RobotCar

gher performance than existing approaches. We further obtain a 6% improvement in performance by characterizing the semanti
 lization, at 15 Hz, on a resource- constrained Jetson AGX Orin embedded platform.

d office environment, achieving 90% success rate. We also release an open-source implementation of our work.
main aim of this dataset is to enable research on sensor fusion with the most commonly used sensors for localization tasks. M

ts showed that the average drift error of the nonlinear filter was 0.16 m (s.d. 0.06 m) compared to 0.15 m (s.d. 0.05 m) for the s
er-level tracking accuracy with minimally-tuned open-source controllers.
nts. We augment the filter with a simple innovation-covariance inflation strategy that seamlessly handles GNSS outliers and shif
ent localization with an average absolute accuracy of ~0.3m in real-world settings. Furthermore, our results indicate that we can
ayesian fusion technique for optimal fusion. Extensive evaluations on multiple real-world sites indicate a significant performance
er iterations compared to baseline methods, exhibiting its efficiency.
likelihood function spreads. Experimental results demonstrate the efficacy of our proposed method, particularly when applied to
d and then employ further scoring and weight functions for localization. Given a broad initial guess from WiFi and barometer loc
ow the trained TCN can predict the spin priors with RMSE of 5.27 Hz. Integrating TCN into the factor graph reduces the predicti
need to mount and calibrate an extra sensor on the vehicle.
experiments, the position drift can be as low as 0.9% over 1.8 and 4.5 km trajectories. That allows accurate estimation of 6-DO
demonstrate the capability of Outram in a variety of scenarios in multiple large-scale outdoor datasets. Our implementation is o
arkably, our approach achieves a computational speed that is 4-10 times faster than the existing methods. Through comprehen

ot. The experimental results show our method's outstanding data-driven ability and highlight its superior arrival rate and genera
g highdimensional problems for real-time applications. The derived controller ensures a rigorous lower bound on the probability
onstrate its capability in real-world scenarios.
proposed architecture can achieve real-time control performance, learn complex behaviors via trial and error, and retain the pr
ich is pervasive in conventional planners. Our approach also outperforms the previous state-of-the-art in CARLA on challengin
simulator. Empirical validations on both simulated and real-world object manipulation setups clearly show that our method can m

monstrate dynamics learning and tracking control on a real nonholonomic wheeled robot.
. We evaluate DMPO on a real quadrotor agile trajectory tracking task, on which it improves performance over a baseline MPC
y models the system's dynamics. Online, we use a Risk-Aware Model Predictive Path Integral controller to safely control the rob
ollers for a few common robots, and a number of challenging example tasks. We provide this library as a resource for the comm
y on OOD data relative to baselines.
efits of pre-training on video data with robotic offline RL approaches that train on diverse robot data, resulting in value functions
rover facilitate effective manipulation of the underlying granular surface. We used Bayesian Optimization (BO) to gain insight in
valuates upper bound of safety in each state as awareness of environmental safety for the agent. During online training, a beha
ource code and supplementary materials will be available at https://github.com/arc-l/remp.
ce and then uses gradients of learned feasibility-score models for implicit iterative pose-refinement. The proposed method is ev
d ignoring force goals dropped the success rate from 90% to 45%, demonstrating that force goals can significantly enhance per
tive experiments and ablation studies have been conducted to evaluate the effectiveness and robustness of the proposed appr
ks to show that using the signature is faster and succeeds more often than methods that rely on local geometry or finite-horizon
 o manipulate. Then, we move the camera in the front of the approximated place, scan a less noisy point cloud containing preci
cular). The improved version of E-M4M presented in this paper, I-M4M, is a faster and more versatile MAMO solver with a rich

ach requires only 10%–50% as much memory, while achieving competitive quality
ciency, and adaptability across diverse scales and scenes over advanced dense mapping methods using range sensors. Our sy
our method achieves state-of-the-art results in camera relocalization under varying lighting conditions. Code and data will be m

Furthermore, the techniques described in our work hold promise for addressing various challenging tasks or aspects through su
demonstrate that our method outperforms existing NeRF-based mapping methods in geometry accuracy, texture realism, and t
ect the true errors of our object maps, can be useful for downstream robotics tasks such as active vision. We perform extensive
ent-level descriptors through neighborhood aggregation using graph convolution layers, which improves robot localization based
a quadrupedal robot. Videos and more information is available at: https://scene-action-maps.github.io
are utilized. Compared to the grid-form particle-based map, our map enables continuous occupancy estimation and substantial
sensors. We demonstrate that our volumetric belief fusion can increase the precision and recall of moving object segmentation
als in high curvature areas. We evaluate our approach on various data, both self-recorded and publicly available, acquired usin
maps. In experiments, we show that our method can overcome the severe visual distortion of omnidirectional images and grea
ussian Process Occupancy Map (GPOM) and Bayesian Generalized Kernel Inference (BGKOctomap), our method achieves a h
ry efficiency. We demonstrate two real-world applications of D-Map for real-time occupancy mapping using a high-resolution LiD
erified on SemanticKITTI, our proposed framework yields promising performance on online DOR of map construction compared
ition, we present the benefit of jointly using RGB-D and LiDAR within our unified method. We finally release an open-source CU

is essential for underwater robots to perform tasks in intricate environments.

deractuated UVMS without pitch control would rotate during manipulation tasks. To validate the overall design and passive pitch
l full mission cycle tests proved the vehicle’s ability in semi-autonomous operation and robust domain crossing and verified
neous yaw changes of 15â—¦ in 0.4 s, a minimum turn radius of 0.85 m, and maximum pitch and roll rates of 3.5 and 1 rd/s, re
.4 Hz, showing a 30.8% increase. Finally, multimodal swimming of the robot is demonstrated by configuring various locomotive
w of the complete procedure. This work represents an important step towards the use of legged robots in marine operations, an
enerate the needed pitch and roll moments. Finally, we propose a terrain-adaptive locomotion control for the robot, which uses
re, path safety is considered for the entire transition between consecutive states, utilizing a novel approach for continuous spati
a diver operated camera rig, conducting a survey of a coral reef. The method achieves real-time performance, running at 24 FP
ologies for 6D object pose estimation, our pipeline provides estimates that are 8% more accurate. We also demonstrate the rea
cularly for feature-based methods. Our method demonstrates significantly better performance in generating correspondences fo
fication in this domain. The proposed model not only surpasses peers in segmentation metrics but also matches ensemble mod
t effects into a volume rendering framework with end-to-end differentiability. Experimental results on both synthetic and real-wo
ntation for visual servoing by AUVs inside underwater caves. The proposed model and benchmark dataset open up promising o

ng up to 8.2% Bpp savings on point cloud benchmark datasets with different lasers.

ubmerged or interacting with the free surface.

and low-risk. Compared with the adaptive robust controller, this article provides an ingenious and enlightening strategy from the
r control but is still highly maneuverable from additional buoyancy control devices. ReefGlider bridges the gap between the cap
with the traditional MPC method, our proposed MPC-CBF strategy reduces tracking errors by 17.82% and 40.26% in simulation
er provides sufficient technical details and completeness, and through a series of experiments validated that the SHOALBOT ha
deg). We validated our models with different in-water calibration sequences, lenses, and housing setups, as well as with comp
an AUV-based MBES dataset. To facilitate future research, both the code and data are made available online.

sit the dock based on the imposed constraint. The effects of increasing the number of AUVs, increasing the number of points o
g minimal travel time and energy. A variant of our framework that automatically adapts its risk sensitivity is also demonstrated to
zon planning strategy is implemented, managing a goal-reaching and uncertainty-reduction tradeoff. Simulation experiments in
foilborne mode, and submarine mode. We elaborate on the design principles and methodologies of Sea-U-Foil first, then valida
ch for more efficient autonomous missions involving side scan sonar, such as underwater exploration, marine archaeology, and
ed to test its trajectory-tracking performance in fast-flowing water. The experimental results show that the manipulator achieves
secutive loop closure, and a proximity based loop closure, which reduces the overall drift. The framework is validated in three d

ve water clearance errors were found to be between −2.3% and 3%.

range of 5 meters in clear water with minimal false positives.

to perform loop closures and reset the localization drift. The underwater caves present challenges to VIO due to the absence o

by rendering the problem tractable. We substantiate the efficacy of our method through various simulation and experimental stu

re presented to validate the performance of the proposed estimation method, effectively estimating the CoM position and show
ight performance, or integration of tactile sensing with F/T sensor feedback, leading to around 16% improvement in position tra
ping module with high elastic energy. We build the dynamic model to analyze the aerodynamic effect during the jumping phase
e perching site immediately after unperching. We validate the proposed framework in experiments where the tiltrotor successfu
ble-suspended load to closely mimic real-world systems. Comparisons between simulations and real-world experiments with dif

of a coaxial rotors UAV. Hence, a power consumption managing pipeline is proposed to extend the power tolerance of the UAV
eractuated aerial vehicles and sagging-prone cables, and through hardware experiments.
sons of perching and disentangling maneuvers demonstrate energy savings that could be further maximized with the use of the
ine. Subsequently, an optimization controller is formulated to follow the reference trajectory by incorporating multiple practical c
he MAV, the measured external forces, and the predicted drag forces.

d plane, are investigated and shown to have strong dependencies on geometric parameters including thruster-to-thruster spacin
CCC path, which is a candidate path for the classical Markov-Dubins problem, is not optimal for the weighted Markov-Dubins p
tching the attitude and tangential velocity. Extensive tests are conducted to perch on static and moving surfaces at various incli
hod as being applicable for trajectory tracking for Monocopters such as the SICARO which has an extended range of flying capa
directions. In particular, we derive two essential conditions to be met, such that the robot can passively align with flat work sur
ectives for future aerial platform development, particularly in terms of simulation-based approaches.

d edge in real-world testing, highlighting end-to-end reinforcement learning's potential. The performance drop observed from sim

on time, suggesting that the proposed co-design method could be a useful addition to the aircraft engineering toolbox.
otor, the quadrotor compliantly responds by executing a $180^circ$ flip reverse flight to handle such drastic disturbances. Num
we present data collected during one hundred real-time high-speed yaw-tracking flight experiments. These results highlight the s

execute physical experiments, ensuring safe operations and preventing platform damage.
anos to assemble poles (mass of 3kg and length of 2m) on top of each other. In this scenario, Geranos demonstrates an impre
s that classifies and learns accurate forces and moments caused by different robotic arms and interactions. Moreover, the force
ance and drone controllers. The approach and the impact of the system integration are successfully verified empirically, both on
ch module. The functionality and robustness of the system and its controller are validated using physics-based simulations for

algorithms. The proposed framework contributes by enabling UAVs to navigate more autonomously and reliably in critical missio
ed, which capitalizes on the features of the magnetic spring. The proposed control algorithm was validated through several exp
o exhibit the JR-effect

uency response. This asymmetry introduces steerability by a rich set of control commands: rotations with simultaneous forward
and confirm stiffness to scale according to theory. Stiffness ranging from 0.15 N/mm to 1.02N /mm were measured for an overa

conflated in literature, are elucidated. Experimental procedures detailed in previous work are discussed and a novel standardis

ally, the practical aspects of this gripper are described through a quantitative evaluation.

releg-wheel mechanisms. Furthermore, the robot experiences an additional angular velocity because of an asymmetric arrange
16 kg, and the height from the ground to the hip center is 786 mm. The proposed mechanism has been proven to be effective t
, in order to increase towing system’s pulling force and tacking success rate. Three connection modes are proposed, comp
s validated that the number of pointclouds and average intensity of data could be restored to at least 97% and 67% after being

l its lengthening and steering by a feed mechanism. Finally, its capability in a constrained environment is demonstrated by navi

n capability. This work may provide insights into the development of the field of bio-syncretic robots and even tissue engineerin
es made in the design.
of 0.56 mm and signal-to-noise-ratio (SNR) of 12.2 are obtained for the ultrasound imaging at high synchronization along with a
ct to track a time-varying trajectory, maintaining the non-contact with distance constraints. Furthermore, a straightforward plann

on procedure and dynamic compensation of eye motion. In this work, we present a real-time marker-less method for 3D pose t
PVIC when the user used low volitional intent, but yielded higher peak plantarflexion, peak torque, and peak power when the u
ext visualization that renders an automatically identified instrument-aligned B-scan in the context of the 3D OCT volume. Our vi
ward strategy for detecting haustral folds using TSER's stereo camera, which allows for precise identification of their position an
gardless of the breadth and disparity of research directions identified. This paper gathers and analyzes the most significant of th
rated through experiments on a patient's dummy torso, highlighting its potential for achieving safe behaviour and accurate force

eenings. To demonstrate the method's effectiveness, we evaluate our proposed method for the diagnosis and monitoring of abd
on error within [1.0, 2.0] mm using only 2D microscope frames.
e tree graph is optimized by maximally maintaining continuity throughout each rib. To validate the proposed approach, we man
nd chordae. To validate the proposed method, we performed human trials with five healthy subjects and measured the detectio
es on the pure Doppler images to provide a region proposal. Secondly, benefiting from the Doppler signal, this work first introd
e, and shows the promising accuracy which is suitable for CMF robotic or navigation systems.
d. The final force sensing error is less than 7.1%, and the drilling state identification accuracy is at least 93.9%. The designed s
mpliance control of the surgical instrument in contact with the tissue to minimise the risk of tissue damage. In this work we dem
imal root mean square error percentages for vertical and feed force are 1.85% and 1.94%, respectively. Moreover, no instance
as the first step of addressing the issue of prolonged operational time, we built a prototype controller and evaluated the accurac
geons. The experiment focused on drilling around critical structures following cortical mastoidectomy, utilizing dental stone pha
nfection robot system, especially highlighting the ability to detect high-touch areas and navigate in the environment while avoid
a dry lab. Experimental results demonstrated that the proposed framework can keep localizing the hidden lymph nodes, and re
action wrench's gain is adjusted with the gripping force on the integrated tactile handle, ensuring that unintended movements c
t curvature assumption. The result shows that the proposed model can get more accurate results, especially when the driving w
s remotely to control the safety operating force threshold. Finally, the experimental setup was built and a series of evaluation ex
c arm features flexure hinges, eliminates clearance issues, and can be fabricated using an origami-inspired robotic approach. F

RMCR increases by 193% compared with regular MCR. Besides, MRMs on the RMCR can achieve active attraction and sepa
oped robot. In experiments, the robot is used to release drugs after anchoring in the lesion site. Finally, ex vivo experiments are

ng such deviations is crucial when integrating neural networks within surgical robotics, as inaccurate estimations can pose seri

mum speeds of 54 mm/s and 40 mm/s in the simulator and ex vivo environments, respectively. It was found that changing the
ent and closed-loop control were tested experimentally.
ching actions and mean task time, which effectively enhanced the surgical training.
SEs in accuracies and in numbers of high-performance models, hence, were the only supervised network used during experim
rforms 3D deconvolution across time. In order to facilitate training of such deep learning networks, we introduced a new data sy

s for real-time motion monitoring and prediction, which has the potential to form a solid basis for enhancing the automation, saf
 rol algorithm using a PHANTOM Omni haptic device and a force-sensing surgical instrument equipped with Fiber Bragg Grating
be obtained and adjusted online. The overall performance of the operators in terms of time cost for task completion is significan
nder high-frequency electromagnetic stimulation. These findings illuminate a promising path toward amplifying the therapeutic c

t allows the formation of a stable, compliant triangular structure through the interaction of internal magnets. This adaptable con
measure the catheter forces by detecting the coordinate of marks on FTG surface. The calibrated FTG gets the average and m
rs, and compare their performances in both free space without interactions as well as constrained environments with contact fo
otumor. ASTR monitors the NIRF markers and gathers spatial and color data from the samples, enabling planning and executio

ple surgical tasks and the converse in complex tasks. Participants preferred the proposed robotic system due to its user-friendlin
accessibility, as evidenced in our analytical and experimental findings from 22.4% to 99.0% of the required working area on an a
the magnetic navigation system has the potential to significantly impact the efficiency of the clinical workflow and improve com

model demonstrates that the whole system could successfully perform electrocoagulation procedure inside the depth of skull in
easuring the similarity between two temporal sequences, was integrated into the centerline-guided method. The proposed appro
erise and validate the model’s accuracy and the steerability and navigation capabilities of the new micro-catheter.

dity and generalizability of the proposed method.

urgical context to evaluate the derived inverse dynamics model showing a 30.1% workload reduction by the NASA TLX questio
ich have application potential in robotics, biomedical engineering, and environmental governance.

ments. Experimental results show that the microrobots can be controlled properly for designated assignments. Furthermore, the

particles with anodic/cathodic enzymes, and a magnetic layer containing magnetic nanoparticles. The fast self-propulsion of con

of such techniques in experimental research.

t the visual Region of Interest (ROI). After quantitative evaluation of operator results in terms of success rate, efficiency, reset t
ssfully evaluated the knot tying system on 2.0 mm wide bionic blood vessel with 30 mm long #8-0 micro-suture. We finally teste
pth map is optimized by a network refinement method. The experimental results show that the MAE value of the depth predictio
eduction in autofocus time and a 15.8% improvement in lighting uniformity. Moreover, this imaging system demonstrates robust
error of autonomous single cell dissection is no more than 0.61 um.
manipulation tasks.
ves a top speed of 75.0 cm/s and a relative speed of 18.8 bodylength/s. To the best of our knowledge, the 2-DoF origami transm
human head (e.g., brain, eyes, and ears surgeries) or within adjacent regions of human skin (e.g., surgeries in the bladder and
bjects. The CSG network employs innovative modules, including collateral multi-scale convolution, cross-scale feature map gui
he auto-calibration method, as well as the robustness of the visual servoing method against occlusion, blur, and low illumination
referable. The results indicated that the coordination calibration technique provided the positioning accuracy of the capillary tip
3-DOF cartesian robot equipped with inertial piezoelectric actuators. This method can be used to extend the range of any type
ual servo controller equipped with a dynamic sperm motion observer is employed to achieve precise tracking and positioning of
by feature pyramid network. An automated centerline-based tail morphology measurement method is also proposed, in which a
 onal raster scanning and improved image quality in ROI than the standard spiral trajectory, which demonstrates the effectivene
pplied voltage. The method was able to shorten the cost time of delivery and posture adjustment before oocyte injection. Our s
ee-of-freedom control system, we demonstrate the operator’s capability to use X-ray Fluoroscopy images to keep the UMR
ered mixed objects, respectively. The Sim-Suction policies outperform state-of-the-art benchmarks tested by approximately 21%
rved object. By combining LOMDPs with off-the-shelf Markov planners, we outperform state-of-the-art-solvers for both object-o
d experiments demonstrate that our method is capable of generating whole-body trajectories in real-time in challenging environ
nments in a zero-shot manner. It shows robustness to unseen subpolicies, failures in their execution, and different robot kinem
that have not been previously demonstrated, with end effector speeds and accelerations up to 2.0 m/s and 7.9 m/s^2, respecti
meters. The approach is evaluated in simulation and in real experiments on a mobile manipulator and compared to several base
ping. Through extensive experiments conducted on the Habitat and Isaac Gym simulators, we find that our method attains a goo
, and the results show that the stability and robustness of robot systems have improved using decoupled control.
obot performing mobile grasping in cluttered scenes with obstacles. We empirically analyze the contribution of various utilities a
at (i) optimize the control performance and (ii) improve the efficiency of online learning sample collection. We demonstrate the e
he dynamics of clay, and is able to create a variety of simple shapes. Videos and additional figures are available on our project
g the future with a computational effort compatible with real-time feasibility, effectively handling the vehicle at its limits. Furtherm
oach through experiments on a simulated off-road track, showing that our proposed model-controller pair outperforms the base
mework. This provides both the ability to improve fundamental behaviors and perform motion tracking to precision.Notably, we a

a-driven residual is developed, which predicts output dynamics more accurately than either model alone, and without prior expe
w for a small mismatch between expected and predicted values. To train the model, a dataset using a real-world quadrotor carry
mark real-world multi-terrain SSWMR dataset. Our results show that the model generalizes to three different terrains while sign

ood pressure by 2.8 mmHg and 1.7 mmHg, respectively, as well as calming their heart rate by 8–10% on average.
the application of integrating into ChatGPT also shows huge potential. The ESP-dataset and all benchmarks are released at url
egmentation masks and bounding boxes. We validate the efficacy of our datasets through extensive synthetic-to-real experimen
t, grasp quality is predicted using a linear array network (LAN) and global visual images as inputs. The conditions such as gras

h community to promote a deeper understanding of human motion control and better design of interactive robots.
s can be mapped to 3D space to guide the 3D decoding with high efficiency and high accuracy. Extensive experiments on pop
oblems of which the following parameters have been adjusted: the fixed duration and the frame to export. We based this adjust
behaviors in LJ-based continuum robots.

ic controller that is capable of positioning the tip, tracking trajectories, as well as performing interactive tasks in the 3D space.
e resulting IK formulation involves only quadratic expressions and can efficiently incorporate a large number of collision avoidan

nto a simulation in Unity for tracking the shape of a cable with a length of 1m and a radius of 5mm. The simulation results show
able control performance for the soft robot without needing to repeat any online parameter adaptations, which significantly impro

simulator, and such method could suppress the hand tremor from the original amplitude of more than 5° to less than 1°. The
ulation and on the physical system. The position is estimated with an accuracy of three to five millimeters for a 130 millimeter lo
one can choose the number of modules to actuate for each throwing direction. Furthermore, the results show that the proposed

l approach, showing an average accuracy of 0.67 mm (0.66% arclength) in target positioning experiments, and an accuracy of
lation studies suggest POE-M's guidance of the key points during the mesh reconstruction process provides robustness and st
res the complexity of chemo-mechanical interactions that drive shape changes in biomedically relevant micromachines. Finally,
ardware, we demonstrate a novel use case for underactuated, elastically coupled systems in "sensorless" force control.

udy can be used in a toolbox for obtaining subject- and task-specific cost functions and assisting the development of personaliz

constraints guide the manipulators to move on zero-perturbation manifolds (which inhabit a subspace of the tangent space of t
thermore, we provide a geometric optimality interpretation of the optimal gait.
ted and in real-world trials, using the CENTAURO robot platform. With these experiments, we demonstrated autonomous capa
f the STL requirement in deployment. A backup policy is designed to ensure safety when our controller fails. Our approach can
in simulation that shows large (up to 100x) speedup compared to a naive approach. This result is significant for applications su
s to find all distinct paths in challenging environments. To this end, we propose the CTopPRM which creates a sparse graph by
ts demonstrate that SVG-MPPI outperforms both the original MPPI and other state-of-the-art sampling-based SOC algorithms
ng a deterministic, local-minima-free path planner, setting apart the proposed planner from traditional methods. Lastly, we prov
ur path planning algorithm and evaluated in various driving scenarios.

ns and real-world experiments demonstrate the effectiveness of our energy-aware ergodic search, which ensures continuous ex
f simulations and real-world robot experiments were conducted to evaluate the algorithm against existing planning algorithms. C
tlasRRT*, a sampling-based planner dedicated to planning on implicit manifolds. Simulations in different operation scenarios we

enging planning problems. Our method surpasses previous works in path planning benchmarks while preserving probabilistic co
reside in the range of control frequencies and open up new avenues of motion planning research.

leads to a complexity of O(n^ 4 ) when reasonable time step sizes are used. To evaluate RETRO’s performance in terms o

aerial systems, environmental sampling tasks using uncrewed marine vehicles, and agricultural tasks using ground vehicles, am
CTMP algorithms. Our proposed framework, as it operates as a constant time algorithm, rapidly generates an initial solution with
nse vegetation. We observe that VAPOR's actions improve success rates by up to 40%, decrease the average current consump
mpute scene-specific costs such as "collision cost" and guide the diffusion to generate valid trajectories that satisfy the scene-s

e main algorithm, using a "fast-forward" approach that foregoes guarantees of asymptotic optimality, implementing heuristics th
cond, an iterative path perturbation algorithm is adopted to locally reduce the number of unnecessary heading changes and imp

Moreover, the proposed hierarchical state validity checking scheme reduces the planning time by 30% compared to pure volum
ng into account the gate's configuration. We validate our planner in real-world flights and demonstrate super-extreme flight traje
t geometric front-ends are sufficient for environments with varying complexities if combined with dynamics-aware backends. Th
certainty. We plan to release our code to benefit the community.
considering dynamic feasibility and anticollision constraints. To avoid pathological hitting, a defensive rule constraint and its co
ol scheme is experimentally validated using an octorotor.
, the approach indicates improvement in a solution to the kinematic orienteering problem and the team orienteering problem in
ty and potential application of the proposed modality.
perimentally verified. Towards this direction, experimental emulation serves as the crucial stepping stone, providing a pseudo m
prove both transient and steady-state position tracking response, reduce energy consumption without requiring any backlash m

e, which have demonstrated that the hovering quadcopter is able to recover from one rotor failure by the proposed controller an
proposed approach is demonstrated by simulated and real-world experiments on a 7 DoF Franka Emika Panda robot.
on control, path following, and vehicle dynamics. We describe the design of JetRacer-4WS, and experimentally demonstrate Je
n and a sequential algorithm, by which the problem is further transformed into polynomial-time solvable surrogates. The simula
otor fluctuation.

h external forces, but also to distinguish load forces at different frequencies and potentially address the issue of mechanical res
tiveness of the approach is demonstrated through simulation and hardware experiments.
anning approaches. We validate the effectiveness of our proposed framework through practical experiments involving a quadru
date the proposed ARC$^2$M scheme, simulations and physical experiments are carried out. The source code is available at h

ments, including steering and climbing a slope. The experimental results indicate that our scheme allows the RW-Robot to perf
performance of the proposed controllers, and also show the feasibility of implementation on high-dimensional robots.
echnique of piezocermamic-based fine positioning stages is introduced and verified experimentally using a representative preci

ded by the velocity of human hand. The autonomous scenarios and controllers switching are accomplished through a motion fu

rface polishing task.

ervo control is applied at the lowest motor level. This paper investigates the modeling, sensing, and control of CDSM within a u
9% improvement in viable task region size and a 21% reduction in motor torque cost. All videos and code are available at https:
l the convergence is reached. Moreover, we implement our algorithm on a tractor-trailer system, taking into account the longitu
to learn than a naive shape transition model directly mapping the shapes. Through evaluation and comparison by simulation a
nt unknown objects on a table to track given paths. The hardware experiments show that with a rough prior model, the multi-co
ng protocol. We show that the operational limit for our model is reached in extreme slip conditions encountered on surfaced ice
f the condition is not satisfied, we will still find approximate solutions by numerical optimization. Coordinates can be defined on

ity of the proposed methods. The prototype that applied our method can reach targets and accomplish tasks, further validating

utions. We demonstrate the effectiveness of IKLink through a simulation experiment and an illustrative demonstration using a p
mulation and real-world, including object manipulation under orientation constraints and door opening with a high-dimensional 6

nd collide 90% of the time in problems generated with an existing planner benchmark tool. We also analyze the transferability o
us to capture information decay due to a lack of measurements and to quantify the maximum attainable information in stochas
ry Material we provide theoretical foundations for our approach. The Supplementary Material and our open source software are
the constraint of the relative angle between the tether stretching direction and the robot’s heading direction to be admissibl
ver-approximating robot geometry, CBF-INC can balance safety and goal-reaching better without being over-conservative. Give
discrete optimal base trajectory planning algorithm to solve several mobile printing tasks in hardware experiment and simulatio
data, demonstrating the generalizability of our algorithm.
in the real world and simulation, generating behaviors we could not achieve with previous methods. The results show that com
ized, so that the UGV can safely and smoothly follow the path to the target location. Compared with single UGV exploration sys

aring it to single person persuasion. The results showed that the robot's gestures increased compliance by the group and the g
n system for the bimanual robot, introduce a measurement confidence estimation technique, and propose an occlusion-resilient
g of multimodal input, multitask planning, and adaptation and processing of the intelligent sensing modules in the complicated s
gnition and fine-weighing operations. This bidirectional tactile interaction framework will enhance the immersion and efficiency o

ing the exploration of rich local semantic information and facilitating effective feature alignment. Furthermore, CAMInterHand e

sly monitors interactions that fluidly change between intentional or unintentional. If deemed unintentional, the proposed intention
demonstrates superior generality to diversified visual and language inputs with a high success rate. Codes and demos are avai
een task completion and the efficiency of navigation and interaction remains challenging for all methods. We further provide mo
n for pragmatic inference. Experimental results show that PROGrasp is effective in offline (i.e., target object discovery) and onli
ency and improving shape discrimination. MMCNN enables unified force and shape modalities, validated through real-contact e
(i.e., clear, ambiguous, or infeasible). We validate the proposed method on the collected dataset and pick-and-place tabletop s

and interface and the monitor system interface. The results showed that although the novel interface presented a considerable l

AV with autonomous visual geolocation. UAV agents send images and relative positions to the ground server, which processes
the accuracy of collaborative SLAM estimates. Furthermore, our system demonstrates its competence in large-scale missions,

the code for public access at https://github.com/minwoo0611/MA-LIO.

segments given in random order and with no initial estimations. Furthermore, AutoMerge performs posegraph optimization after
wn payloads, uneven terrains, etc.
bipartite matching, our approach allows multiple robots to address the same task as a subteam. Our approach also enables vol
tasks, modeled as a flexible size multiple vehicle routing problem with time windows (m-MVRPTW), given the predicted target
 (based on queueing theory and a population model) that may help to support this reasoning.

s high interpretability and flexibility — in real world experiments, we show RoCo easily incorporates human-in-the-loop, where
are. Our work is the first work that demonstrates the possibility of learning neighbor-avoiding and obstacle-avoiding control poli
robots crash and get replaced by new ones, this meta policy will be reused to guide new robots on how to quickly follow the ex
outline the algorithm, sketch a proof of its improvement over a base policy, and evaluate its performance on a challenging prob

unication. Experimental results in simulation demonstrated that, using the proposed visual coordination method, multi-agents u
omparison with the ideal case of isotropic sensors and communication. Despite an increase in the amount of time required to ful
case where we neglect reputation.
sensor measurements or dropped communication packets. Furthermore, we showcase that the system operates on real robot
develop an expert algorithm for controlling a swarm of robots and employ imitation learning to train our decentralized STGNN m
set from coarse to fine. Comparisons between the proposed method and the existing ones through extensive simulation tests pr
anges vary. Finally, we demonstrate how MOCHA can be tightly integrated with the planning stack of autonomous robots. We sh
trate proof-of-concept with a team of a VOXL m500 drone and a Clearpath Jackal ground vehicle, providing a complete system
ss passive docking of cells that generalizes to any polyhedra with radially symmetrical faces. Future work will allow the cells to

ssion execution quality while preventing faulty assignments. To evaluate the system, we simulated a find-and-decontaminate m

e number of hyperarcs scales quadratically with the number of robots and linearly with the number of objects. In contrast, the n
pable of handling sensor blinding through leveraging an inter-agent range measurement, despite a dynamically uncertain enviro
ance across simulated runs through two datasets and compare to the performance of an algorithm representative of most sche
tors. Our proposed approach is assessed using both a publicly available dataset and field experiments1, confirming its efficacy
at the approach of runtime configuration of the MPC drastically decreases computation time, while retaining task execution perfo
monstrate our method can outperform other methods by approximately 10.5%-27% on success rate and 70%-75% on inspectio

intent representation’s ability to encode and facilitate negotiation over points of contention.

rence combined with Thompson sampling for decentralized multi-agent decision making. We compare different action selection

d in the state of the art.

odified by adding encoding of robots' state graph, and two Multihead Attention based decoders whose output are used to constr
l network distribution, and ore-car dumping. The objective is to describe these processes and provide insights on some of the c
nding theoretical methods to an industrial application by facing the typical practical issues such as graphs difficulties (load/unloa

ol Barrier Functions (HOCBF) as a mean to maintain a minimum level of observability for the targets. Simulation results are rep
f geometric conditions. An application on how to use the knowledge of all singularity cases in a formation for singularity-free con
n, we derive a multi-agent model-based policy gradient method to gradually update the robots' patrolling strategies towards the
 imulations that vary surface pattern and initial robot poses. The optimized algorithm parameters show up to 55% improvement
h that first 1) uses A* to determine a sequence of block placements and removals while ignoring inter-robot collision, then 2) id
significantly improved performance in terms of time efficiency (5 times faster), grasp quality metrics, and success rate (5%-10%
on by others.
interaction forces and proximity to conductive objects, and its utilization in various tasks was also demonstrated successfully.
public datasets show that our method outperforms state-of-the-art methods regarding grasp motion on various metrics.
le feedback to determine appropriate force, position, and actions, resulting in more accurate imitation of the expert’s action

not labelled in the datasets. As a result, we propose a classifier to identify the pattern of the expert behaviour within a mixed da

a balanced solution between control performance and landing safety. Finally, extensive simulations are conducted to substanti

iveness of these controllers.

age the reader to view our demos at https://aalpatya.github.io/gbpstack.
p fashion utilizing a probabilistic finite state machine (PFSM). Finally, the efficacy of the proposed scheme is experimentally va
s also provided. CloudGripper-Rope-100, a dataset of more than a hundred hours of randomized rope pushing interactions and

ot exploration.
enarios with high communication delays.
or accuracy by 80.1% and 67.3%, respectively, compared to iNeRF while performing at 34fps on consumer hardware and not re
b.com/PJLab-ADG/VeloVox.
n experiments on the public radar point cloud dataset View-of-Delft (VOD). With IoU thresholds 0.5/0.25/0.25 the average pred
ediction, while the capability of DC is implicitly embedded into the detection Transformer through a depth denoising mechanism
d contrastive learning, our framework then reduces the weather gap and acquires instance features that are robust to weather
he effectiveness of our symmetry-aware training procedure and demonstrate that our approach is robust towards inaccurate cam
oughout the model's training. The framework demonstrates comparable or better performance in KITTI 3D object detection benc
er previous state-of-the-art RGB-based approaches, especially in terms of rotation accuracy. Code: https://github.com/goldoak/
ence (42 FPS).
ction methods. To achieve this goal, we adopt a semantic-segmentation network with a sigmoid head that simultaneously provi

lving the performance gap with respect to the teacher.

-agnostic network modules that can be plugged into existing point cloud transformers for end-to-end training. We evaluate our m
ances the performance of the temporal fusion method MPPNet. We also conduct extensive experiments on motion model selec
E), which utilizes spatio-temporal virtual (STV) grids for fine-grained representation. We also present the attentive pillar aggreg
r model can achieve the state-of-the-art performance on public benchmarks. Especially on CamVid, comparing to baseline, our
on the KITTI, Waymo, and nuScenes datasets, and the results demonstrate the state-of-the-art performance of Shift-SSD in bot
efficiency.
e efficiency in runtime.
number of pseudo-labels per image based on model confidence. Our method outperforms state-of-the-art approaches on seve
 the BEV, and (2)a multi-scale feature fusion mechanism for LSS-based view transformation. Extensive experiments on nuSce
arge-scale nuScenes datasets, and we observe that utilizing such predictive information significantly improves the overall detec
omain. Finally, it is proved that the improved Yolov5 algorithm improved mAP@[.5:.95] by 6.42$%$ and [email protected] by 9.3
ply it to control robots in real-world environments. In experiments, our method improves boundary and structure measurements
cted on three challenging drone-to-drone detection datasets. Our results demonstrate notable improvements, with F1 score enh
nd significant performance gain (up to 9 AP) across multiple state-of-the-art models and datasets with a negligible additional la
ormance on OOD detection for semantic segmentation benchmarks.
ccuracy in an uncertainty-driven manner with minimal, non-disruptive interactions (ca. 2-3 per scene). We demonstrate the effe
ignment pipeline that rectifies inter-agent frame misalignment. This method leverages a zero-shot image segmentation model fo
ines. Furthermore, it offers real-time inference which is well-suitable for robotic manipulation applications.
o label quality in comparison to other autolabelers; (2) high evidential uncertainty estimates; and (3) lack of clear interpretability
ferentiable physics module can predict contact information for pose refinement. We also show that our approach generalizes to

o exploit their different expression capabilities. Finally, we utilize long short-term memory to gradually classify dynamic objects.
a recent appearance-based classifier and approaches the performance of offline methods. We demonstrate its generality on a n
d on VI-SLAM data; we examine the extent to which both object and environment characteristics impact performance, along wi
e. SemanticKITTI and nuScenes. In addition to being real-time, the proposed method achieves state-of-the-art results on nuSce
state of the art.
mmetry groups based on multi-sensor fusion. The proposed framework extracts local equivariant features of objects in point clou

I-DPS show that EfficientDPS can achieve the best trade-off between speed and accuracy than the state-of-the-art methods.
es generated by the HFP module, the Pyramid Attention Integration (PAI) module is introduced to augment local semantics and
y. Our project website is available at: https://ai4ce.github.io/CID/

ctor-agnosticism of our WLST framework. Notably, it outperforms previous state-of-the-art methods on all evaluation tasks.

mances among single framebased radar-camera fusion methods in the nuScenes 3D object detection benchmark. The code wil
te that both OVAAL and OVAAL+ lead to significant improvements, with mean Intersection over Union (mIoU) gains of 4.55% a

a general solution to the selection of multiple frames while taking into consideration the possibilities of disc intersections. Final
on measures like entropy or CoreSet, size-based sampling does not require a pretrained model, thus addressing the cold start p
ng. Experimental results demonstrate that the ATP exhibits superior prediction robustness and accuracy in perturbed environm
d their rate-of-changes are taken into account. A Python implemention confirms the real-time execution capability of the control
n terms of peak power, torque efficiency, and tracking capability.
ymmetric stair preconditioner provides up to a 34% reduction in condition number and up to a 25% reduction in the number of re
m solvers by at least 10x for a majority of solves and 3.6x on average.
roach through simulations and experimentally validate its performance in dynamic waypoint scenarios for time-optimal trajector
e space, the proposed invariant descriptors for motion and force trajectories may prove useful for the recognition and generaliza
he same linear time complexity from a single forward-backward pass. We experimentally demonstrate that (1) we achieve bette
nstrained trajectory tracking, obstacle avoidance, and postural optimization. Two mobile manipulators are used, each having 10

the prevention of unnecessary clutch transitions. We empirically demonstrate the effectiveness of our proposed method on a d
uses locally optimal feedback policies obtained from DDP. We show through simulations and experiments that the method can
chniques. This work paves the way toward a more systematic integration of force sensors in model predictive control.
kinematics than state-of-the-art robotics libraries.

ate our approach on a two-agent quadcopter system on hardware under tight input constraints.

the quadrotor as well as a hybrid system, the planar pusher. To the best of our knowledge, this is the first time that a SOS-base

tor, to further validate the approach.

tems, including a car model with 6D state, and a 3-joint planar manipulator. Our results show the great capabilities of CACTO in
ons and experiments.
being unable to know the noise's realization a priori. We validate our algorithm in simulated scenarios of (i) an inverted pendulu
alidated through some comparative simulation results of a biped walking robot with external forces.
ate information. Based on the small-gain theorem, the notions of strong unboundedness observability and input-to-output stabi
as been tested on two different underactuated systems: the torque-limited simple pendulum and the cart-pole. Extensive simula
ses the exploration performance in terms of ergodicity and scales well to real-world problems. We compare our method in kinem
le problems.
nvironments. We also experimentally compare CODAK with previously published white-box solutions to evaluate the penalty of
multi-step behavior when compared to alternative learning techniques. Moreover, our approach successfully accomplishes the
active, including any time the convoy waits. We refer to this problem as the Assisted Shortest Path Problem (ASPP). We prese
hat distributing and assigning coverage responsibilities to agents depending on their sensing and motion models leads to 40%,
blation models in our designed environment. Demo videos for this work can be found at: https://sites.google.com/view/samarl
S-KATS's performance. Our work is directly applicable to many similarly scaled traffic and multi-robot scenarios with directed lan
ansmitted information, on average, by approximately 50%, while maintaining satisfactory performance.
sk assignments for minimizing makespans.

on of the benchmark framework with two different environments and three centralized coordination methods (two MAPF algorith

n simulation, we demonstrate a $60%$ improvement in success rate, an average $1.8times$ reduction in mission completion tim

results with robot dynamics such as unicycle, double integrator, and car with trailer in different settings show that our method is
to reason about each others' path intentions, thereby enhancing the level of cooperation among agents at the whole system lev
e time required to achieve complete coverage. For validation, we perform ROS+Gazebo simulations in five 2D grid benchmark w

dels. We validate the effectiveness of our MIP-based MCPP planner through experiments that compare it with two state-of-the-a

s subsequently deployed in an outdoor one in order to evaluate its robustness against noise. Throughout the outdoor experimen
n. The performance of the proposed method is verified by exploring 3D simulation environments in comparison with the state-of
ctively detect objects across various scales and angles. Through experiments, it was demonstrated that incorporating this synth
ctor of 10 while maintaining acceptable detection performance. Furthermore, our fusion architecture reaches 94.6 mAP on the t
ee challenging datasets: KITTI, EuRoC, and KAIST-VIO, and two highly accurate SLAM backends: Full- BA and ICE-BA show
obstacles in 3D space and continue the flight smoothly with the help of the collision-aware map. In addition, we will open-source
-96%, constituting a 16-29% improvement compared to the best heuristic methods. We demonstrate the functionality of our app
magnitude compared to state-of-the-art incremental GMM-based mapping. In addition, the proposed approach yields a superio
asure by the fraction of the explored space as a function of the trajectory length.
y in path generation. We conduct extensive benchmark and real world tests, where FC-Planner computes over 10 times faster
simulation and 26.4% in real-world experiments.
on-making capabilities to each UAV to switch between different execution modes, the proposed strategy is shown to strike a gr

ork, which integrates onboard sensing and a neural network controller, achieves a success rate of 87.36% in real-world experim
e that best matches the teleoperation input while satisfying the safety constraint encoded by the CBF. In contrast to existing pe
o the array. This asymmetry is direction-dependent, allowing us to predict the headings of flow from two different sources, like w
event mismatches utilizing point clouds' statistical features. In addition, our system includes an optional and auxiliary learning-b
and real-world experiments, we demonstrate that the trajectories generated by our method significantly improve state estimation
obile manipulation approach, demonstrating its feasibility.
ent induced by updrafts, which improves the precision further. Compared with a particle-filter-based method, we can determine
of 14.8 frames per second (fps) and power consumption of only 2.9W while the CPU and GPU temperatures are maintained be
orrespond 700k plants and confirm a link between a robotic seeding policy and germination.
ntional methodologies, which rely on static scene snapshots, our method continuously captures the evolving scene geometry, re
ool manipulation task.
achment. Throughout all examined scenarios and experiments, our approach remarkably improved the robot's performance. by
improvements. Extensive experiments are conducted on the GraspNet-Planar benchmark and physical environment, and supe
se training two additional networks: a residual NeRF learns to infer residual RGB values and densities, and a Mixnet learns how
ation. Finally, we take the set of 3D edges with beliefs greater than a threshold and apply a post-processing step to reject false
gnificant commercialization potential.
or robotic hand-over.
y, while the remaining 7% required more attempts.

y conducted on two newly-constructed lowlight grasp datasets and the proposed method achieves 98.9% and 91.2% detection

ticulated objects, in particular objects in which the articulation is not apparent from vision alone. We implemented our system o
rom the frictional limit. To find this value, we use a high-density camera-based tactile sensor to measure the 3D deformation of
on a synthetically generated dataset, SynWidth, which contains randomly generated width-mate objects with symmetric or near

h human subjects under controlled conditions. Over 900 mean-picks-per-hour is reported on a single-arm system. For dynamic
ntinuous and robust harvesting capabilities. The result underscores the potential of harvesting robots to bridge the labor gap in
 % for 10 unknown objects, and demonstrate the method's capability to grasp objects in structured clutter and from shelves.
ng mode. Our developed learning-based model achieves an accuracy of 93.33% in classifying 9 different types of liquids (8 liqu
periments on a Stretch RE1 robot, we study the effect that IVFP has on performance - both in terms of effective data throughpu
g between shape categories rather than just specific objects.

t redundant manipulation times. A standard visual localization pipeline at the object level is developed to capitalize on the adva

st time, but solely a point cloud acquired from a single RGBD camera. Our resulting policy first predicts the desired point to cho
roaches such as finite differences. The pseudospectral parameterization also provides a means of trajectory data compression

ifficulty of the clutter. A website dedicated to the benchmark containing information on the different tasks, maintaining the leade
roject are available at https://irvlutd.github.io/SceneReplica.
ion models; 2) A set of new bias-free metrics for measuring diversity, by incorporating the characteristics of a given scenario, a
ass LiDAR-based detectors in downstream applications. We believe the proposed metrics and the new benchmark dataset will f

are-accelerated robotics.

We assessed the system's effectiveness in coaching skills through robot interaction in a sparring experiment and a survey involv
ontact point displacements in the stimuli direction and, simultaneously, sets the desired stiffness based on the muscle activatio

ed by exploiting varying virtual mass element. Finally, experimental results are provided to demonstrate the performance of the
simulation setup.
ons. In addition, we assessed the knee indicators to identify the variation in motion performance. The subjects performed 120 s
s, eight for each CDLE configuration, were tested. Results showed varied gait adaptation among the three groups such that the
nd used in a series of pHRC experiments. Our preliminary findings suggest that PC is more precise and robust than the rating s

on and on a real system. In addition, we can perform evaluations on a larger scale compared to prior work. With our newly intro
he key component is a real-time intent recognizer that helps the robot respond to human actions. We describe the details of the
tal in quantifying the performance of the (neuro)musculoskeletal system. Thus, the combined use of human modeling and robo

that base compliance has a significant impact on task accuracy as it can easily excite the system if an inadequate control strate
wer, showing agreement of trends and validating the use of a simulation model to predict physical human-robot interaction beh
and damping) to better fit subject-specific needs and characteristics. Experiments with twelve healthy subjects exhibited a sign
or lidar-based multi-agent exploration, which can direct one agent to repeat another agent's trajectory adaptively based on the q
e the first results that provide valid safety guarantees in such a setting.
o Unmanned Aerial Vehicle (UAV) and a quadruped robot Unitree A1.
he results show that the proposed controller greatly reduces the number of constraint violations compared to state-of-the-art ro
ning. The efficacy and computational efficiency of our approach are demonstrated through numerical simulations.
belief hypotheses.
ng us to establish a trade-off between computational efficiency and performance. We demonstrate how these bounds can both b
d, including factored variants designed to improve scalability. Experimental results on a set of benchmark domains demonstrate
g efficiency for uncertainty-aware agents navigating in the graph.

constrained partially observable robotic domains, showing that it can plan successfully in continuous CPOMDPs while non-hier
high uncertainty both on 3D unseen regions and its involved 2D rendered pixels, compared with recent methods. Furthermore,
transition, we introduce the notion of context-awareness, i.e., we store local environment information for each executed transitio
e-art planners in dynamic and large-scale environments and is shown to be more effective at both exploration and collision avoi
tion in trajectory portions residing in non-traversable regions. Our video is here: https://youtu.be/3eJ2soAzXnU
h is an ordered search guided by an informed heuristic. We test IBBT in different planning environments. Our numerical investig
ally robust Conditional Value at Risk constraints. The resulting formulation offers computational efficiency as well as robustnes

R program. Our method, Dynamic Replanning via Evaluating and Aggregating Multiple Samples (DREAMS), outperforms other
g the uncertainty associated with measurements to weigh their contributions. We evaluate our perceptual maps in a simulated
ervable, time-varying environments. Our results demonstrate superior performance over standard methods, highlighting the fram
ing the SLAM ASP as a partially observable Markov decision process (POMDP). Our experiments show that dynamically selec
s benchmark methods in terms of both search efficiency and estimation accuracy.
he solutions of similar planning problems as experience to speed up planning queries and use it to efficiently construct the datab
ctiveness of the massive data augmentation in combination with deep ensembles and the proposed scenario-based planning a

ollover and extremely high slip. Simulation study confirms that the proposed algorithm achieves up to 20 % reduction of the pro
collision avoidance in the braking zones. Finally, the non-convex optimization problem is reformulated into a second-order con
goal with minimal cost on a single EG. We show that RL efficiently solves problems with up to 20/4 or 25/3 nodes/agents, using

ores all topological routes, until the best 𝑘 of its leaves reach the goal. It is proven that different branches of the tree explore
e prioritized planning method is given for planning paths for multiple robots residing on an LWCS.
alition changes are needed. In this work, we propose to let agents learn to iteratively build cooperative schedules to solve such
emonstrate that our method can reduce the effect of potential delays from contingencies. We show that our algorithm is comput
nflict-free local paths. In addition, we also explicitly consider motion/communication uncertainties that are unavoided in practica

less electrical power is consumed when the spring mechanism is used, which implies a lower-powered actuator is sufficient to
ld label bone peaks in the raw US signals. As a result, our method achieved sub-millimeter precision across all eight bone area
d and electrode placement. The empirical results yielded from our study underscore the efficacy of our approach, particularly in

series is used to model the impedance parameters of each joint as continuous functions of gait cycle progression and task. Fou
n particular, it is shown that the AFP mechanism with 54W power provided in toe-off phase can reduce the peak power of the m
al that even brief exposure to a low-amplitude force field, or to uncompensated residual friction and dynamic forces, applied at th
heme provides an accurate classification of intuitive hand motions, enabling the amputee to perform versatile finger movement
sers during walking. In the experiments, the BSE reduced the activation level of the back muscles by up to 40% during lifting, w
ers restoring force and flexibility to the prosthetic wrist, allowing it to withstand external disturbances. Furthermore, the propose
e to transfer large torque loads of up to 57.4 Nm. This work contributes to the design and development of more flexible and ada
 to validate the robustness and reliability of our approach. The proposed algorithm achieved high accuracy in detecting transitio
eness of our approach, simulations, and experiments were conducted. Our push recovery experiments employing the exoskelet
cross subjects (P1: 52% range 5%, P2: 49% range 6%, P3: 39% range 3%) suggesting that, for expert users of some devices, t
subject could walk comfortably and continuously at their preferred walking speed, achieving normative ankle torques and joint t
ess the effectiveness of the integrated system, which merges object recognition and gravity compensation within a built-in struc
found at https://theairlab.org/gicandidates/.
d in this paper is shared within the libPeR library for Perception in Robotics: https://github.com/PerceptionRobotique/libPeR_bas
spins), that commonly occur during UAV flights. In real-world experiments, we demonstrate the accuracy and applicability of Ra
ewpoints on the KITTI dataset, and at different periods of time localising between KITTI and KITTI-360. We achieve accurate m
neficial for improving the robustness of the LiDAR-SLAM systems.

in order to learn to reduce the bias based on the constructed map consistency signal. Complementary to the removal of the bia
e is available at https://github.com/dcmlr/groundgrid
of-the-art results compared with other recent self-supervised methods. Moreover, it shows a good capability of being applied in
n system that is versatile and robust. We deploy the system on a Micro Aerial Vehicle and show how it navigates through an ind
ughly 26−43% compared to standard navigation methods without wall-following.
e several paths to avoid obstacles, because MPPI assumes that a distribution over an optimal control input sequence can be a
tionary and non-stationary scenarios.

rk: 1) coil placement during transcranial magnetic stimulation and 2) drill and injector device positioning during femoroplasty. R
m a single demonstration video, and of incrementally updating them when new demonstrations are available. To reproduce man

s://rb.gy/3ov2r8.
results show significant improvements in terms of the torque magnitude the device transfers to the user, enhancing its efficienc
he structural components of the system, which motivates future redesign considerations. Overall, our method provides a framew
stics were evaluated based on the joint torque generated by the actuator. The actuator utilized pneumatic pressure to generate
and encoding the forces using Gaussian Mixture Regression, and reversibility is accomplished by means of Reversible Dynami
group of healthy volunteers using the TWIN lower limb exoskeleton.
verage error over the workspace of 5.5 cm. Beyond the application of FES, our method can be extended to other control system
ct. Our results indicate an increase in tracking performance with lower overall assistance, and less muscle fatigue when the hyb
performance. The system includes sound effects and verbal feedback from the robot to keep users engaged throughout a sess
algorithm. As a result, our method outperforms the existing state-of-the-art methods to a large extent. Furthermore, we validate
bsite: https://sites.google.com/view/safemdp
RL algorithms, signifying a substantial advancement in the field.
m and competitive multi-agent racing, against learning-based and control-theoretic approaches. We empirically demonstrate the

safety and then use it to introduce an aggressiveness parameter that allows for fine-tuning the mentioned tradeoff. The experim
ces the memory footprint by as much as 14.2x and 16.7x across tasks from the Atari 2600 benchmark and the DeepMind Contr

hensive feature representation to deal with complicated road scenes. Extensive experiments conducted on the OpenLane datas
emerging problem of vertical mismatches in automotive AR displays. To tackle this issue, we begin by defining the problem and
nvironment. Our methods can be easily added in standard policy optimization to improve learning efficiency, which we demonst
storing or transmitting the data to public servers. Nevertheless, these gradients contain sufficient information to potentially expo
k featuring several high-performance MARL algorithms across typical scenarios for robots with continuous actions, which highli
Our experimental results in the D4RL benchmark and a real-world manipulation task confirm that our method can benefit more
of our method through extensive simulated experiments. To the best of our knowledge, REFORMA is the first robust RL approa
gy consumption efficiency by 14.2× and 15.3×. To the best of our knowledge, ReactHD is the first HDC-based framework
on the estimated flexibility to perform challenging DLO insertion tasks. Our pipeline trained with diverse insertion scenarios ach
ing a graph neural network, 2) an action policy, and 3) a query policy which decides whether or not to query for the event detec
o haptic and auditory sensory modalities. Through extensive experiments on a dataset where a humanoid robot interacts with 1
work in challenging scenarios, including real-world cabinet-like environments with arbitrary household objects. The results show
ur key idea is to encode the neural radiance field with multiple MPIs facing different directions and blend them with an adaptive

haustive set of experiments and ablation studies to identify the performance of our approach, while also discussing limitations t
der architecture based on standard Transformer, where the encoder operates only on the visible subset of patches to learn loc
on POCR achieve better performance and systematic generalization than state of the art pre-trained representations for roboti
tial improvements in our framework over existing robot hanging methods in the success rate and inference time. Finally, our sim
for online manipulation task classification and task-level perception.
rom SlotTransport and performs relational and spatial reasoning to predict the future appearance of each slot conditioned on ro

-world datasets such as KITTI & EuRoC and compare the refined features with existing feature descriptors. We also evaluate o

onment together with high-resolution local information to train a neural network. HIO-SDF achieves a 46% lower mean global S
nd real settings, approaching the state-of-the-art robotics-aware representations. Appendix and videos: https://sites.google.com

pproach are verified via synthetic and real-world experiments, demonstrating our method’s ability to generalize effectively a
d outperforms multiple baselines while providing generalization to novel objects and scenes. Code and data is available at https
he same evaluation criteria. Its robustness in estimating poses is greatly improved, especially in an environment with severe oc

not disclose their datasets, making MMAUD an invaluable resource for developing accurate and efficient solutions. Our propose
milar performance on seen objects. We further show competitive results on the real dataset OCID. With its lightweight design (0
to state-of-the-art methods for unseen object instance segmentation.
xcels in all SSC metrics and shows great temporal consistency.
st depth estimation in diverse weather. Besides, to reduce manual intervention and better adapt to different models, we designe
uate our approach using a high-fidelity simulator that considers realistic traffic, communication bandwidth, and vehicle sensing a
re plagued with outliers. We then propose a simple approach based on Expectation-Maximization (EM) and establish theoretic
compared to State-of-the-Art algorithms.
of eight planar surfaces from a wide range of viewpoints, and show that our method outperforms the proprietary-distance-estim
ur VG4D achieves improved recognition performance. To enhance the 4D encoder, we modernize the classic dynamic point clo
er of downstream planning tasks that are specified through abstract (language) prompts and require complex reasoning over s
ses and sampling-based regularization for rare classes that alleviate the need to manually select thresholds and can dynamicall
of the proposed joint design.
ee-degree-of-freedom (3 DOF) solar tracker which reduces the area of its shadow by reducing its height as it tracks the sun. In

s successfully attached to various slender objects of different geometries with a maximum holding force of 92.67 N (up to 135.1
ating all possibilities for screw dependencies, deriving a simple equation for each case, and then describing each singularity typ

es is that the robot is backdrivable since it uses direct-drive or quasi-direct-drive actuators.

kills (160266 tasks). We show that a high-capacity model trained on this data, which we call RT-X, exhibits positive transfer and

n leverage large-scale human videos for learning, the translation module only requires a small amount of in-domain data, and c

d to learn FT-NCBFs by minimizing a loss function constructed using the derived conditions. Using the learned FT-NCBF, we s

vide the numerical example of a Dubins car, utilising HJB reachability analysis along with the algorithm to obtain the maneuver
or both in simulation and experiment.
cified using signed distance functions (SDFs). We formulate the cloth manipulation task with safety constraints as a constrained
can be used for safety verification. Finally, we present results from hardware experiments demonstrating our approach's ability
to withstand the impact momentum (26.376 kgâ‹…m/s) on par with a reported average effective momentum from straight punch
omalies to preserve system safety. We validate the proposed approach on an autonomous aircraft taxiing system that uses a v
ncoder (CVAE) to learn a state-conditioned disturbance distribution, and find the resulting probabilistic safety controller exhibits
earned IPC failed to do so.

t move backward over the horizon, and assuming that such set satisfies a condition that is weaker than control invariance. Safe
times as accurate for the same computation time, or 6 times as fast to reach the same accuracy.
the practicality of our approach.
le, and fast collision detection method based on joint angular velocity band-pass filtering, a method exploiting the unique low in

backup strategies for human-robot interaction. An LSTM classifier uses data inputs from the robot, human, and safety algorithm
ine rotation and translation by aligning equivariant features aggregated over the graph hierarchy. To estimate scaling, we lever
tional standard normal distribution. Second, we learn a richer and more expressive latent pixel representation to aid OOD detec
ariables which is amenable for (P-)QR operations in the SRF propagation and update and prevents unnecessary computation.
rack3D achieves new state-of-the-art performances, improving by 6.00% on NuScenes and 14.13% on Waymo dataset.
hieved tracking and mapping accuracy and low-latency performance of our system.
standard visuomotor policy training. Goal-conditioned policies learned with UVD-discovered subgoals exhibit significantly impro
offset when it approaches the apple near its equator.
he plane point cloud and strawberry point cloud, constructs a coordinate system on the strawberry point cloud, generates an ax
ed GrainDet, that includes over $140K$ images for three types of grains: wheat, sorghum and rice. Comprehensive experiments
of our contributions is to extend LoFTR's training method, which uses matching images of a static scene, to a dynamic scene o

ed leaf and discard inaccurate leaf instances. The experiments show that our network yields higher segmentation accuracy on s

ed without the need for inter-robot communication.

work for leaf instance segmentation. We also introduce a novel automatic postprocessing that considers the difficulty of correct
s during SWP analysis. Taken together, these constitute the core building-blocks toward enabling complete robot autonomy in p
mulated as a function and the success picking rate of tomatoes is 68.4%.
eld is achieved with only a few labeled images. Furthermore, we show that models with pre-training learn to distinguish between
n be obtained in a fully automatic manner from the skeleton graph. In the second stage, we exploit the skeleton graphs in deve

To show the improved performance, we conduct plug-pulling experiments with a quadrotor-based aerial manipulator using the
a detailed analysis of the role of our proposed cost and constraint functions for cooperative occlusion-free navigation. Second, w

} incrementally construct a 3D scene graph of agricultural environments from data taken onboard a mobile robot. We validate a
s and point clouds in comparison to frequently employed solutions within the ROS ecosystem.
the tangent direction of the picking point by estimating the pose of the sweet pepper peduncle. Finally, the robot and proposed
le to perform automated transplanting accurately, but there was an occasional problem of offset error from 0. It was confirmed
sult using Mean Squared Error (MSE) and Standard Deviation (SD). We are able to achieve accurate results with the MSE valu
s. Though not fully robust and requiring improvements in throughput, our system is the first to operate on fruit trees and represe
xperiments in agricultural fields help validate our method's efficacy in acquiring accurate by-tree information on-the-go and in re
g accuracy compared to state of the art and ablations. We also provide quantitative results on data collected by a real robotic s
ay sampling, which directly estimates the local gradient direction for viewpoint planning to overcome occlusion and improve perc
ach task, it is feasible to perform transplantation of saplings and harvesting of leafy greens using a cobot, without task-specific
mproving its performance on various downstream tasks. Additionally, we introduce the novel Fraunhofer Potato 2022 dataset co
ance of the crop-weed segmentation task in new fields and on different robotic platforms without additional manual labeling.
ption stack, a grasp planner to pick up multiple logs at a time and an extensive experimental campaign conducted outdoors, on
ng and optimizing results. Different cases of demonstration represent different situations and potential issues.

ed on real datasets. The technique is able to identify and leverage hypotheses which show a medium or strong correlation to re
stock under the presence of an external predatory threat, where robots act as predators. An analytical proof and empirical resul

semble a small structure.

ormance in detecting changes in semi-static environments compared to existing methods.

ment learning algorithm affects convergence speed and adaptability. Our results demonstrate that it's optimal to match the sam
hniques for improving the performance of an RL method in our problem setting: adversarial domain sampling and domain param
and classical methods to interpret the IE data and reveal subsurface objects; 3) a GPR data processing module to reconstruct
 sks interactively. In extensive experiments, we show that ERRA can complete various long-horizon manipulation tasks specifie
rasp-Anything successfully facilitates zero-shot grasp detection on vision-based tasks and real-world robotic experiments. Our

pen-source software at https://github.com/krzentner/language-world/

nstrated on several tasks in simulation as well as real-world environments.
nning of complex natural language tasks in scene graphs of virtualized real environments.
y 76.49% with higher executability compared to SayCan. Our approach enables embodied agents to follow natural language co
ents. Our code, data, appendix, and video are publicly available at https://sites.google.com/view/graspgpt.
aining and deploying deep reinforcement learning based methods in a challenging real-world environment with a swarm of robo
azebo frameworks.
ovided to validate and assess the proposed control solution.

nlinear system, reducing OCP complexity, and enhancing response speed. We provide empirical validation through two numeric

t substantial deviations. We validate our approach through testing in simulators, encompassing a 1:43 scale race car and a full-
industrial manipulator in simulation, as well as on an external force estimation task on a real quadruped robot. These experime

anoid robot robustly reacting to perturbations to keep its balance.

ased safety thresholds. These findings emphasize the promise of our proposed approach for ensuring the physical safety of hum
olution of the DS with Conic Control Barrier Function (CCBF) to fulfill the constraints. We validated our approach through exten
a GPU at runtime, making it suited for time-critical robotic control. Additionally, a neural approximation has been proposed to ac
s pass through extra dense layers, resulting in the prediction of the robot's 14 arms-joint angles. Compared to an only keypoint
nt planning algorithms and show that using our approach produces safe, fluent, and prosocial interactions. We demonstrate the
monstrating the lane change scenarios, indicate the proposed method outperforms the existing work in both predicting the hum
detector results in a notable maximum reduction of 30.09% in inference time and 26.53% in total time per frame. Additionally, th
on a 7DoF Franka Research 3 manipulator.
ected, and 7% higher robustness with 4% training data selected. Our method also significantly improves efficiency by at least 2
odes the direction the human is facing with respect to the camera coordinate frame. This allows automatic setpoint computation

t the controller reduced the peak gain of the frequency response by approximately 2 dB for various robot shapes. Robot experi
AR mission goals and the path they should follow to explore the area. Our results, i) show that first responders can rely on a SA
ndle scene changes (e.g. blocked pathways), adaptively updating the obstructed part of the global graph from traversable to no

mpling schema.
d a customized communication protocol. The results and the feasibility analysis show the possible applications and limitations o
— we fit sensor data into mentally constructed meaningful worlds). With respect to ChatGPT’s ambitions, it can imagine p
s, including an object-based system and a pretrained scene classifier, in experiments in a photorealistic simulator.
dings indicate that LLMs significantly improve the grounding capability, especially for complex language queries, making LLM-G
ons. To conduct a comprehensive evaluation, we collect data in a variety of driving environments and perceptual conditions and
ork (Hi-VoxelNet) with Hybrid Voxelization that leverages the advantages of cubic, cylindrical, and spherical voxelizations for hy
 lts compared to fully supervised baselines while using less than 0.3% of the ground truth labels, paving the way for learning com
notably outperforms state-of-the-art methods on both Dark Zurich and Nighttime Driving datasets.
mic Representation Enhancement Module is proposed to dynamically remove noise in the input representation via Dynamic Gro
and a Turtlebot with a RealSense camera suggest that Lifelong LERF can persistently adapt to changes in objects with up to 91

of-the-art semantic mapping quality within 2D-3D networks-based systems and matches the performance of 3D convolutional n
oor and outdoor environments, which enhances the versatility of the proposed exploration algorithm. We validate our exploratio
ainst leading zero-shot methods highlight Open-Fusion's superiority. Furthermore, it seamlessly combines the strengths of regio
o merge multiple semantically similar instances, even if they are spatially distant. To this end, we regress 6-DOF bounding box
ation model to 4D by reusing queries that decoded instances in previous scans. This way, each query decodes the same instan
ediate LSP to retrieve conducive contextual memories from the GSP. By utilizing the MP3D dataset, the experimental results in
generated dataset is then utilized to train the proposed CECI model and for extensive validation of the proposed method. We e

bling the features extracted from RGB images to contain both semantic and spatial geometric information. The unsupervised de
tion benchmarks within specific few-shot segmentation (i.e., 2/3-way 5-shot) settings. Experimental results demonstrate improv
ormance on the new moving instance segmentation benchmarks, including diverse environments, and provides model-agnostic
ntic fusion to an indoor object search agent improves its success rates.
res from onboard sensors. The BEV-level fusion mitigates the coordinate differences between features obtained from onboard
ngle modality or complementary masked modalities. We achieve state-of-the-art performance over three RGB-T semantic segm
onstruct a multi-layered scene graph that includes lane information, the position of static landmarks and their assignment to cer
ate that the suggested methods enable a broad BEV perception with remarkable performance gains.
paths for navigation. We validate AGRNav's performance through benchmarks in both simulated and real-world environments
urious radar points. We build a new dataset in the real world. Extensive experiments show that mmEMP achieves competitive p
inspired by self-driving companies and research institutions. Extensive experimental results of representative algorithms on nuS
lity. Secondly, as different modalities exhibit varying levels of importance under diverse lighting conditions, we propose a novel
ion even for those points not in the FOV (field of view) of the camera. In this way, RGB images are not required during inferenc
modality within the feature space. The model retains the style information in the shallow layers while simultaneously fusing the h
ate-of-the-art model to a new height. Code will be made available.
SuperFusion on the nuScenes dataset and a self-recorded dataset and show that it outperforms the state-of-the-art baseline m

considering temporal and spatial correlations between measurements. The graph structure is designed for flexibility, making it
reflect the motion confidence, and further segment the moving object through adaptive threshold on the time image. In tempora
n improvement of 1.24% in mAPH value(L2 difficulty) and the Argoverse2 Dataset with a performance improvement of 7.40% in
ethods by reducing the mean absolute error (MAE) of depth estimation by 25.6% and 40.2% on the challenging nuScenes datas
bjects, which is required for training the spatial anticipation ability based on semantic information. Additionally, an algorithm that
its enhanced robustness in the presence of noise and low-lighting conditions affecting the RGB images.

ents. Firstly, we propose an attention-based interactive fusion (AIF) module that applies spatial priors of events and frames as t

ur method reduces the median localization error by 71.5 % in translation and 65.5 % in rotation and is robust to mapping failure
nsation module. It alleviates the disturbance caused by complex scenes as well as the particularity of the robot platform. Experim
 on uses a crude optical proximity sensor, the influence of reflectance is minimized. Analyses and real-world experiments confir

mic contact information. Through the grid-based multi-material additive manufacturing technique, the affordability and productiv
nables our approach to estimate the direction of arrival (DoA) of sound based on TDoA information computed by our robust-TDo
ent with a small single-board computer.
nd physical robot scenarios reveals that our method successfully replicates unseen fonts and styles while achieving integrity in u
nvironments.

effectiveness of our approach in controlling a quadruped robot in confined, rough terrain. By achieving this, our work extends t

nd recover a quilt over an infant.

cision and generalization. Experiments in simulation and real-world verify the effectiveness of our method in speed (maximum 4

ilitates the robot to be an assistive agent, allowing it to either accompany doctors or visit patients individually. Ultimately, ARIS
an have a large carrying capacity.

s. The generality and effectiveness of the framework are evaluated in real world experiments where it solves various perceptio
considered optimization techniques. Experiments have been conducted considering the cases of robot deployments. The statist

esults demonstrate that SocialGAIL outperforms baseline methods in terms of goal-reaching, intermediate state faithfulness, tra
dar sensors using DRL. Finally, we conduct extensive Sim-to-Real experiments to confirm the realism of the simulator, by deplo
d (FEM) to model deformable solids, the Incremental Potential Contact (IPC) approach for frictional contact and a robust secon
vortices and corresponding flow fields generated by real fish kinematics were found to be more stable than those arising from s
ld data, with 5% performance gap. When finetuned with real data, the model achieves 14% higher average precision than the m

ons. By leveraging expert knowledge gained from the real-world data, the simulation model is manually tuned to replicate comp
keypoints are further used for efficient scan-to-map registration and mapping. Experiments on KITTI dataset demonstrate that
vided). The dataset consists of 16 sequences in a large car park and 6 long-term trajectories around a university campus that p
ttps://github.com/SenseRoboticsLab/DISO.
P codes for the community.
eometric self-similarity as well as obscurant occlusion. For the benefit of the community we release the datasets presented: http
We show that our system can adapt to different camera configurations and allows real-time execution for typical robotic applica
at href{https://github.com/SJTU-ViSYS/Ground-Fusion}{https://github.com/SJTU-ViSYS/Ground-Fusion}.
enhance the stability, performance, and applicability of neural SLAM in real-world scenarios. Project page: https://hero-slam.gith
nces, demonstrate that our method outperforms recent NeRF-based SLAM method in both localization and reconstruction.
pose-only measurements, making it general for any feature-based SLAM. We also give a practical example of how our backend
equence. This yields a real-time application of our approach for a typical robotic 3D LiDAR sensor. We perform extensive exper
balance the time consumption and the correctness of representation. Based on the extracted and maintained local geometric i
ed to existing alternatives. Moreover, we demonstrate that our map filtering methodology achieves higher similarity rates than t
ower. We evaluated IPC on standards benchmarks against several state-of-the-art methods. Although it is simple and relatively
cross diverse computer vision tasks, including image classification, semantic segmentation, and stereo matching. Taking into a
cts reality. Next, we use object information to enhance the data association of feature points and consequently update the map.
ose features, thus significantly reducing the model complexity and boosting the speed in edge devices. Built upon the novel fea

blicly available datasets justify our improvement. We release our pipeline as an open-source software for future research in this
gorithms utilizing the mapped wall planes and free-space clusters. We tested S-Graphs+ on multiple datasets, including simula
database images per query, as opposed to just a single match. Practical code examples in Python illustrate to prospective pra
ocity while managing Doppler velocity uncertainty. Based on our observations, we propose a dual imaging radar configuration t

r and stereo configurations on EuRoC dataset has reduced by 45.3% and 36.2% respectively.
https://github.com/davidskdds/DMSA_LiDAR_SLAM.git.
oint cloud map based on the spot uncertainty model of LiDAR points. The experimental results show that our method outperfor
mates trajectories competitively with state-of-the-art LiDAR SLAM methods, while also producing dense maps competitive with e
AR-inertial odometry pipelines but is significantly faster in computing the odometry. The source code of our implementation is pu

to evaluate various Dynamic SLAM formulations.

R(-inertial) odometry. Experiments on challenging environments such as corridors and forests demonstrate the high accuracy a
r theoretical properties such as the null spaces and sparsity. Based on these theoretical results, this paper develops an efficien
which necessitates their interaction with the ground. We validate the effectiveness and robustness of DORF across various type
online the triangle mesh of large-scale scenes, just relying on a standard CPU without GPU acceleration.
ry of emph{certifiably optimal} sensor designs in practice. Results from synthetic experiments reveal that sensors placed with O
the proposed framework into a modern feature-based system, we introduce HPF-SLAM. We evaluate the system on EuRoC-M
camera-object relative pose as the measurement model. To verify the performance of the proposed algorithm, we present the
he minimization of the largest eigenvalue of the adjacency matrix, so it suppresses the upper bound of the maximum degree of

olely depends on the UWB transceivers mounted on-board. In addition, an extended Kalman filter (EKF) SLAM is used to solve

r MW spaces, including open spaces, demonstrate that the performance of the proposed method is comparable to that of other
t our method outperforms the baselines across all three datasets, showcasing robust performance and notable generalization c
overing office and home environments. It successfully categorizes rooms with varying layouts and dimensions and outperforms
iorate and integrate errors, resulting in accurate camera pose estimates. Though our framework retains compatibility with tradit
be directly used in downstream tasks such as navigation or exploration.

ciency in methods involving reconstruction tasks. We deploy our method on a UAV and the results show that our method can pl

ssian process regression with a physically motivated prior. Its performance is evaluated on the MVSEC dataset, where it achiev
 (OC), or other consistency enforcing techniques. This leads to greatly improved estimator behavior for significant error and une
hat the proposed method effectively enhances the localization accuracy of odometry, thus confirming the effectiveness of the m
Moreover, these non-cooperators play a game with the DRL-based navigation policy, then we can improve the robustness of s
ework at scale, we release the MESSYSURFACES dataset which contains images of 70 real-world surfaces that need to be cle

onomous robot navigation.

d and interpret the relationships between agents over time and space, thereby making more informed decisions. Compared to p

us on the cognitive and emotional drivers behind these interactions.

d method can achieve a mean success rate of 92% on the user's natural language-described preferences after receiving 33 rou
ncept to different types of robotic systems and electromagnetic devices, further enhancing the capabilities of reconfigurable ele
ere conducted on a two-segment cable-driven continuum robot prototype, confirming the effectiveness of the proposed planning
hen the proposed gripper is controlled to the ultra-sensitive state, a flying ball with only 1 m/s could also be captured. This work
he soft actuator. Experimental data show that the actuator was flexible during bending and the stiffness can increase more than

r, the sample efficiency also allows us to learn directly on the hardware within 3 hours.

g the capability of a closed structure gripper to adapt mechanically to fruit-shape variability. Moreover, a shelf life assessment o
miniature camera on vertical walls to perform detection and surveillance tasks. This work provides a reliable structure and contro
me by dynamically varying the size of the robot. The crawling and bounding gaits enable Tetraflex to move in smaller incrementa
ntrol is presented. A prototype soft manipulator was developed to validate the proposed design experimentally.
on a machine learning model. In addition, our design successfully demonstrated its application in fruit sorting, which proves its r
to reduce the inertia of the manipulator. The information from the cameras and force sensors is used to estimate the extent of t

y substantially improves the maneuverability of pneumatic soft crawling robots, thereby providing new routes for the potential a
% (17) being successfully picked and 38.5% (10) sustaining damage. The movements of the robot arm in the harvesting seque
ontrol system. This circuit system can be powered by a small lithium battery and output a periodic voltage to drive the motion of

tors to precisely control all antenna positions. We characterize the response time, accuracy, and repeatability of a single actua
h have high-resolution tactile sensing. This design also potentially helps researchers discover and explore more soft-rigid tactile

king principle, fabrication, and optimization of a robotic gripper based on this shape-memory silicone sponge. This sponge-jam

on. A series of experiments are conducted to validate the effectiveness of the proposed method and the robot's capacity to acc
ch is further used to construct a surrogate model for fast prediction of target responses. To achieve an optimal actuation capab

ng angle of 240°. The robot can also increase stiffness by increasing internal air pressure to resist deformation caused by ext

MAS = 2, validating its on-body functionality.

er laminates. Finally, an extensible soft gripper based on flat tube actuators and a robotic wrist module are developed, and thei
ed mechanism’s straightfor- ward structure facilitates easy integration with a wide range of tip mounts, steering mechanism
ver 120% axially, but only 3% circumferentially. With the addition of controlled layer jamming, this film enables a steering metho
of each segment. Results show that a single module can bend up to 30° omnidirectionally, its stiffness can increase up to 95%
, a feature uncommon in previous designs. The findings presented in this work are key to the development of larger scale conti
ly, we conduct experiments on a linear stage to validate the 3D contact point localization capability of the proposed whisker sys

outlined. Its possible extension and further use are also discussed.
-soft hybrid robotic glove with force sensing function. The test results show that the robotic glove has good fit and high output fo
meters for each application. Using rapid manufacturing processes the optimized actuator is constructed and physical testing vali
ared to the most flexible stage (100% alignment). Additionally, the human subject testing demonstrated a stiffness increase of u

d sustaining grip force experiments, respectively. The results suggested that the proposed method could provide a new approac
of external force disturbance has 11% relative error within a 5 N range. The combined sensing and modeling approach can be
contact-rich tasks, and enables imitation of tasks where the demonstration and imitation is performed with different body-schem
) demonstrate that our dynamic model can precisely describe their complex dynamic responses and the EERLSMC can achiev
puts to predict applied tip forces. Experimental validation was performed using the STIFF-FLOP robot, a pneumatically actuated
ctively, using feature extraction and a deep neural network. The proposed actuator has two degrees of freedom (i.e., bending o
ely simple and low-cost. Potential applications include solar tracking, and in the future, firefighting smoldering fires. We envision
f the principal curvatures and the minimal distance between surfaces as penalty functions. The effect of external forces includin
oncentrations, we found that the lowest crosslinker concentration yielded the fastest response time of the valve at an ambient te
n. Overall, the sensor represents an improvement towards minimizing electronic hardware, and cost of the sensing solution, wit
can be switched at runtime to handle highly diverse control tasks. This paper presents the architectural concepts as well as firs
a CLI tool for simplified interaction with container images during the development phase. Within the scope of this paper, we em

e-based gas dispersion simulation. To illustrate GSL-Bench's capabilities, three algorithms are compared in six warehouse sett
TL in a realistic simulation environment (AI2-THOR), and evaluate its performance across a series of cooking recipes. We dem
we demonstrate how our system can autonomously program robots for various assembly tasks in a real-world project.

e images, and in real time injects corrections into the local cost map, thus enabling the robot to traverse areas that would otherw
g integrity. This work demonstrates how machine learning-driven processing can enhance existing data products to maximize th
opens new possibilities for controlling floating platforms in zero-gravity settings, promising advancements in space exploration.
y can be derived. Each MU then seeks to follow its assigned trajectory by using a decentralized PID controller that outputs the M

h a series of experiments in both simulated and real-world environments. The source code, along with recordings of real-world
e a new dataset, SPADES, comprising real event data acquired in a controlled laboratory environment and simulated event dat
tions. Functionality includes constant-time queries for height, obstacle presence, and slope details, boasting improvements in r
 comotion method is elucidated based on transition support points. Finally, simulations and experiments are conducted.

dations of these models and field tests of the system on slopes, gravel, rough grass, and on water.
er and data were tested for water resistance and resilience to environmental factors, as well as power and data transfer.The ob
aring platform can successfully perch on a flat surface with a velocity of up to 72.5 mm/s and with an approach angle misalignm
me facing up under the action of equivalent pendulum during the locomotion, which makes it possible to carry workloads such a
using a planar example. The criteria considered include: (i) Design Complexity, (ii) Trajectory Errors, (iii) Required Torques, an

m and cadaveric soft tissues with the da Vinci Research Kit. The results demonstrate the robustness of the method to occlusio

either using the recurrent memory layer or are trained with a difficult physics environment. Further project information is availab
cs that enables the automatic generation of large and diverse datasets for 6D pose estimation of surgical instruments. Among th
y possible. We introduce Surgical Gym, an open-source high performance platform for surgical robot learning where both the p
stic, we also design a cross-task pretraining paradigm that uses multiple training objectives associated with data from diverse t
orm for procedures like biopsies. In addition, a tailored master device has been developed to control the extensible bronchosco

mporal compensation. We evaluate our approach on the public SurgT benchmark, which is generated from Hamlyn, SCARED, a
sidual gap between the simulation and observation and improves the simulation's predictive capabilities. The effectiveness of th
ental results show that our method efficiently detects relevant subsurface anatomy with shorter trajectories than the comparison

dels achieve higher accuracies and edit scores than MP recognition models. But, using MPs enables the training of models that

peting registration accuracy similar to other works. Results indicate that our Iterative PnP is suitable for future vascular interven
on and hardware forward kinematics accuracy and precision are characterized by errors of 14.74 +/- 8.87 mm or 26.61 +/- 17.0
ble curvature and stiffness. Experimental results agree with the model, paving the way for use of this model in design optimizat
acquisitions, respectively. Our results yield an average RMSE ranging from 0.58 ± 0.21 mm to 0.66 ± 0.20 mm depending on
etter further proposes a novel integration of magnetic actuation and magnetic localization in MSRC systems, where fluoroscopic
edict distributed external forces on the articulated manipulator body with an average 0.09 N error. FSO can also estimate loads
rmed in different breath-holding positions on a respiratory motion simulation platform, and experimental results show that the m
formation flight and reach a target while maintaining flock formation.
nly minor parameter adjustments. Our findings demonstrate the potential of this approach as a flexible solution to tackle such ta
action between swarms dislocated across multiple labs worldwide, encouraging scientific collaboration and advancement in mu
orphogenetic engineering experiments. The results revealed that swarms of Morphobots can form patterns via physical interacti

ctions communities and demonstrates various collective behaviors with strong potential to be realized in macro- and micro- len
ow that automatic design can leverage mission-specific interaction strategies to enable an effective coordination between the tw
fboard control algorithms and high-level fleet commands. To validate the proposed simulator, we provide a case study of decen
rate meaningful generalization to harder transport scenes with training only on simpler versions of the task.
 Franka Emika robots we then show the effectiveness of our approach using model predictive control in real world experiments.

e robot backtracking. Then, feasible trajectories with adequate obstacle clearance are generated by topological graph search.
eness of the resulting solver and that a round robin baseline strategy used previously for TAMP becomes a special case. Expe
oked in prior research, markedly enhances the system efficiency by obviating the prolonged measurement pauses and leveragi
ion-free solutions, while outperforming the current state of the art on metrics of success rate, planning time, path length and nu

sons, we show that an agent utilizing historical multi-modal information surpasses all the compared methods that neglect the his

source TSP solvers OR-Tools.

pick-and-place tasks, and it was compared against state-of-the-art approaches TM-RRT and Planet, showing better performan
nes when the budget is tight by collecting measurements leading to smaller reconstruction errors.

and efficiency of robot use using the dataset for contact on two examples. Finally, our experiments provide a set of injuries that
ement domain, where we demonstrate performance benefits relative to a state-of-the-art task planner.

benchmark problems and comparing the results obtained using the proposed method with those obtained using other state-of-
, which easily leads to repeated exploration and reduces the overall efficiency of the system. To solve these problems, we prop
We evaluated the proposed interface and the related teleoperation framework with naive users, assessing the teleoperation perfo
ion by up to 42% in the force precision and 63% in the success rate, demonstrating the potential of the proposed approach to e
the physical interactions with the manipulated objects or the environment, whereas the very low weight of the arms (1.0 kg LDA
eep reinforcement learning (DRL) agent is trained and deployed at the operator side to seamlessly switch between autonomou
es to assess the system's effectiveness, including a performance evaluation of key components within DTMR and a quantitative
d, showing that for an online estimated OBM, the accuracy of the force feedback was improved by up to 44 percent while using

nd high kinesthetic feedback output satisfying requirements. Further, POstick-VF has strength with unlimited workspace at the e

building at least a tower of 3 cubes in 10 minutes, with only 5 minutes of training beforehand, regardless of the control modality

se the physical plausibility of the placement as well as contextual soundness while considering user preferences. Code, video,

in a simulation with various assembly tasks to show the effectiveness of online estimation.
g up to 105% of the robot's weight. By simultaneously adjusting their pitch and using their arms, the pilot adjusts the robot pose
 n-robot interaction. Finally, we evaluate our system subjectively and objectively through human-subject experiments. Autocomp
ur method to different fine-tuning approaches. We construct several DA variants and validate them through extensive experimen
ncourage consistent predictions for each object in the semantic mask. The knowledge learned from modal-specific prior is then
licy from the source domain to the target domain without the need for additional fine-tuning. Our approach has been tested thro
on. The implementation is available at https://github.com/auniquesun/PPT.
ction to construct depth-aware information, extracting target domain-specific knowledge in Voxel and BEV feature spaces. It th

ess. To implement our idea, we designed a novel multi-level residual network with a filter module attached, where each level of
daptability to the target domain. The efficacy of the proposed method is demonstrated in reaching tasks and peg-in-hole tasks c
e dance show where dancers encompass individuals with disabilities and able-bodied individuals. Dancers can practice dancing
. In addition, the proposed foot performed as a mechanical filter by reducing the rate of change in the GRF. Furthermore, the re
ility of this robot.
ped with a small manipulator or a flexible grinder, the robot can perform operation, detection, or machining tasks while adsorptio
ome hardware problems caused insertion motion failures and checked the manipulability of the manipulator. Our experiments co
which is highly capable even in the absence of complex algorithms. The future incorporation of more advanced sensors and te
rooms and through non-metal occlusions, which is a clear advantage over cameras, mmWave radar, Infrared sensors, etc. It h
ut also critical situations. The cooperative effort between humans and machines ultimately leads to a turning point in environme

s, especially in real-time and diverse settings. We propose a universal 3D human recovery engine with a well-designed robust m
s. Implementing Deep Learning across fields such as HRI opens novel avenues for achieving heightened efficiency and friction

beliefs as on-demand explanations in response to different types of questions.

nalyze the behaviour of the moving target and generate the trajectory. Alternatively, an map (optional) can be integrated to prov

, successfully tracking the target moving horizontally and vertically. The AUV succeeded in estimating with an RMSE of 0.23m a

m is optimized for persons with dementia through large, easy-to-read/press buttons on a touch screen, spoken instructions, and
ing the clip delivery system and another for managing the transesophageal echocardiographic probe. Key technological advan

uman-likeness. Our method was implemented on an actual human-robot setup. We extracted dynamic human features and gen
upporting pivot, resulting in a sideway lean and subsequent turn. The third method introduces a torque optimization technique r
nment that replicates lunar conditions. This video introduces the blockchain-based Multi-Robot System (MRS) for distributed m
ping control by detection of object contact with an embedded load cell and the prediction of a desired open hand movement from
ner is parsed in Unity and visualised for the operator with a projector. When areas are clear, the robot continues its task autom
dges. It also has wheels for ground movement and to assist with transitioning over ledges and corners. EEWOC has demonstra
n real-world scenarios within a pilot home for the elderly. Collaborations with nursing schools and caregiving companies such a
nt. In the second part of the video, we demonstrate the combination of the nitinol-embedded soft robotic gripper with resistive b
markably, it survives a 20-meter freefall without sustaining damage. The incorporation of onboard sensors enables real-time sta
nd half, Slip-in Manipulator, can generate motions to travel under the weight of the human body and can generate bending moti
126Mpa, and remained undamaged and continued to function normally in the laboratory. Currently, the buoyancy regulator has
g the winding method of tendon, which make the load for each degree of freedom can be shared by two motors at the same tim
ollaboration with the robotic arm, from 40 to 135 degrees. Three levels of stiffness were investigated and surface electromyogr

ndoscopy. Our second technical contribution is an NRSfM method which exploits a sparse surface normal set. For that, we prop
differentiable rendering with the Kalman filter. This synergy allows for simultaneous estimation of the robot's joint angle and pos

vel MOT benchmark specifically tailored for the agricultural context, based on video sequences taken in a table grape vineyard,
he number of UWB anchors, i.e. adding them to the factor graph when their ranging measurements are available and discarding
roach for determining the success of the initialization by evaluating the residual of the normal epipolar constraint. Extensive eva

more efficient at goal-conditioned exploration. Finally, we show sim-to-real transfer to robot hardware and XGX performs over t
ad a higher mean average trajectory error and became unstable in these challenging environments. This paper makes an impo

achieve smooth estimation. A high-performance accelerator that optimizes the efficiency of using heterogeneous devices is app
ty. At the same time, the nonlinear MPC suggests optimal throttle settings, striking a balance between goal attainment speed a
ef, CMN extracts a global heuristic vector that guides the planner to find a local navigation action. Empirical results demonstrate
e seamlessly integrates 2D and blinker markers within a single marker unit.We have developed prototypes of 2.5, 5, and 10 cm
odel to offer multiple tradeoffs between accuracy and speed to suit different applications. Experiments using a Jetson Orin NX e
g benchmark datasets, Touchdown and map2seq.
a 76% higher success rate in indoor and 23% higher in outdoor navigation tasks with higher computational efficiency.
tor with a remarkable zero-shot navigation capability. We extensively evaluate our method using the AI2-THOR simulator and c
on, we introduce new performance metrics that offer insights into the efficiency and effectiveness of various agents in object go
cess rate in the local path-planning task. To perform long-horizon object navigation, we design an LLM-based planner to utilize
ime adaptation approaches. Our code is available at https://github.com/csiro-robotics/GeoAdapt.
rrains and accurately identify obstacles. The semantic information is represented in 30 classes using an RGB colorspace that c
efs. We compared our method with complete coverage and random walk methods, showing that our approach is more efficient
nertial-proprioceptive-tactile measurements from the robot’s observations to a representation space and performs a nearest
rrect global image with statistical features, which helps guide the model to refine sharp texture and structure. Extensive experim
nly demonstrates the cross-category flexibility and instance awareness, but also show remarkable robustness to domain shift an
wledge, RenderOcc is the first attempt to train multi-view 3D occupancy models only using 2D labels, reducing the dependence
n cost-volume-based methods. In particular, our self-supervised method outperforms previous supervised methods on the KITT
o-shot goal-conditioned manipulation, underlining its practical applications in robotics.
boLLM not only outperforms existing baselines but also substantially reduces the engineering burden associated with model se

D depths of the environment without extra labeling work or 3D data. Extensive experiments on the KITTI dataset, KITTI-360 dat
ntinuous fashion and eliminate the temporal layering problem. Results from rigorous validation through quantified metrics at all s
ultiple time steps using cross-attention-based fusion modules. We evaluate our method on the physically-realistic cluttered bin-
s research.
t a novel dataset on endoscopic submucosal dissection for surgical activity tasks. Extensive comparisons and ablation experim
 elf-supervised losses on real-world data in two additional learning stages. By adopting this comprehensive strategy, our method
xperiments are conducted on a new dataset featuring both image frames and event data captured under extreme conditions. Th
pability of modeling dynamical systems. Overall, this work is a significant step towards developing advanced controllable video g
or encourages the fused images to resemble visible images holistically, while the other ensures that the targeted objects in the f
point cloud once and automatically adapting to all surrounding cameras. Our refinement balances the accuracy and efficiency

on and the Street Image Generation. The Neural View Transformation phase converts the BEV map into aligned multi-view sem
mprovement in rendering speed. Codes are available at https://github.com/VISION-SJTU/Lightning-NeRF.
ficantly facilitates other research in the related fields. The code and dataset will be publicly available at https://github.com/Zeror
f-supervised network alongside an economical single-beam LiDAR, striving to achieve or surpass the performance of more exp
ge. We believe our work lays a strong foundation for future exploration within the research community.
tion task on the Cambridge Landmarks and 7-Scenes datasets demonstrate the superior predictive accuracy and benefits of ou
ns of the model. Experiments demonstrate the effectiveness of our method in various simulated and real-world tasks, confirming
d bimanual manipulation tasks from a small number of human demonstration videos. Videos and source code are available at h

package for training and testing machine learning-based view planning frameworks, custom view dataset generation of arbitra
ecover corrupted messages in situations where CopperTag data is incomplete. The experimental results illustrate that CopperT
raphs between agents, leading to accurate relative pose and time. We validate emph{FreeAlign} on both real-world and simulat

uded images. Thus, we further propose a 2D matches based sampling strategy to preclude the occluded area. Experimental re
module, considering the fusion of global and local elements. The experimental results reveal that the proposed model outperform
ti-frame self-supervised monocular depth estimation methods. Adding MAL into previous state-of-the-art methods leads to a re
es these high-quality features to produce clear results. Additionally, a controllable operation is introduced to improve feature m
FV images and the BEV semantic segmentation. Extensive experiments on the nuScenes dataset show that our semi-supervise
world datasets, which demonstrate that the system significantly improves annotation efficiency compared to manual annotation
ule, which incorporates event camera data into a cross-modal learning architecture based on SAM, with only limited tunable par

ve cross-modality knowledge transfer while preserving the unique strengths of both modalities, utilizing a spike-oriented uncerta
ge radar data as supervision. Our experiments demonstrate that PillarGen outperforms traditional point upsampling methods in
rk follows an effective coarse-to-fine procedure to evaluate the FoV visibility over the target frame. In addition, the deep Q-learn
orm precise Cobb angle measurement. Furthermore, we devise joint loss functions for our segmentation and regression netwo
images) from 162 different vehicle models of 43 makes depicted in front view. It is split into 8 subdatasets to investigate the in
ves the accuracy of the keypoints prediction. Then, the predicted contour of the previous stage is taken as geometric prior that
envision that the soft acoustic end-effector will facilitate multiscale acoustic manipulation in interdisciplinary applications, espe
s available at https://www.youtube.com/watch?v=cx5THPyIKjA
nthetic datasets, highlight the competitive performance of our method in terms of recall and accuracy. Furthermore, we demons

al tissue tracking on the SuPeR dataset as well as in the context of relative pose estimation on a repurposed version of the SCA
model (LLM) and an adaptive weight assignment methodology. We first develop a new multi-teacher CL framework that levera
ation accuracies exceeding 99.27% for instruments and macular holes and an average accuracy of 98.97% in recognizing surg
 the Atlas humanoid robot, reaches a target from an unknown starting configuration, demonstrating that this controller achieves
e centroid of both the hand and object, the deformation of the object, and the occlusion rate of the hand. Furthermore, to preve

tation using quaternions, generate new interaction matrices, and conduct comprehensive comparative experiments in simulate

aches. We validate the efficiency of our approach through numerous experiments with elastic deformable objects of various sha
nsive experiments show state-of-the-art performance for DCPT on multiple dark scenario benchmarks. The unified end-to-end l
our research represents a step forward in quadrotor visual tracking, moving from task-specific methods to more versatile and a
eme is introduced, considering the concurrent movements of both the drone and the objects. Within MDA, an Adaptive Feature

te their boxes from frame to frame using a flow-based motion model, refine the propagated boxes with the box regression modu
mation learned at different stages and ease the optimization, for the search area, we densely connect the initial input sparse pill
we propose to separate label-making and training into two distinct stages. In the first stage, we use the pre-trained model to est
mproved tracking performance.
zation uncertainty estimates from probabilistic object detectors. Experiments on the Berkeley Deep Drive MOT dataset show th
obots of the same family.
mark, surpassing all preceding efficient trackers, while running over 100 fps with ONNX on the Jetson Orin NX edge device. Mor
reducing cumulative positioning error in agricultural open fields.
heel or ground changes and improve the accuracy of prediction under new control inputs, but can even improve tracking accur
nce range to update candidate points' inverse distance. This adjustment is based on the convergence ratio of the inverse distan
so perform online calibration of GNSS-IMU extrinsic parameters to mitigate the possible extrinsic parameter degradation. The p
imulated and real-world datasets to evaluate our method. The results demonstrate that the proposed method outperforms SoTA

urthermore, we demonstrate the effectiveness of our pipeline through onboard closed-loop quadrotor aggressive flight and larg

ssion SLAM) on Hilti SLAM Challenge 2023 with 1.7 times the score compared to the second place.
with impaired vision safely along the current track. Our system is deployed on an embedded device, Jetson Orin NX. Through
ssistance. Moreover, experiments conducted with three healthy subjects showed that the VGFL reduced the surface myoelectr

keleton weighs 3.9 kg with a maximum torque of 20 Nm at the knee joint. The paper describes the mechatronic design, mathem

is fed back into the ProMP iteratively to enhance the individualized features for as long as the training process continues. The
ers a promising solution to prevent low back pain during weight-lifting tasks and provides guidance for future back exoskeleton d
and cost-effective motion analysis systems.
k, we establish the mathematical expression of the HELIE transfer during exoskeleton assistance. This will promote the researc
ee in both the Sagittal and Coronal planes, with the aim of minimizing joint misalignment without the use of inherently soft mate
c terrain and achieve prescribed locomotion tasks for a wheel-leg vehicle on sinusoidal terrain in simulation.
e a simple LQR-based controller. The approach is validated in real-world experiments where we show the execution of trajector
of CPG control in transformable wheel-legged robots.
he planned trajectory as safe if rollover and slip are not predicted and unsafe otherwise. The proposed PTCM is demonstrated
special cases with particular servo constraints. Structural deformation control laws are developed and experimentally validated
f a positive definite Lyapunov-like function semi-negative definite. In the next step, another control input is designed by the bac
ith by a naive approach that keeps the feet still during the airborne stage. To the best of our knowledge, this is the first time tha
ting the necessity of whole-body optimization compared to the baseline locomotion MPC when interacting with heavy objects. E
a physical robot.

emonstrating its effectiveness in collision avoidance and whole-body manipulation tasks. Project page: https://sites.google.com/

ances the whole-body telelocomotion performance in unfamiliar environments but could reduce performance in familiar environm
 iency. The dataset, code and pre-trained models have been released in our GitHub.
0 and PointSegDA for 3D point cloud shape classification and semantic segmentation, respectively. The results demonstrate th
-ID, accurately tracked each selected target throughout the experiments in all cases except for two cases. In contrast, the MOT
an be computed in parallel for all potential candidates. We conduct extensive experiments on 5 large-scale datasets to demons
art in several zero-shot action recognition tasks across modalities, significantly surpassing the vanilla skeleton zero-shot metho
d place recognition while maintaining a small model size. Code and supplementary materials can be found at https://github.com
ty of our trackers compared to other state-of-the-art trackers, with enough speed for real-time applications.
Ptextit{n}P. This approach showcases that the image descriptors extracted by pre-trained ViTs are well-suited to achieve a nota
ng the preservation of spatial information. Finally, meaningful action event frames are extracted by analyzing the flow and cluste

acking have proved the high accuracy and fast response for both theoretical stiffness and position adjustment.
h a comparable SEA implementation. We show that BSA outperforms SEA 1) in terms of maximum velocity at low final times an

ory of these springs and demonstrate their utility in performing a net-zero mechanical energy cost lifting task that requires force
clearance. Simulated and field test results demonstrate that the approach generates obstacle-free, smooth, and feasible trajec
nts are taken into account to plan a flyable trajectory. The problem is formulated as the multi-step optimization model that acco
cedastic uncertainty. The open-source code is released at https://gitlab.com/ReI08/perception-energy-planner.

ompared to the pre-trained baseline, requiring only SI{22}{second} fine-tuning on an ultra-low-power GWT GAP9 System-on-C

control of a manipulator subject to base moving disturbance.

e drone reaches 100% reliability at 0.5 m/s in a generic and previously unexplored indoor environment.
cal robotic sketching. Furthermore, we demonstrate our robotic sketching system’s capability to generate complex images

platform, and keys to success in terms of its scalability and flexibility.

ons to caption the various sources of disturbances, and 2) Smoothed Streaming Meta-Learning, which learns to capture the evo
and refine target coordinates based on uncertainty variance computed by the object detector. Finally, the integration of PPM an
e devices such as NVIDIA-TX2 or Xavier NX during autonomous navigation. Integrated into the robot navigation system, extens
erate syntactically correct problems with more than 99% accuracy and valid plans with more than 58% accuracy.
imator by up to 68.1% in force estimation accuracy, utilizing only 1.4% of its network parameters. Furthermore, our method sho
ED panels, MCDEN demonstrates competitive results with a 99.2% detection rate and rapid real-time detection capability.
s various articulated objects as a unified textual description containing kinematic joints and contact location. Building upon this u
or. We examine the performance of ISR-LLM across three distinct planning domains. Our experimental results show that ISR-LL
arable accuracy outcomes.
ce compared to existing methods in modeling spatio-temporal relations, motion prediction, and identifying time-invariant latent f

motion features of humanoid robots, ensuring the agent imitates the motion data while maintaining goal orientation without add

at, we introduce an adaptive simplification technique for a probabilistically constrained setting. Such an approach provably retu
an effective tool for sensorimotor learning under uncertainty.

dels, fostering interdisciplinary collaboration between artificial intelligence and human-computer interaction.

gh-fidelity cutting simulator and a robotic simulator), learning compliant control policies on the calibrated simulation environmen
ial and envisioned goal scenes are matched to formulate robotic action policies. Experimental results demonstrate that SG-Bot
ask success rates across all four tasks and scales better to more agents. We further demonstrate the hybrid frameworks in 3D s
perform those relying solely on traditional language instructions.
performance decline. This emphasizes the critical importance of research in this direction. In this concern, we spotlight the auto
n. Our results, both in simulation and real-world scenarios, confirm that our approach adeptly manages intricate tasks that dem
fferent prompting styles across 3 robots in a high-level sorting task, and present a collection of empirical observations: (i)includ

environment and planning setups. Our experimental results show that the proposed replanner can perform on par or even bette

ment learning. We show Gen2Sim succeeds in learning policies for diverse long-horizon tasks, where reinforcement learning wit

orrect both syntactic and semantic errors via autoregressive re-prompting, resulting in significant improvements in task complet
cuting skills succeeds in only 44/60 (73%) episodes. Extensive perturbation experiments show that ASC is robust to hand-off er
e two ARMAR robots search through their LTM for past object locations in episodic ego-centric images and robot state data. Ex
aborative planning and exploration tasks involving a drone and a user equipped with a MR headset.

d experiments, including comparison with other approaches and dynamic target tracking.

ase about 26K images and LiDAR scans along with annotations. Baseline code and supporting scripts are available at https:// c

y updates roadmap information based on the environment changes and avoids dynamic obstacles. Simulation and physical exp

the controlled system into a learned latent space. Then, it constructs a reduced form of Morse Graphs representing the bistabi
 ents on a collaborative robot with 7 degrees of freedom have demonstrated the superiority of the proposed approach in both pa

nd steering angle. We extensively validate our approach using the CARLA simulator on various terrains. Our method demonstra

Scenes dataset.
and explaining, and unsafe action pruning. To demonstrate the effectiveness of our system, we conducted experiments in Virtua
n the CARLA simulator. Results demonstrate that our approach surpasses existing methods in terms of performance and robus

geneous system without the capability of sharing their knowledge. We extensively verify our framework in a simulated multi-robo
s, and the cluster-based Hungarian algorithm (CBHA) in Monte-Carlo simulations with different numbers of agents and tasks. T

as desert. Through the framework, the robots get an understanding of which goals yield the most information gain and, therefore

ds by 12.5%/5.3% in ADE/FDE on ETH/UCY, and the improvement on SDD is 6.4%/7.4%.

We integrate a scene context fusion layer in TTHMS to incorporate the scene context. Comparison and ablation experiments on
valuated with reference data in various scenarios. The results demonstrate that it accurately and robustly determines motion an
corporating visible and thermal footage alongside Controller Area Network (CAN) frames. We demonstrate that nighttime navig
roject website https://jchengai.github.io/planTF.

tperforms existing methods in both occupancy prediction and trajectory prediction in partially observable setting on the Waymo
onstrate that our method is a reliable and practical localization solution in various autonomous driving localization tasks.
distinct behaviours, and there is a challenging trade-off between them. We are able to solve both together using a novel trainin
our framework achieves efficient decision-making and human-like driving maneuvers. Comprehensive experimental results dem
s on the autonomous driving platform. The results show that our method is capable of learning the lane-changing cost function

em by leveraging the advantages of both geometric and learning-based methods. We validate this hybrid paradigm through a p

0 questions and ground-truth responses from the NuScenes dataset.

or inputs. We evaluate our approach on both the Town05 and Longest6 benchmarks using the closed-loop CARLA urban drivin
ar places. We collect offline data and conduct evaluation experiments in three real-world urban scenarios. Experiment results s

nuous representation of 3D space from low-resolution LiDAR and upsample a small number of layers to any number that we wa
ent learning (UnaRL) algorithm. The pro-posed approach, which won 2nd place in the Alibaba Future Car Innovation Challenge
terpretability and transferability of extracted skills.
UMTraDistill only decreases by 1.8%, and it can process 10 m × 10 m elevation map at a speed of 74 fps. Field robotics exp
ve modeling: Autonomous Valet Parking (AVP). Results demonstrate that the proposed system achieves comparable performa
el reward function based on region dynamic guidances, thus effectively addressing the challenges of active gas source searchin

of urban and suburban geographies (https://waymo.com/open/data/motion/). Compared to Waymo Open Dataset (WOD), WOM

ies in diverse kitchen scenes. On average, MT-ACT outperforms prior methods by over 40% in unseen situations while being m

gnals during the online fine-tuning process. In a diverse set of five real robot manipulation tasks, we show that our method can i
a fraction of the compute costs. These experiments demonstrate distillation from ensembles as an effective method for improv
drives, small angle approximations, and controls. We draw insights on current development to inform future research to better d
anding (VTOL) UAVs. This work solves the challenge of self-takeoff for large-scale flapping-wing robots using a designable me

etical and experimental turning radius. In addition, the direct relation between the vertical tail deflection and the roll rate of the o
troller for a robotic fish through only one governing kinematic model.
strategies that increased robot speed by more than 200%. Our study is a beginning step to extend robot mobility beyond simple
ly verified our platform with our previously-proposed mechanism of energy saving ('vortex phase matching') by flow visualizatio
trates the effectiveness of our method in complex missions.
cost of maintaining medium and maximum stiffness accounts for 8.72% and 17.87% of the total cost of transport, respectively. A
can significantly improve the effectiveness, efficiency, and flexibility of swarm robots in complex environments.
erform sim-to-real transfers to the Unitree Go1 quadruped, where we observe robust navigation in a variety of scenarios. Our re

alking in all falling trials. This work presents an example of using biologically inspired simplicity to solve what would otherwise be
ps, even on uneven terrain, enabling the Unitree Go1 quadruped to jump 0.5 m high, 0.5 m forward, and jump-turn over 2 rad.

tiveness and feasibility of the proposed controllers, offering substantial support for real-world applications in robotics.

rial orientation without flipping.

tonomous operation.
eferences. The measured root-mean-square (RMS) values of the tracking errors are as low as 4 mm.
g the initial learning flight has a pronounced impact on the network's performance. A higher density of sample points in proximi
esign is further verified through a series of operation validation experiments. Such a robot has the potential to provide a highly b
ormance. Through structural optimization, the maximum wind speed attained 2.18 m/s. Flight demonstrations with payloads evi
 ransmitted to the sensory area of the BNN, closing the loop. This system presents a platform to investigate how information is p
ricted extension capability of the proposed approach has the potential to increase the application prospects of continuum robots
adrotor, respond quickly with full-scale folding within 0.9 s, climb a maximum slope of 36 deg, and traverse narrow passageway
an appropriate level of body compliance. This allows the robot to 1) locomote more energetically efficiently, as evidenced by a r
on. To achieve its real-time spatial reorientation, we propose a novel quadratic programming algorithm based on a geometric m
s (on a flat surface, on low-dense saline, and on saturated saline) and yielded different module configurations. The process of s

e presented which successfully validate the observer-based output control of the SCP-driven robotic eye.
e project page: https://ootts.github.io/easyhec.

and outdoor environments using various experimental setups validate the effectiveness of our proposed algorithm. The compre
ndom calibration drift with an accuracy of about 0.03 degrees in the yaw angle.

accurate method can provide state-of-the-art results for camera lidar calibration in the field.
https://github.com/OpenCalib/LiDAR2INS.
ng LiDAR and KITTI datasets have verified the robustness and accuracy of our method which outperforms existing works. We
umbersome. To overcome these limitations, we propose eWand, a new method that uses blinking LEDs inside opaque spheres
enarios. This includes cases where the initial mis- calibration is even more severe than what the system encountered during tra
l of agreement with the ground truth and exhibit superior performance compared to the conventional PINN and the non-linear g
wed that our method can effectively perform incremental learning without resetting the rewards of the gathered observations.
ormance of our HIFMNet compared to other state-of-the-art learning-based methods.
earch documents low repeatability and accuracy of the conventional methods for robot-assisted TMS hand-eye calibration. App
rk requires only encoder measurements without using any force/torque sensors, but still maintains accurate estimation capabili
y transformer architectures to estimate accurate calibration parameters from the correlation information. Our method achieved a
xperimental results demonstrate that our proposed method is highly robust and effective in various scenes.
oint-difference calibration. (2) A new feature extraction algorithm is proposed to rapidly locate the target and extract the target c

FM's zero-shot nature enables it to be readily deployed on real-world robots such as the Boston Dynamics Spot mobile manipul
to flexibly handle both goal- conditioned and goal-agnostic navigation. Our experiments, conducted on a real-world mobile robo

ing the sim-to-real gap. Our extensive comparisons with state-of-the-art baselines and real-world implementations demonstrate
and mitigation in an integrated manner through the reward function of RL. Our experimental results demonstrate that the propo

arget rapidly without a preconstructed map.

ultiple time steps. This, in turn, amplifies the challenges associated with training. In response to these challenges, we propose P
entiating through a zonotope collision check, allowing one to integrate prediction and planning within a gradient-based optimizat

acles safely when needed.

gnificant increases in success rates to more than 50%, up to a 9.5% decrease in normalized trajectory length, and up to a 22.6
safe and efficient navigation compared to existing methods. See the project GitHub for source code and supplementary materia

posed rewiring improves the update performance reducing the size of the estimation model. We finally show its batch prediction
nized public dataset. The results indicate that JOB incurs only 1/6 to 1/5 of the computational cost associated with raw LOS che
sions without making overly conservative predictions.
predictions when one or more sensor modalities are unreliable or unavailable in complex outdoor environments. We evaluate o

ow improved planning performance on tasks that require reasoning about physical object concepts, compared to baselines tha
while significantly reducing movement cost and planning time. We further justify the generalization ability of our approach in a r
w that our method achieves sufficient reconstruction quality, outperforming several baselines under limited view and movement
than the available alternatives for LiDAR-camera intitialization. The proposed method can be applied with any combination of se
e that UM3 surpasses existing benchmarks across various categories and objects, while its adaptability in real-world scenarios
upon submission, demonstrating the current state-of-the-art performance. The implementation of our MF-MOS has been release
struction with rich geometric details.
ess to the geometric values without additional inference or volume rendering, further enhancing the efficiency. Our results show

its utility for downstream tasks such as panoptic tracking. We make the dataset, code, and trained models publicly available at

40.25% in AR, and 70.981%/58.06%/43.56% in outliers, when tested on FlyingThings3D (FT3D_S, FT3D_H) and KITTI_H data
e of understanding the scene. Experimental results on nuScenes dataset demonstrate that the proposed CVFormer obviously o
(TID) is competitive with prior arts while using 80% fewer parameters, consuming 20x less memory footprint and running 40% f
s indiscriminately. Experiments on the V2X-Sim dataset demonstrate that ActFormer improves the detection performance from
surpasses existing methods in generating tasks on the KITTI-360 and KITTI-Raw datasets, as well as in the completion task o
speeds. Our rigorously optimized network, when tested on the Cityscapes-3D datasets, consistently outperforms various baselin
on the large-scale nuScenes and the Waymo Open datasets for both 3D detection and semantic segmentation tasks, and it ac
ween the finger and object can be regulated precisely (5% RMS error) in realtime.
ve real-time sensing during the in-hand object pivoting.
em in estimation accuracy. Moreover, this compliant haptic method works effectively even when the manipulator's approach ang
ntact estimation and the effects of online estimation on MPC performance.
s TacShade’s ability to reconstruct the shape of the contact objects through one image for object distinction to avoid the lo

er to implement. Real-world experiments are carried out to validate the proposed method with two differential-drive robots, Hus
 regression models from a single sliding stroke of the fingertip against the object surface, and that strokes that span a larger ran
omains and show that LEAGUE outperforms baselines by large margins. We also show that the learned skills can be reused to
d framework and regularization, we achieve consistent performance improvements in diverse TTA scenarios, especially on dyna

sing both the distilled data and current task data, with policy distillation performed based on uncertainty metrics. Our method is

while bypassing gradient-based optimization, thus relieving the learner from catastrophic forgetting. Empirical evaluation in a s

t-austin-rpl.github.io/Lotus/
ps was demonstrated through flight tests. Finally, the feasibility of the undocking mechanism was verified by mid-air separation

posed system focusing on robot motion, window frame detection and move-over mechanism. The results show that WiBot can
is performed in ROS, allowing ready portability to different applications.
nts in a realistic simulator.
me when UGVs have to wait to guarantee safety. We validate our system through extensive simulation with 7 UGVs and real-wo

ore, we present a neural network model for image-based force estimation to demonstrate the dataset’s applicability.
old speed up in labeling time and a considerable gain in performance while training deep neural models on automatically labelle

yxssm/PointSSC.

hmark our dataset for the spatio-temporal OOI classification task by adopting state-of-the-art approaches from related areas of
extensive dataset of RTS-based measurements to date. Additionally, we provide the precision of all poses for each experiment,

haring this dataset, we aim to foster and iterate new mobile robot algorithms quickly for robot actual working dynamic environm
niques to tele-operate a 7-DoF robot arm and collect a diverse range of human-robot collaborative manipulation data which we
d analysis of numerous baselines and investigate gaps between ML-based and traditional methods. Find the nuPlan dataset an
on effort -- the TBD Pedestrian Dataset and show that our collected data is larger in scale, contains richer information when com
with human oversight even if imperfect while also providing a single performance metric. We demonstrate a single video-conditi
generated scenarios on baseline autonomous driving methods, demonstrating our model’s ability to effectively create critic
ormation. Moreover, we also provide a corresponding human demonstration video and a language description for each robot se
ore, our approach is based on two key contributions. First, we collect a large dataset where an indoor mobile robot interacts with
marize future research on risk identification. Our aim is to encourage collaborative endeavors in achieving a society with zero c

ediction models form the foundation for studying how road conditions and surrounding entities affect driving behavior in comple
et is the first dataset that addresses the sensor-related gaps in the domain of 3D object detection in real traffic. Furthermore, we
ordinates currently perform best, outperforming both the keypoint-based method and the refinement-based method. Our datase
ata for specific weather conditions or sensor setups from a database in future. The dataset, pre-trained models for offroad perce
The results show that our solution excels the existing methods, and the network demonstrates its context-aware grasp on obje
ts challenge and thus value for VPR research.
ine results.
antitatively compared with the existing state-of-the-art methods of 3D reflection removal.
e matching and label graph aggregation strategies for pseudo-label reassignment with different task confidence. For this novel

hich contains labeled vehicle spray data in highway-like scenarios. The dataset will be available upon acceptance (see supplem

e selected S points. Additionally, coarse features containing high-level semantic information are incorporated into the child poin
RI tasks. The demonstrations showcase the real-world applicability of our advancements and may inspire more HRI use cases i
domain-specific parameters (DSP) and task-relevant parameters (TRP). Specifically, DSP exhibits sensitivity to outputs with sub
ments show the validity of STNet, which can reach 98.91% slip detection accuracy. Meanwhile, the ablation studies confirm the
heterogeneous inputs and cross-attention can be a powerful approach to fuse them. Our model is validated on the KIT Whole-
n planning tasks as an exemplary downstream application.
d outperforms other baselines by a large margin. In addition, we illustrate the usefulness of our method in real-world robotic ap
s, highlighting its potential for advanced scene understanding and manipulation.
el for realtime keypoint detection and used the inferred keypoints as control features for an adaptive visual servoing scheme. W

a mixture of real and synthetically generated images to predict the two-dimensional optical flow from a single image. Finally, we
ng SoA approach, our model results in a R^2 improvement from 32 to 47% on the horizontal image coordinate and from 18 to 5
h point-level and cluster-level pseudo labels, our PCTeacher achieves state-of-the-art performance on two benchmark datasets
2 times fewer parameters. Code, data and interactive results are available at https://claussmitt.com/pagnerf
d in indoor depth estimation. Our code will be available at https://github.com/AnqiCheng1234/GAM-Depth.
ssue as a conditional 3D keypoints generation task. To bolster cross-camera adaptability, we introduce the Normalized Camera
try learning of the LiDAR point cloud into Bird’s Eye View (BEV) based lane identification. Experimental results indicate tha
nce. Our open-source code and pre-trained models are available at https://github.com/Blurryface0814/PCPNet.
stantly reacting to dynamic obstacles and environmental changes. At the same time, it provides a simple way to define new mob
and provide a novel indoor 3D lidar dataset, collected during our experiments, which includes our automated annotations.
, occlusions, and noisy depth measurements, as well as generalization to unseen objects both in simulation and real-world robo
erformance on the standard trajectory prediction metrics, while significantly outperforming state-of-the-art baselines in terms of m

n under 5 touches, highlighting that vision-based tactile sensors are highly effective for fabric texture recognition.
el choices, respectively. Our results show that employing chemical synapses yields noticeable improvements compared to elec
 he effectiveness of each module and present our state-of-the-art results.
h approaches employing global optimization techniques.

ach in generating human-robot interaction behaviors that can be controlled via text commands.
emantic-guided residual data. Additionally, we introduce two modules for extracting and coding residual features. And incorpora
trajectory length over time. Additional results and an open source implementation is available at https://jstmn.github.io/cppflow-
lum), (2) simulation benchmarks (Safety Gym), and (3) a real-world agile robotics platform (RC Car) demonstrate great superio

sks and in a demanding real-world scenario, where we use a Kuka LBR Iiwa 14 robotic arm to perform the hitting movement in

s of our experiments demonstrate a notable improvement of grasp performance in comparison to conventional online learning m

emonstrating competitive results across three distinct robot manipulators.

r with 91.65% accuracy in our test task. Ablation studies were also performed to quantify the amount of data ultimately needed
ential adjustments for their integration and evaluative comparisons to state-of-the-art. In physical experiments with novel objects

n visuomotor policy learning, with a specific focus on improving the robustness of intricate assembly tasks that require both geo
econfiguration or retraining. We evaluate our framework in both simulated and real-world environments and show that it achieve
ual student policy can successfully learn to grasp and lift a challenging object, which was not possible with a single-view policy

elf-supervised learning paradigm using an algorithmic supervisor in a physics simulator. Real-world experiments show that our p
m/view/foldingbyhand
7 unseen objects. Comparative experiments show that the proposed HAGrasp outperforms open-loop baseline Contact-Graspn
are demonstrated by conducting real-world experiments using novel objects without the need of retraining.
real-world applications.

uate our approach on picking box-shaped objects of various weight, shape, and friction properties from a constrained table-top
thout supplementary training. Our approach does not need any real-world manipulation examples nor fine-tuning in the real wo
ith comprehensive experiments in SAPIEN simulation and real-world Franka robot. Results show that GAMMA significantly out
imulation and on a real robotic bin picking setup demonstrate the model’s ability to generalize across various object catego
methods. We show that our method surpasses or performs on par with those methods on robot manipulation tasks. Finally, we
real via seamless sim-to-real transfer. Our method significantly outperforms SOTA networks and even surpasses the performan
correction stage.
m more and diverse viewpoints typically improves pose estimation, it also increases the manipulation time. To address this trade
s for both generated datasets and real-world scenarios, presenting a promising direction for improving the adaptability and flexi
results. In addition, real robot experiments demonstrate the effectiveness of our method with a success rate of 94% and a clutt

eviously unseen, challenging "6-DoF" objects. Raw footage of real-world validation can be found at https://youtu.be/bwPf8Imvo
s like odometry estimation.
as the pseudo labels and generates class-aware feature (CF) maps for knowledge transfer between the two students. Second,
h prototypes. CrackNex outperforms the SOTA methods on multiple datasets. Additionally, we present the first benchmark data
the performance degradation issue caused by the use of binary point cloud Transformer modules. We propose a novel binariza
of downstream radar point-based registration tasks.

ng as well as for a robot grasping downstream task.

her than policies using tactile and vision-based rewards and 135% higher than policies without tactile observational input. Robo

urther the academic discourse, we make available a demonstration video, the trained models, and the source code for public a

g, and throwing actions. Extensive experimentation in both simulated and real-robot scenarios substantiates the effectiveness o
ge of objects successfully. With extensive experiments on a Yale Model O hand, we show that the proposed system can provid
the fusion model and embed the pre-trained model into a reinforcement learning framework for downstream tasks. Experimenta

omputing torques to be exerted by the arm and hand to close the loop. Simulation and experiments with a Shadow Hand have b

ons, experiments, and comparative studies are conducted for validation.

ow that model-based motion planning can perform comparably to RL with dramatically
e, it provides a flexible and generalizable solution for various manipulation tasks. As examples, we analyze the results on 7 han

and experimental results demonstrate that the proposed scheme can generate reasonable and efficient paths for R-UAV in the
ollision failures by up to 8.5 times by selecting the best available machine (2) Target Tracking: FogROS2-LS also enables robu

atform verify the applicability of the methods. A video with all results is provided in the supplementary material.
d with manual heuristics or end-to-end predictors, it is shown to both improve online adaptability and retain the quality guarante
g pitfalls. The effectiveness of our algorithm has been validated through extensive numerical simulations in high-fidelity environ
ization and Mapping (SLAM) and power grids.
experiments of the Point-to-Point Motion are conducted to test the robot's performance. Experimental results reveal that the ave
the effectiveness of the proposed control strategy in attitude control.
om- pare our approach with existing methods and demonstrate its superior efficiency, requiring less than 15% of the time of the

they scale better than the widely used O(nd^2+m^2d+d^2m+m^3) LTL algorithm of Featherstone.
ucting RIM using flexible mechanical system models over rigid body models are also addressed by comparing the effective ma
ed. Accordingly, RIDER equips robots with the ability to generate navigational behaviors by predicting environmental changes,
trajectories similar to those trained in simulation and narrows the sim-to-real gap. Experimental results show that our method, t

effectiveness of our method, yielding an average translation precision of 0.48 mm and rotation precision of 0.45 degrees. Conse

ptually-degraded environments. The key novelties of this work are the designing of the hierarchical probabilistic map matching fr

lution for the problem and also propose various heuristics to efficiently compute solutions for large problem instances. We eval

pected post-impact behaviors. The monitoring framework fuses a (bandpass) momentum observer with impact-aware control to
ng flows, tailored to the data domain by taking its structure into account.
mary tabletop manipulation actions. In order to detect these failures, we propose FINO-Net [1], a deep multimodal sensor fusion

m and actuator models, it is sufficiently general to be applied to other systems.

provement in explored volume increment performance. For validation, we conducted field tests using UAVs in real subterranean
door testing environment and shown to autonomously perch and unperch from a steel cable. The grippers force application was
existing CPU implementations. Because few GMM-based frameworks are open-sourced, this work seeks to accelerate innovati
on and real-world scenarios with the Searcher robotic platform. Comparative experiments with expert operator control and state
-world experiments, underlying the gain in performance and reliability.
lve picking and placing various objects, such as plastic bottles, bags, and cans. The results demonstrate successful teleoperate
ex source location scenarios. With regard to computational requirements, the method achieves update rates of 10 Hz on a Ras

work contains a novel dynamic map module that can simultaneously track dynamic obstacles and represent static obstacles ba
le, and convenient means of increasing the efficiency of their solar energy system.
human labeling.
Franka Emika's FR3, and Kinova Gen3. The experimental results show that most of the energy (60- 90%) is consumed by the e
ike battery replacements, and then facilitate direct relaunches. Importantly, our design places all operational responsibility on th
ocations within a user specified map update cycle. We present a simulation study to demonstrate the alternating exploration-ex
ng and point measurement data. Our approach minimizes the number of days and locations for sampling, while preserving the
obot moves, indicated by systematic deviations to a ground truth terrestrial laser scan.

isible. This enhanced awareness contributes to better navigation performance in environments where such predictive capabiliti
alization and force estimation. In the grating identification task, ViTacTip demonstrated an accuracy of 99.72%, surpassing TacT
ance shows promise in embodied tactile sensing and human–robot interaction.
dditionally, benchmarking our system across CPU, GPU and Loihi2 hardware platforms resulted in a 96% reduction in power c
ors and three object geometries, and highlight its practicality for the task of USB stick insertion with a compliant robot arm. The
d planner, generating shorter and more reliable pushing trajectories despite large disturbances. The simplicity of our training en

nd Low-cost while facilitating a wireLess deployment. Evaluations and experiment results demonstrate that the proposed L3 F-T
ntly enhance the capabilities of underactuated fingers and potentially broaden their future usage.
of tactile sensors more accessible to roboticists by allowing them the flexibility to easily customize, fabricate, and calibrate cam
shared tactile latent space for more accurate and adaptable tactile networks, showing an improvement of up to 84% over the sin
ments demonstrates the accurate state estimation performance of AllSight.
ss the 6D force, which can generalize to unseen objects. To promote the development and applications of vision-based tactile s
d has the potential to assist humans in detecting defects on curved surfaces of infrastructure such as pipelines.
construct a 3D view of a scene just before grasping and perceive the tactile imprint after grasping, allowing for monitoring of the

nce at different indenting forces. Then, we computed the optical flow as a tactile flow approximation and its divergence to estim
t in the bi-reorienting task due to the sim-to-real gap. Finally, we demonstrate the generalizability and robustness of this system

hieving high accuracy in terrain classification tasks.

dataset are available at https://github.com/leolyliu/TEG-Track.

ercentage error of 4.15%. The sensor was integrated into two robotic 3D-printed soft grippers to enable real-time monitoring of d

of the tactile simulator to provide a real-world framework. Potentially, the framework can be used to train, for instance, reinforce
single image without relying on stable lighting conditions. Finally, we perform surface reconstruction from normal and point clou
e sequences to generate realistic marker motions during the normal and lateral sensor deformations. We evaluated and compa
al-world demonstrations.

nt from falling from the gripper. The experimental results demonstrate the improvement of our contrastive visual-tactile model pr
e focus on two distinct tasks: material classification and grasping success prediction. Our results indicate that MViTac facilitate

e-scale data acquisition process. Simulations demonstrate the va

generalizability of our analysis on various multisensor systems.
posed framework. The experimental results demonstrate that the proposed framework can achieve accurate contact angle estim
n resolution between the axes, a method for improving the resolution isotropy without changing the difficulty of machining is als
antom with sinusoidal motion. To the best of our knowledge, this is the first study to validate adaptiveness of RUSS under dynam

ts. Supported by realistic experimental evaluation using a Franka Emika robot, our tactile robot programming approach for the

ms of expressiveness. By incorporating our robustness definitions into a trajectory planner, autonomous vehicles obey traffic rule
h and offers experimental evidence of its shortcomings. Finally, it presents valuable insights to promote best practices when em

The proposed iterative method leverages graph structure to generate a more efficient search space. The effectiveness of our m
. We demonstrate our approach on a real F1/10th autonomous vehicle that navigates in indoor environments and transmits rich
horizon planner for Markov Decision Processes that preserves guarantees over MITL temporal robustness. We show the scalab
sks. An environment-agnostic LTL pre-training scheme is further incorporated to facilitate the learning of the Transformer modu
ynthesize strategies to achieve complex tasks with guarantees. Further, we discuss the current computational limitations and im
eveloped, in which the automaton states of local and global task are jointly considered to generate a feasible planning that satis
onstrate the validity, the effectiveness, and the transferability of IIRTS.
-and-deliver and navigation tasks.
rface painting test, validating the developed model.

peration that is robot agnostic.

irregular and visually undetectable scenes. Experimental results demonstrate the validity of the framework.
eatly alter the intrinsic stiffness of the grasp produced by these fingers. A notable feature of the proposed new design is that on
accuracy compared to the baselines. Moreover, our method can successfully perform grasp transfer to unseen object categori

Attacks-on-Grasping-Networks
set with the additional reference sources and propose an evaluation strategy that simulates coverage for different reference sou
stribution of features. Finally, we applied the developed algorithm to a physical robot for grasping. Experimental results on the V

proposed for making grasps more robust to domain randomization, resulting in a transfer ratio of 84% on the Franka Research 3
grasping motion learned from human demonstration. The kinematic synergy primitives have a low reproduction error of 3.9% o
ot handover.

n cluttered table-top scenes. We show the effectiveness of the proposed method by extensively evaluating it against state-of-the
or states, object pose, video, and grasp success data are provided for every trial.
enable MoMo to simulate future action outcomes and detect out-of-distribution states and high-risk situations preemptively. We
lization. For the generalized grasping tasks with dexterous robotic hands, our method boasts a success rate nearly 20% higher
ntact-rich dexterous manipulation skills directly in the real world. We identify key ingredients for leveraging demonstrations in m
pproach in generating feasible grasp poses and achieving successful grasps in dynamic environments. By integrating these ad

s confirm the efficiency of the proposed scheme.

CBs with large surface mount parts. The proposed method offers a promising solution to enhance the efficiency and reliability of
he completed object shapes from various directions and positions in real-world scenarios. Our work opens up new possibilities
would not be apparent from overall success rates, can be identified and evaluated.

etitive solution for robots that need to manipulate objects and check their mechanical qualities at the same time.
st results demonstrate that this novel self-adaptive gripper is comparatively highly adaptable for grasping irregularly shaped obj

ables five-mode grasping in comparison to pneumatic actuators. We believe that bionic soft fingers present a promising solution

y of the gripper. Obtained results are promising and suggest that this approach could be particularly advantageous in the contex
odology scales well with increased co-design parameters, rendering the co-design of complex, high-dimensional manipulation

ch is further augmented by tactile data augmentation methods.

orementioned informations, the joint angles can be reconstructed in high precision based on the inverse kinematics models of

rly to synthetically generate different onboard vibration energy distributions for tactile sensing. Experiments demonstrated that p

s tissues in 25 out of 27 trials, illustrating the potential of jamming-based mechanisms as bidirectional interfaces capable of pro

entation provides a perceptual sensor model for a particular tactile scene. Models trained on actively collected data outperform
a multi- segment version of the system and show the display’s ability to withstand loads during contour following in haptic e
outperforms the non-interactive data collection methods for scene reconstruction.
iations. This approach is experimentally evaluated against a standard zero-torque controller with a stationary single-degree-of-f
er exploration of robotics in new societal applications, fostering connections and collaborations across diverse fields.
en considering different robot hands with dissimilar kinematic structures. Furthermore, the subjects were able to successfully pe
ameters that maximises such coefficiency. Results proved that by acting coefficiently the robot could meet the individual prefere
TN). FN are then subcategorized into either missing annotations or catastrophic labeling errors. Images lacking labels are augm
for a Human-in-the-Loop agent.
tions noticeably improved and crashes minimized by day 10 of training, with a training plateau observed by day 15.
with a random, a cyclic and a greedy (short-term optimization) policy, for different user profiles. We show that our approach out
mental findings reveal that we can enhance the generalization of the estimated reward function without requiring additional hum
navigation policy. By incorporating part-time human interaction, robots recover quickly from their mistakes, significantly improvin

e our framework using two months of real-world electric micromobility operation data in a city. Experimental results show that o
d experiments are implemented to verify the performance of our system. The maximum velocity can reach 6.0 m/s for a novice
oads to evaluate the release and compensation time, and the effort reduction with respect to non-powered exoskeleton case. R
n compliance in the face of unexpected contact caused by unplanned obstacles, which provides a guarantee for safe manipulat
sassembly tasks.

state-of-the-art deep learning methods like LSTM, GAN, and CNN-based approaches, by approximately 86.46% when predictin
o the original SARL policy, our modified policy maintains similar navigation times and path lengths, while reducing the number
ate with their path preferences in RCEs. Finally, we conduct thematic analysis of custom open-ended questions to gauge intere

ared various navigation methods in BSON using a newly proposed metric called socially-aware navigation score. Through BSON
best of both worlds. In particular, IJP jointly optimizes over the behavior of the ego and the surrounding agents and leverages d
motion tendency and motion uncertainty of dynamic obstacles. Qualitative and quantitative simulations demonstrate that the pr
ncrete examples and quantitative evaluation.

This outperforms baseline models in both recognition accuracy and generalization. Furthermore, an ablation study demonstrat
g the state-of-the-art machine-learning algorithm by 6% on a more difficult version of the task. The fact that our classifier obtains

esides, traversing curtains in a medical scenario was tested. All experimental results demonstrated the proposed frameworkâ€
omputationally `force' the network to learn channel dropout variations and thus learn robustness to channel dropout. The propo

se due to robot reverse psychology and leveraging its potential to boost team performance. The experimental results demonstra

d recognize the cued directions with an average resolution of 19.8â—¦ and 19.6â—¦ at different speeds, respectively, and there

ble at https://emprise.cs.cornell.edu/morpheus/.
ion along non-constant curvatures subject to low-frequency disturbances. The MIntNet was able to accurately predict future po
e to estimate the path and actions at waypoints on an undescribed path, if enough information is available. Experimental result
display effective task mental model alignment, 18% more victim finds, and 15 times more efficient guidance plans then current o

s scored low on expectancy of use for work in the future at their current state of development.
the abduction-adduction direction compared to supination-pronation. The findings of this paper indicate that sEMG encodes sig
on and also demonstrate it in real-world confined cabinet-like settings with multiple unknown objects. Our results show that the
also decreases the overall activity of lower limb muscles by up to 25% and thigh co-contractions by up to 40%. Participant feed

controller. Experimental results demonstrate the effectiveness of the proposed method in achieving human following and obsta

he belief MDP (derived by reformulating the POMDP) with a probabilistic labelling function. Then a product belief MDP is const
ith a robot collaborator quantitatively and qualitatively.

rlayed onto the environment, and hand gestures are used to add or delete 3D map features in real-time. Our system is impleme
mpared to using only real training data. The proposed method can be applied to a variety of dynamic tasks and different LSTM-

0, and one composed by only a collaborative manipulator, i.e. a UR5e.

k. The proposed shared control framework is verified both on a virtual platform and then on a real dVRK to assess its usability a
bot increases.
Emika Research 3 robot.
pular deep RL method are vulnerable to changes in policies of other agents and that the proposed framework improves the robu
ompared to direct teleoperation.
e sequence of online corrections in a knowledge base and retrieves that knowledge in settings with new task or object instance

al results show that the learned trajectory can be reproduced and generalized to several targets and adjusted online according
superior tracking accuracy and trajectory smoothness under external disturbances. Subjective responses underscore users' per
controllers, showcasing its potential for applications in planetary exploration and other scenarios requiring dexterous telemanip
table plans are found in most cases. With this framework, we intend to add flexibility and generalisation to HRC plan generation

he robot to switch smoothly between the two HRC modes. An optimal final control input is determined by a quadratic program (
es while ensuring safe interaction.
nformation, correcting inadvertent errors, and sifting through in-consistent movement decisions. The performances on the repre

r work in both simulation and a user study, which analyzes the feedback given and its implications. We also use the total collec
paintings and drawings that match the input text prompt more clearly than FRIDA, both from a blank canvas and one with huma
y of wearable robots for PVI. The source code has been made publicly available at https://junweizheng93.github.io/publications/
both in laboratory settings with 12 healthy blindfolded subjects and a proof-of-concept demonstration in a real-world scenario.

covery, which resulted in high energy consumption and instabilities maintaining an upright posture, concurring with prior NASA
plementary materials are available online.
nce, with task order differing by experimental condition (rising competence, falling competence, or failures at the midpoint), aski
redictions. Based on the hypothesis that human trajectory patterns also manifest laminar flow (that represents predictable motio
 our architecture with a physical mobile manipulator robot and perform extensive system evaluations in a laboratory environmen
that our robot can stably cooperate with human users to create diverse styles of portrait drawings. In addition, our portrait drawi
who trained using the VR-2DD task compared with those who were trained with the simple task. We also carried out active (onlin

m 3D models by using a consumer-grade robotic arm. We evaluate the proposed framework by taking 3 standard models and a
github.com/cvg/hololens_ros.
e planning, and on-site learning within a few minutes.}

nd dynamic running. Notably, even in the absence of transition motions in the demonstration dataset, the robot showcases an e
ironment. Finally, we validate the effectiveness of synthesizing EBG and NC-MPC for humanoid locomotion on non-flat terrain
orphic encoding scheme to reproduce essential features of the NWR e.g., spatial summation effect and temporal summation eff

e design the three-layered structure by analyzing the shape differences of hundreds of scanned human models. An energy-bas
strate that our method achieves low root mean square errors in torque tracking, ranging from 0.05 Nm to 2.5 Nm, even in the p
effectively bridge the gap from simulation to reality and outperform other methods significantly for point cloud-based 3D object d
mplementarity. Our method achieves state-of-the-art performance in the highly competitive nuScenes 3D object detection datas
n blocks for transmission efficiency. Experiments show that EMIFF achieves SOTA on DAIR-V2X-C datasets, significantly outp

method outperforms existing state-of-the-art methods for multi-class road defect detection and segmentation methods.
each agent, coupled with a cross-attention mechanism for the final fusion. Extensive experiments on OPV2V-HPL demonstrate
mainly digested by a lightweight 2D BEV convolution network and compensated by integrating the 3D voxel features interpolat
rformance, and improvements even compared to previous multi-modal solutions.
age backbone to the monocular semantic occupancy prediction framework with low cost of hardware. With these advantages, o
ects of the 3D MOT problem. We quantitatively evaluate ShaSTA on the nuScenes tracking benchmark with 5 metrics, includin
nts on the public OPV2V and V2XSet datasets underscore FDA’s effectiveness in point cloud-based 3D object detection, p
d task-aware discrepancy. Our extensive experiments on the OPV2V dataset demonstrate that these attacks substantially unde
ds with real-world perception inputs, allowing us to compare the performance of conventional and end-to-end methods for the fi
on in practice.
nd meanwhile enhance the performance of 3D object detection network.
to extract complementary and beneficial cues for collaboration task. From an information-theoretic perspective, the proposed a
i-sensor Kalman Filter. Our algorithm learns to estimate measurement uncertainty for each detection that can better utilize the t
. In situation 2, We observe that scale is the strongest factor in improving closed-loop generalization regardless of the choice o
prior baselines and is capable of generalizing to composable new tasks and adapting to Out-of-Distribution (OOD) tasks.
making process. To that end, we incorporate the correlations among dimensions of the state into the policy, providing more inf

y inferring an appropriate ranking, the reward function is able to guide reinforcement learning by indicating when task progress i
superior performance of our querying method over the state-of-the-art information gain method.
n. Furthermore, the framework also can be applied to garment unfolding, and experiments suggest that it exhibits generalization
significantly increase inference frequency without losing decision quality. The impact of each optimization alone has also been
 ss of our approach in providing upper-bound estimates for escape energy in quasi-static manipulation scenarios. Two real-worl

W of the spherical mechanism can be increased by up to 3.5 times. Finally, experiments are carried out to show the proposed ro
and the removed features and avoid creating new dependencies. The methods are evaluated on the Allegro robot hand in a rea
on space and safe, guiding behavior over which RL policies are trained. Behavioral dynamics enable bang-bang-like RL policy a

ents on a 16 degrees-of-freedom robotic hand grasping and manipulating objects with different
application was presented to demonstrate the robot’s capability to perform a wall painting task.

navigation tasks among heterogeneous agents, assisting the cooperative task completion by HVs in a smart city. Experimenta
monstrate that our proposed SCRNet outshines existing state-of-the-art methods, highlighting the potential of SCRNet as an effe

cute a parametric trajectory (spiral or raster) for an autonomous area coverage and meanwhile command the base in order to k

hieves better highway driving quality than the state-of-the-art on high-fidelity closed-loop simulations.

nt types of task specifications.

oject website: url{https://sites.google.com/view/active-mde}

h real vehicle driving logs. Furthermore, our closed-loop simulations of several benchmark scenarios demonstrate how the beh

nuPlan dataset.
ds. We then design multimodal planning networks which enables MNP to learn and predict multiple optimal solutions. Simulatio
mproves the planning performance of an MPC method, and also outperforms prior diffusion-based methods.

e fixture calibration.
adation model is identified, we practice the mathematical flexibility of the HMM framework to estimate several of the no longer a
ask boards, describe a method for deriving the relative task complexity (RTC) from timing data, and compare robot solutions wit
ates using the proposed model transformation are demonstrated based on experimental data for the flexible manipulator with re
d integration. The proposed system eliminates the need for human intervention, facilitating automated garment production. This
 acing can optimize the overall process by reducing uncovered gaps and overlaps between coverage lines.

nnection and thus avoiding unnecessary downtime.

nse and large instances at a scale of more than one hundred without significant loss in efficiency.

ed by 99 times, and scenario database utilization rate increased from 0.2% to 36%).
he deployment of autonomy at scale. While this work focuses on AVs, the scalable supervision framework may be of independe
havior, showcasing its versatility in modeling human drivers in diverse highway scenarios.
etrics, and further investigate the contributions of proposed modules via comprehensive ablation studies.
ment. We also examine the proposed method in CARLA simulator and perform thorough comparisons to demonstrate the effec
trajectories from real-world driving data and conveniently train an SCBG model generating trajectories based on the input court
servation. This approach can more accurately identify risks that the driver is not aware of and provide warnings only when nece
ion results using RetinaNet and FCOS on the nuScenes dataset demonstrate that training the models with our safety-adapted
nts confirm that our approach produces excellent performance on both SemanticKITTI and SemanticPOSS benchmarks, achiev
mo Open Dataset.
he effectiveness of our proposed algorithm on the OpenCOOD platform. Our results demonstrate a substantial reduction in com

simulation and experimental results to validate the effectiveness and robustness of the proposed solution.
A simulator, employing an autonomous driving control agent and a steering wheel maneuvered by licensed drivers. Experiment

rom environment to environment; 2) images are ubiquitous through satellite imagery, in-car camera systems, and traffic monito
The results of our experiments suggest that the proposed methodology can attain driving scores of 49.21%, coupled with an imp
n separate training of the two modules. Additionally, we find that tree-structured policy planning outperforms the conventional s
encoder to obtain time-dependent latent variables for a motion policy predicting the future trajectories. Our results show that th
hat randomly drop non-causal agents throughout training. Finally, we release the causal agent labels as an extension to WOMD

ths of DRL and conventional methods in generalizable control.

that can run sequentially given the user's computational resources. The algorithm then applies instantaneous assignment (IA) f
chance of AV failures in these scenarios at each iteration. Subsequently, by re-sampling from historical scenarios based on the

-of-the-art performance with the lowest inference latency and smallest model size. Experiments also demonstrate that our fram
ting a motion planner standalone without a decision module.
Prototype Matching (MPM) (searching the appropriate leaf node and the links to the root node) and top-down Leaf Link Aggrega
n process (POMDP). An optimal warning generation framework is proposed as a solution to the proposed POMDP. The simulat

nse traffic, demonstrating satisfactory practicality and generalizability.

ficiency and adaptability of trajectory prediction algorithms in AVs.
and provide high-quality multi-modal predictions in real-time.
n under the backpropagation neural network, CNN, and CNN-LSTM are compared. Moreover, the online intention estimation an
e/detry-lab/public/robot-trajectron.
 d images. We fed our magnified images directly into a simple ResNet18 network for recognition, achieving a competitive score
maneuver takes place. We release our model configurations and training scripts on https://github.com/ykung83/Driver-Intent-Pr

art performance, but also significantly improves performance by up to 39% in challenging scenarios, such as high ego-speed, co
of hand motion before at least 50% of the trajectory is completed. It also successfully exploits gaze fixations when the human o
e claws within 20 milliseconds after reaching closed configuration. We demonstrated the effectiven
c manipulators based on an inherent self-sensing capability of electrohydraulic actuators. The proposed architecture can serve
ry for providing ankle assistance during the push-off phase of walking, while maintaining low stiffness (4 - 12 Nm/rad) for planta

e performance is greatly improved by optimizing the transmission path.

obot is made 4 times more compliant than it is nominally using tendon actuation only.

of diverse modeling techniques. By assessing the differences in reconstruction losses, we illuminate the merits of each approa

ensional reference paths with different path characteristics. We show that our method manages to quickly traverse the reference

Parallel manipulator - Cable Driven Parallel Robot to validate the results through both simulation and experiments. A visual repr
ational effort by up to 92%.
other existing segmentation methods may benefit from using this trajectory representation to improve their invariance.
e ABAQUS implementation of the finite element method (FEM), and has speed comparable with the real-time FEM framework S
e and in real scenarios with KUKA LBR iiwa 7 robot.

s indicate that the optimal time planned in the virtual passive-joint space is decreased by 2.8% and 8.1% compared with the ac
spectrum of a reduced Hessian matrix of the potential energy.

t wrench-twist capability and enhances the robot hardware effectiveness without any ha
d-lab/ricmonk and the video demonstration of the experiments can be viewed at https://youtu.be/hOuDQI7CD8w.

he conservativeness of the energy function and outperforms baselines in terms of controller performance while maintaining safe
and quantitatively.
res. We also show that with the proposed method, skill learning is feasible and our safe manipulation tools can be transferred to
LMs, trained on a vast corpus of text describing typical human behaviors, encode substantial world knowledge, including proba
ces to modalities it does not direct attention during the task, resulting in a consistently high success rate (≒90%).
idity plastic-made humanoid robot, to validate its effectiveness.
stimates, particularly under a limited amount of new experiments, and improve the robot’s ability to predict and prevent acti
rst- and second-order dynamical systems. In the latter, motions are 3, 4, and 6 dimensional, of first and second order, and are
movements for new task scenes. Notably, EditMP enables all robot joints rather than the robot end-effector to avoid hitting comp
on, only a single image of the desired goal configuration is required at inference time.

and Locally Weight Regression to encode and learn the motion and impedance required to execute the task. To explore what i
generated by our semi-supervised LfD model on the physical robot hardware using the KUKA iiwa robot arm. We also validate
n sum-of-squares optimization. Simulation results show that our approach significantly saves time and resources in finding comp
different simulated control tasks and show that our method results in safer control policies compared to other classical and learn
r of objects. RLAP furnishes significantly higher goal reaching rate and shorter planning times compared to the state-of-the-art
ments with payloads from 0 to 3 kg are conducted. During the experiments, the operator force is measured to evaluate the prop

ith respect to expert ratings, as well as individual characteristics of excision forces. Results show that the proposed framework

n even with the presence of disturbances. Project website: https://airexo.github.io/

New motion trajectories, which mimic those sketched by the user, can then be generated from the probabilistic model. We empir

how that PromptAdapt is a strong domain-adapting policy that outperforms baseline methods by a large margin under a range o

in shorter time periods than other state-of-the-art planners even when using unreliable demonstration data collected by a single
cantly improves off-the-shelf offline RL agents' performance on some of the most challenging D4RL benchmark tasks. For insta
more, in the context of tabletop manipulation, we demonstrate the applicability of our novel skill discovery approach to acquire sa
task is within 25% of the optimum, at the expense of reduced volume of material removed per pass. A key advantage of our ap

owing it with potent representation learning capabilities. It provides a robust estimate of the Q-function with minimal variance an

By distilling the learned latent dynamics model from the teacher to the student model, TWIST achieves efficient and effective si

Experimental results on multiple tasks show that WODTEM can improve sample efficiency.
rewards to guide the low-level control policy. We validate the proposed method COMPOSER with five snake robot tasks, includ
 g. The dataset is available at https://theAirLab.org/TartanDrive2.
tage training strategy aimed at consistently enhancing the performance of both functions. Additionally, we propose a novel dom

its efficacy on multiple legged platforms.

verify the robustness and feasibility of our method, and the experimental results show that our method outperforms the state-of
estimation methods demonstrate the superior accuracy and robustness of our approach. Additionally, we investigate the transfe
moothness by 27% compared with traditional methods.
effectiveness of our approach across various metrics. We also presented experiments that demonstrate the capability of MATR

proposed to learn compact representations which disentangle robot poses and object positions from high-dimensional observat
ic representations for downstream robotic pushing manipulation tasks.

experiments demonstrate that our proposed method outperforms RNN-based approaches.

d at https://sites.google.com/view/language-deformable.

nts show that our method achieves an 86.4% (open plane area) and a 73.8% (room) success rate on evasive objects, outperfo
ng diverse objects compared to the existing method while sustaining data efficiency. Our method can solve some tasks with su
has more efficiency, higher success rates, and demonstrates more realistic behavior.
ts adaptability to unseen furniture configurations, and its robustness against visual disturbances. Ablation studies further accen

asks that prior work considered. We show that our method outperforms prior works that do not use equivariant architectures or
and a vision-based grasping task with a 7DoF manipulator. Finally, we show that the vision-based grasping task trained in simu
rimental studies demonstrate that MultiAC6 generalizes well, without retraining, to DLOs with different lengths or materials.
LAM scenarios: 1) a 120km city-scale trajectories with LiDAR-based inputs, 2) a multi-visited 4.5
higher performances in longer-range tasks. Moreover, a penalty term is introduced in the reward function for the high-level poli

show that RaLF achieves state-of-the-art performance for both place recognition and metric localization. Moreover, we demons
. The consistency of the local map can enhance the adaptability of our method to challenging scenes. Extensive experiments d
ng to utilize the physics information, the agent can transfer this knowledge across different tasks and navigate in an unknown e

eater training efficiency. The trajectories collected with the approach confirm the effectiveness of the proposed adaptive reward
ompletion rate.
data augmentation and mirror loss function. We provide a theoretical foundation for using augmented samples in an on-policy
bject-to-arm correspondence in each task assignment round. In the experimental part, we adapt some search-based methods t
 hrough guided online safe RL training, which outperforms both the offline DT policy and online safe RL algorithms. Experiments

ological Mapper and a novel RL-based Hierarchical Topological Planner (HTP). The Topological Mapper employs a visual enco
sian process. Our method can correctly adjust the support surface prior to interacting with the terrain and is extensively tested o
to stabilize the quadruped under external perturbations. Quantitatively, in simulation experiments, we cut the failure rate up to 4

e policy retains the robot's planar mobility, enabling rough velocity tracking. Finally, zero-shot Sim2Real transfer is achieved on
tacles, pushing objects, crawling under a table, pushing a door open with its leg, and holding it open while walking through it. T

ith a minimal burden of reward shaping.

rent robot platforms. We trained the policy and implemented our approach on the Unitree A1 robot, achieving impressive results
ength, do a handstand and run across tilted ramps, and generalize to novel obstacle courses with different physical properties.
e locomotion behavior in simulation and on the quadrupedal robot ANYmal. Videos of the experiments and code are available a
on. Videos are available at https://risk-averse- locomotion.github.io/.
ion tasks, revealing that policies trained with EIPO consistently achieve higher task performance than PPO comparisons while
ked with planning waypoints to reach a faraway global goal location, and an RL-based policy trained to follow these waypoints
olds to maneuver successfully.

of the benefits of utilizing rolling contact in the stance foot, specifically the reduction of necessary positive and negative work thr
y simulations with the humanoid robot TORO.

iveness of our proposed controller is evaluated through extensive simulations and physical experiments conducted on a rat rob
uped (on hardware) and a humanoid (in simulation).
comotion over high obstacles.

G) in the spinal cord, effectively coordinating both frequencies and amplitudes of the CPG to produce rhythmic output (Rhythm

e available at https://sites.google.com/leggedrobotics.com/pedipulate.
rld testing, our policy achieves proficient command-following performance with only a two-minute data collection period, and ge
of quadruped robots. We validate our proposed method through simulations and real-world experimental trials on a large-size qu
 A1 robots with different payloads on variable terrains and in the presence of disturbances. It is shown that the distributed MPC
y supports the training of control policies for robots with diverse dimensions, eliminating the necessity for robot-specific adjustm

p convergence of the trajectory under an LQR tracking controller. The generated convergent trajectories recover more effective
demonstrated the ability to balance and steer the skateboard.
mics simulations and real hardware implementations on a quadrupedal robot provided by Skymul Inc.

present our results on different quadrupeds and bipedal robots in simulation and showcase how our method allows the balanc

enerate task-specific descriptors. Our experiments show that our representation can allow for better robustness than current so

We evaluate MBFusion on three architectures, including HDMapNet, VectorMapNet, and MapTR, to show its versatility and eff
mators significantly reduce the data transfer and thus improve energy efficiency. As shown in our extensive experiments, the pro
able at: https://github.com/USTCLH/OCC-VO.
ng. Finally, these functionalities of NF-Atlas are validated. Thanks to the sparsity and the optimization based formulation, NF-A
alized by our approach.

registration process. The generated maps consume only a compact size of 450 KB/km and remain adaptable to evolving envir
the-art methods, our method is more efficient and robust, and reduces the running time per frame by more than 60% on averag

implementation. Moreover, OctoMap-RT constructs a map with 0.52% higher accuracy, in terms of the number of occupancy g
models, and datasets will be available at https://github.com/szturobotics/Stair-detection-and-modelli ng-project.
we evaluate our approach on zero- and few-shot tasks, demonstrating unprecedented performance for generalization to new en
stics of the robot platform. Based on these two strategies, the proposed method overcomes the limitations of existing methods,
d interpretable continual learning paradigm for navigation by perception.
udies show that the proposed knowledge distillation technique surpasses other counterparts. Implementation of our method has

zero-shot transferred to scenes with unseen maps and unheard sound categories.
e InstanceLoc, which enables near realtime localization of instance masks given a trained label field and an off-the-shelf image
licly available large-scale datasets. Results show that the proposed method can track the robot pose even though the map sca
ot localization tasks. Specifically, the mean localization error is $0.503$ m for all cases.
synchronous capture of camera and LiDAR data, granting us the flexibility to manage reconstruction detail levels across variou
ically degenerate scenes. By using the additional photometric information, our approach shows drastically improved robustness

o a priori knowledge of whether the viewpoint is similar or opposing, and also demonstrates competitive performance in similar

n. This strategy effectively bridges the gap between SLAM and real-world coordinates while also substantially reducing accumu
titive performance on KITTI depth estimation.

. Experimental results demonstrate that our method outperforms state-of-the-art methods, achieving 85.3% Recall@1 on the M
ted effectively, and the online calibration notably improves the pose accuracy.

mental results show that, in 85% of scenarios, the proposed localization system can estimate its pose within 3m in large indoor
ethod achieves state-of-the-art performance.
sts, we demonstrate that the method offers a better trade-off between accuracy and computational cost.
oves data efficiency. Experiments on public datasets demonstrate that our method could select the samples that bring the most
at https://github.com/ga1i13o/JIST
nt LPR methods. The experimental results show that our CCL consistently improves the performance of different methods in di
oot Mean Squared Error compared to our VIO SLAM baseline. Code example: https://github.com/AlexS28/OptiState

le upon acceptance.

e representations, which are then aligned to estimate a pose that matches the pose label. This formulation allows for the active
achieving new state-of-the-art performance levels.
al descriptor by an attention-based aggregation module, to query the compressed Lidar map in the vector space. A transfer lear
and and Oxford RobotCar benchmark localization datasets, which include up to 27k places. We found that VPRTempo’s ac

characterizing the semantic properties of these features, uncovering unique domains which encapsulate datasets from similar e
rs for localization tasks. Moreover, the full dataset or some parts of it can also be used for other research areas such as channe

5 m (s.d. 0.05 m) for the state of the art, meaning that the benefits in terms of tuning and flexible configuration do not come at th

es GNSS outliers and shifts without requiring coding of a whole set of exception cases. We use real-world data from the Ardupi
esults indicate that we can achieve comparable accuracy even with infrequent range measurements down to 1Hz.
a significant performance enhancement, with results showing $sim25%$ improvement over the best baseline.

articularly when applied to position estimation utilizing GNSS pseudorange and carrier phase observations. The multiple update
om WiFi and barometer localization, our experiments show that our global localization and the weighted point to line ICP pose tr
graph reduces the prediction error of landing positions by over 63.6% compared to a baseline method that utilized an adaptive e

ccurate estimation of 6-DOF ego-motion over long distances also in feature-sparse mine environments. These results are usefu
s. Our implementation is open-sourced: https://github.com/Pamphlett/Outram.
ods. Through comprehensive numerical evaluations across four state-of-the-art marginalization archetypes, we not only affirm

ior arrival rate and generalization compared to other state-of-the-art methods like the PPO-based navigation method.
bound on the probability of satisfying the safety specification. Finally, the efficacy of our proposed algorithm is demonstrated th

nd error, and retain the predictive properties of the MPC to better handle out of distribution behaviour.
t in CARLA on challenging dense traffic scenarios when evaluated at realistic speeds.
how that our method can manipulate different objects with a single demonstration and significantly outperforms the baseline in b

nce over a baseline MPC algorithm for a given computational budget. It can outperform the best MPC algorithm by up to 27% w
er to safely control the robot under its current understanding of dynamics. We demonstrate in simulation and in the real world th
s a resource for the community, describe its design choices, and present experimental results. Perhaps surprisingly, we find tha

esulting in value functions and policies for manipulation tasks that perform better, act robustly, and generalize broadly. On seve
ion (BO) to gain insight into successful turning gaits in high dimensional search space and found strategies such as differential
ring online training, a behavior correction mechanism is adopted, ensuring the agent to interact with the environment using safe

he proposed method is evaluated using a set of YCB-objects and a suction gripper, demonstrating a success rate of 95.2% in s
n significantly enhance performance. The appendix, videos, code, and trained models are available at https://force-sight.github.i
ness of the proposed approach, which highlights the high success rate of our solution for retrieving everyday objects.
geometry or finite-horizon planning. Finally, we demonstrate using the signature in the real world to manipulate a cable in a sce
oint cloud containing precise local geometries for manipulation, and learn affordance on such point cloud to propose fine-graine
MAMO solver with a rich action space. We discuss the impact of the improvements we make in an extensive simulation study

ing range sensors. Our system's code will be accessible at https://github.com/HITSZ-NRSL/RIM.git.

s. Code and data will be made publicly available.

asks or aspects through subsequent migration.

acy, texture realism, and time consumption.
ion. We perform extensive evaluations on indoor and outdoor real-world datasets, achieving substantial improvements over sta
es robot localization based on segment-level retrieval. Using real-world data, we show how our proposed map representation c

estimation and substantially improves the mapping performance at different resolutions.

ving object segmentation and even retrieve previously missed moving objects in an online mapping scenario.
ly available, acquired using various sparse LiDAR sensors. We show that using our method for normal estimation leads to norm
ectional images and greatly benefit from the wide field of view (FOV) of 360-degree cameras to reconstruct various scenarios w
, our method achieves a higher Precision-Recall AUC for the evaluated buildings.
using a high-resolution LiDAR. In addition, we open-source the implementation of D-Map on GitHub to
map construction compared with state-of-the-art methods. We also validate the proposed framework in real-world environments.
elease an open-source CUDA/C++ implementation.

l design and passive pitch concept, we evaluated the testbed during underwater experiments in energetic conditions at a wave
main crossing and verified the vehicle’s endurance during each mission stage.
rates of 3.5 and 1 rd/s, respectively, were recorded. Our results suggest that Snapp’s swimming capabilities have excellen
guring various locomotive pattern of the articulatedcompliant legs, such as swimming forward, retreating, lateral swimming to th
s in marine operations, and contributes to highlight the importance of multidisciplinary collaborations among roboticists and scie
for the robot, which uses thrust to compensate for the interaction force.
roach for continuous spatiotemporal collision checks. The proposed formulation provides safe performance even in environmen
ormance, running at 24 FPS on a NVIDIA Jetson Xavier NX, 160 FPS on a NVIDIA RTX 2080 GPU and 7 FPS on a single Intel
e also demonstrate the real-world usability of our pose estimation pipeline on an underwater robotic manipulator in a reaching ta
erating correspondences for sonar images which will pave the way for more accurate loop closure constraints and sonar-based
so matches ensemble models in uncertainty predictions, all while being 2.5 times faster.
both synthetic and real-world data demonstrate that our method effectively restores true color from underwater imagery, outperf
ataset open up promising opportunities for future research in autonomous underwater cave exploration and mapping.
 ghtening strategy from the perspective of harnessing waves.
the gap between the capabilities of thruster-driven AUVs and gliders. These combined characteristics make ReefGlider ideal fo
% and 40.26% in simulations and field experiments, respectively. Although the control effort slightly increases by 7.78% and 4.20
ed that the SHOALBOT has excellent movement ability in the amphibious environment.
ups, as well as with comparisons with other state-of-the-art camera models. Additionally, we demonstrated the performance of

ng the number of points of interest in the mission area, and varying sea state on the mission duration are also analyzed.
ity is also demonstrated to improve ASV safety in highly congested environments.
Simulation experiments in a realistic underwater environment validate this approach. Experimental comparisons against a full be
ea-U-Foil first, then validate the effectiveness of its tri-modal locomotion through extensive experiments.
marine archaeology, and environmental monitoring.
the manipulator achieves a trajectory-tracking error of 1.03 mm in static water and 2.91 mm in flowing water at 1.2 m/s, which
work is validated in three different test-environments, namely a pool, a test-tank with a gantry for ground truth motion, and the flo

VIO due to the absence of ambient light and the confined nature of the environment, while also being a crucial source of fresh w

ation and experimental studies, covering ground-free and ground-bound configurations as well as both motion-only and manipu

e CoM position and showing satisfactory constant-attitude flight performance.

mprovement in position tracking accuracy. Our algorithm achieves 93.4% accuracy in real-time texture recognition, which furthe
during the jumping phase and realize a stable transition to flapping flight. This work lays the foundation for outdoor flight withou
ere the tiltrotor successfully performs perching and unperching on/from a vertical surface during flight. We further show effectiv
world experiments with different vehicles' configurations show the fidelity of the simulator results with respect to real-world settin

ower tolerance of the UAV. Specifically, a theoretic model for the coaxial rotors is proposed to analyze the induced velocity and

ximized with the use of the pod or of tether winding. These approaches can reduce energy consumption to only 22% and 1.5%,
orating multiple practical constraints of aerial manipulator.

thruster-to-thruster spacing, thruster protrusion from the fuselage, and inclusion of flanges or strakes. Peak thrust enhancemen
eighted Markov-Dubins problem. Moreover, the obtained list of candidate paths for the weighted Markov-Dubins problem reduc
 g surfaces at various inclinations. Results demonstrate that our proposed system enables a quadrotor to reliably perch on mov
ended range of flying capabilities.
ely align with flat work surfaces achieving full alignment through the rotation along non-actively controlled axes. Additionally, the

ce drop observed from simulation to reality shows potential for further improvement, including refining strategies to address the

gineering toolbox.
drastic disturbances. Numerical and real-world experiments demonstrate that the proposed robust control strategy has superior
hese results highlight the superior capabilities of the proposed MPS-based scheme when compared to a benchmark controller c

os demonstrates an impressive load-placement accuracy of less than 5cm.

tions. Moreover, the forces arising from robotic arm motions are delicately used as part of the drone’s power to save energ
erified empirically, both on a swarm of real nano-quadcopters and for large swarms in simulation. Overall, we provide a framew
ics-based simulations for different assembly configurations.

nd reliably in critical missions.

dated through several experiments.

with simultaneous forward and backward locomotion. We also observed and controlled a new side locomotion phenomenon not
re measured for an overall device length of 140 mm, including a maximal stroke length of 22 mm. The results confirm excellent

ed and a novel standardised procedure is proposed for robust characterisation and fair comparison of different actuation system

of an asymmetric arrangement of the front and rear Mecanum wheels, which differs from conventional symmetric arrangement
en proven to be effective through force test and distance jump experiments.
odes are proposed, compared and tested. The best one is applied to the sailboats towing system and improves its performance
97% and 67% after being cleaned.

nt is demonstrated by navigating and grasping an object in a curved pipe with a variable internal diameter.

nd even tissue engineering.

nchronization along with a low target lost rate (15.8%) and constant-force tracking (2.50$pm$1.02 N).
e, a straightforward planner is introduced to assess the adaptability of non-contact manipulation in a cluttered unstructured env

ess method for 3D pose tracking of trocar, achieved with only a single monocular camera. Our experiments show promising re
nd peak power when the user commanded high volitional input in late stance. This additional torque and power also allowed the
he 3D OCT volume. Our visualization is enriched by augmenting the uncertainty information on the instrument-aligned B-scan. T
fication of their position and depth. The TSER achieves an average motion speed of 9.8 mm/s in a bellows tube that contains s
s the most significant of these, with a focus on the technical constraints of the first-generation dVRK, which both existing and pr
haviour and accurate force control during ultrasound procedures, even in cases of contact loss.

osis and monitoring of abdominal aorta aneurysms as a representative use case. Our experiments demonstrate that our propos

posed approach, we manually extract the US cartilage point cloud from one volunteer and seven CT cartilage point clouds from
nd measured the detection rate of observation points used to evaluate the image quality of parasternal long-axis view. The resu
ignal, this work first introduces an online artery re-identification module to qualitatively evaluate the real-time segmentation resu

ast 93.9%. The designed sensor has higher accuracy, is compatible with magnetic resonance imaging (MRI), and accurately ide
mage. In this work we demonstrate that we can achieve a maximum motion error for the end effector (EE) of our ophthalmic rob
ly. Moreover, no instances of dura mater damage are observed and the robotic system exhibits a high level of autonomy as it p
and evaluated the accuracy and task space with simulations. The results indicated that by using the fingertip and wrist motions
utilizing dental stone phantom and cadaveric models. Our experimental results demonstrate that our proposed haptic feedback
e environment while avoiding dynamic pedestrians.
dden lymph nodes, and reduce the number of collisions (21% and 48% reduction rates using two different force models compar
unintended movements caused by accidental contacts are prevented, thus safeguarding operational safety. The proposed met
pecially when the driving wire displacement increases. For a 10 mm parallel robot, when the displacements of the two pairs of w
d a series of evaluation experiments were performed. The experimental results verified the feasibility of the designed haptic inte
spired robotic approach. Furthermore, using a novel backlash-free coupling method, the robotic arm could be easily attached a

active attraction and separation under simple magnetic fields. The reprogramming process of the RMCR is theoretically investig
y, ex vivo experiments are carried out on fresh porcine intestines.

estimations can pose serious risks to the patient. We present a robust method that not only improves the precision upon existin

s found that changing the control frequency of the panel can help the robot to pass through different morphometrics, in particula

work used during experimental validation and they yielded an average accuracy of 96.5%. For unsupervised classification, both
e introduced a new data synthesis pipeline that used physics-based catheter insertion simulations, along with a convolutional ra

ncing the automation, safety, integration, and overall efficiency of robot-assisted micro-surgeries.
d with Fiber Bragg Grating (FBG) sensors attached to the SHER 2.1 end-effector. This adaptive sclera force control algorithm a
sk completion is significantly improved, while the variance of average speed and the total distance for robot operation is reduce
mplifying the therapeutic capacities of capsule endoscopes without necessitating a size compromise.

gnets. This adaptable configuration is uniquely designed to withstand the dynamic squeezing environment prevalent in real-wor
TG gets the average and maximum errors of force sensing approximately 37 mN and 223 mN, respectively. Finally, in the exper
vironments with contact forces at different locations. The experimental results in terms of dynamic shape-force sensing accurac
ling planning and execution of robot trajectories in accordance with the proposed glossectomy workflow. The system successfu

em due to its user-friendliness and effort reduction.

uired working area on an adjustable phantom model.
workflow and improve complex endovascular aortic repair procedures.

nside the depth of skull in a robotized manner effectively.

thod. The proposed approaches eliminate the need for visual feedback and thereby minimizing the risk of radiation exposure fo
micro-catheter.

by the NASA TLX questionnaire.

nments. Furthermore, the introduction of air bubble enhances ultrasound imaging, facilitating further in vivo applications. These

fast self-propulsion of continuously rotating 30 µm prototypes by the steering in a glucose solution was demonstrated as expe

ess rate, efficiency, reset time, and the Index of Pupillary Activity (IPA), an optimized motion strategy based on the ``palpation"
cro-suture. We finally tested the mechanical strength of the knots for potential medical assessment.
alue of the depth prediction error is less than 0.3 $mu$m, which proves the accuracy and effectiveness of the method. This dee
stem demonstrates robust versatility and can serve as a reference for other optical systems.

, the 2-DoF origami transmission mechanism is the first parallel mechanism designed for the insect-scale robot and the singula
rgeries in the bladder and some blood vessels), as well as the surgeries on small animals. But they are insufficient for operation
oss-scale feature map guide, and multi-scale feature fusion, to preserve essential sperm details despite their small size. Experi
, blur, and low illumination.
curacy of the capillary tip on the order of 10 um. The imaging technique could detect the contact to the elastic objects and cell
nd the range of any type of AFM with a dual XY stage setup. Thus, it opens the door for high-resolution long-range AFM by add
acking and positioning of the target sperm three-dimensionally. Experimental results revealed the system achieved a success r
also proposed, in which an outlier filtering method and endpoint detection algorithm are designed to accurately reconstruct tail e
monstrates the effectiveness of our proposed method and its potential to be deployed in other point-by-point scanning systems.
ore oocyte injection. Our study provides an easy-to-use technique for oocyte manipulation without contact, and it has the potent
y images to keep the UMR coupled with the external field during wireless teleoperations, resulting in a success rate of 76.6% wh
sted by approximately 21% in cluttered mixed scenes.
t-solvers for both object-oriented POMDP and MDP analogues with the same task specification. We then apply the formulation
me in challenging environments. We will release our code to benefit the community.
and different robot kinematics. These capabilities open the door to a wide range of downstream tasks across embodied AI and
s and 7.9 m/s^2, respectively. Notably, we demonstrate a projectile avoidance task in which the robot avoids a thrown ball whil
compared to several baseline methods. It is shown that it outperforms standard MPC and an existing AMPC approach on sever
t our method attains a good balance between observation and manipulation, yielding high performance under various grasping
pled control.
ibution of various utilities and parameters, and compare against representative baselines both with and without active perceptio
on. We demonstrate the effectiveness of our method through a series of cha
e available on our project page at: https://sites.google.com/andrew.cmu.edu/sculptbot
ehicle at its limits. Furthermore, the controller can adapt to changing environments by training the NN during the race. The MPC
pair outperforms the baseline and ensures robust high-speed driving performance without control failure.
to precision.Notably, we are capable of optimizing the gaits of the Purcell swimmer using approximately 10 cycles per link, whic

ne, and without prior experience with the system. An adaptive model predictive controller is developed with the composite mod
real-world quadrotor carrying a slung load was curated and is made available. Prediction results are presented and corroborate
ifferent terrains while significantly reducing errors in linear and angular motion predictions. As shown in the attached video, we

% on average.
hmarks are released at url{https://dingrui-wang.github.io/ESP-Dataset/
ynthetic-to-real experiments on NTU RGB+D and Toyota Smarthome.
e conditions such as grasp width, depth, position, and angle are also predicted. Second, with the grasp width and depth of the o

ctive robots.
nsive experiments on popular benchmarks 3DPW and Human3.6M illustrate the favorable trade-off between the accuracy and c
port. We based this adjustment on the estimation delay, the muscle saturation, the joint kinematic mean power frequency, and e

e tasks in the 3D space.

umber of collision avoidance constraints. Our experimental results demonstrate >98% solve success rates within complex, high

e simulation results show that our proposed approach achieves an average length error of 1.07% over the continuum's centerli
s, which significantly improves the overall system's operational efficiency with reduced computational complexity and easier con

5° to less than 1°. The proposed method paves a new way for tremor suppression exoskeletons.
ters for a 130 millimeter long pneumatic robot.
 s show that the proposed learning-based real-time controller achieves performance comparable to an optimization-based non-r

ments, and an accuracy of 0.72 mm (1.11% arclength) and 0.72 mm (1.01% arclength) for tracking a square trajectory and a circ
ovides robustness and stability to the pipeline. POE-M reduced the maximum Chamfer distance error by 23.1% compared to th
nt micromachines. Finally, we propose an approach of numerically optimizing control signals by iteratively refining models, whic
ess" force control.

development of personalized rehabilitation technologies.

e of the tangent space of the internal configuration space). To invoke this representation (termed a perturbation map) we assum

strated autonomous capabilities without any human intervention, even in case of collision or planner failures.
er fails. Our approach can adapt to various initial conditions and environmental parameters. We conduct experiments on six tas
nificant for applications such as ship routing, where route evaluation is a desirable capability.
creates a sparse graph by clustering an initially sampled dense roadmap. Such a reduced roadmap allows fast identification of
g-based SOC algorithms in terms of path-tracking and obstacle-avoidance capabilities.
methods. Lastly, we provide necessary algorithms and an open-source implementation of the proposed methods.

ich ensures continuous exploration and guarantees spatial coverage.

ting planning algorithms. Compared to Informed RRT* Connect, RBI-RRT* reduced the computation time of achieving a specifi
ent operation scenarios were conducted and the results show that the proposed planner outperforms Euclidean space planners

preserving probabilistic completeness and asymptotic optimality. We deploy our method on a mobile robot and demonstrate re

™s performance in terms of error, we conducted a comprehensive analysis of its regret bounds, comparing it to an Oracle value

using ground vehicles, among numerous others applications.

ates an initial solution within a user-defined time threshold. Furthermore, functioning as an anytime algorithm, it iteratively refine
average current consumption by up to 2.9%, and decrease the normalized trajectory length by up to 11.2% compared to existi
es that satisfy the scene-specific constraints. Further, instead of a single cost function that may be insufficient in capturing diver

mplementing heuristics that concentrate future samples into improving the path quality of the nominal solution. The algorithms
heading changes and improve the path smoothness. Comparative experiments are then carried out to validate the superiority o

% compared to pure volumetric collision checking and increases safety by avoiding high-risk regions. We deploy the SMUG pla
 super-extreme flight trajectory through race tracks.
mics-aware backends. The metrics we designed faithfully capture the planning difficulty in a given environment. All code is avai

e rule constraint and its constraint transcription method are proposed. A large number of simulation and real-world experiments

m orienteering problem in the monitoring task compared to the existing solutions. The proposed approach has been experiment

one, providing a pseudo mockup scenario by virtual material deposition, helping to identify technological gaps from simulation t
requiring any backlash model information. The effectiveness and simplicity of the developed control strategy are verified throug

he proposed controller and continue to fly even if two adj

ika Panda robot.
rimentally demonstrate JetRacer-4WS’ ability to perform controller auto-tuning and illustrate the advantage of four-wheel st
le surrogates. The simulations using a quadrotor model are employed to demonstrate the effectiveness and advantages of the

he issue of mechanical resonance between the manipulator and the applied load forces.

iments involving a quadruped robot equipped with a roller arm, successfully navigating and transitioning between different leve
urce code is available at https://github.com/LongJin-lab/ARCM.

ows the RW-Robot to perform flexible spatial locomotion from simple user input.
ensional robots.
ing a representative precision motion system comprising a planner stage and a uni-axial fine stage under step-and-scan traject

ished through a motion fusion mechanism, which is achieved via optimizing evaluation functions that are reliant on future states

ontrol of CDSM within a unified framework of hybrid motion/force control. Through experiments and simulations, we demonstra
ode are available at https://sites.google.com/view/ymchen/research/rl-for-roms.
ng into account the longitudinal and side slip effects. The open-loop simulation shows the proposed Koopman bilinear model ca
omparison by simulation and real robot experiments, we confirm that our MBRL method can achieve high data efficiency for lea
h prior model, the multi-contact MPC controller adapts itself on-the-fly with an adaption rate around 20 Hz and successfully man
countered on surfaced ice. DRIVE is an efficient way of characterizing UGV motion in its operational conditions. Our code and
dinates can be defined on these orthogonal foliations and used as additional task variables for control. We show in simulations t

h tasks, further validating the applicability of the snake-like manipulator.

e demonstration using a physical robot.

 with a high-dimensional 6-DOF robot manipulator. In these complex settings, our method exhibits high success rates and finds

nalyze the transferability of attacks to different planners, and discuss underlying assumptions and future research directions. Ov
ble information in stochastic spatiotemporal environments. 2) Second, instead of directly tracking the ergodic trajectory, we intr
open source software are available at https://github.com/TAU-CGL/tr-rrt-public. This research sheds light on effective approach
g direction to be admissible during its movement is provided in this framework. The proposed algorithm repeatedly deforms the
ng over-conservative. Given state-based input, our neural CBF-induced neural controller-enhanced RRT (CBF-INC-RRT) can in
experiment and simulations.

The results show that computing and executing hybrid locomotion plans is possible on hardwar
ingle UGV exploration system, our approach reduces the overall path risk by 26.8% in simulated experiments, showing that the

ce by the group and the gaze behaviour decreased psychological reactance.

pose an occlusion-resilient multi-sensor fusion technique by positioning visual perception sensors in distinct locations on the rob
dules in the complicated system. Extensive experiments are conducted in both simulation and the real world demonstrating that
mmersion and efficiency of interactions between humans and robots.

ermore, CAMInterHand employs the cooperative attention fusion module to fuse all features from multi-view images, enhancing

nal, the proposed intention and attention recognition algorithm can activate safety features to prevent unsafe interactions. We a
odes and demos are available on https://jxu124.github.io/TiO/.
ds. We further provide more findings on the impact of diverse user feedback forms on the agents’ performance.
object discovery) and online (i.e., IOG with a physical robot arm) settings. Code and data are available at https://github.com/gic
ted through real-contact experiments, enhancing EIT tactile systems for human-robot interaction by incorporating physical know
d pick-and-place tabletop simulation environment. Finally, we demonstrate the proposed approach in real-world human-robot int

presented a considerable lag in the time response, it was more attractive than the standard one. However, the user command in

d server, which processes the data using VGMN for multi-agent geolocation optimization, including initialization, pose graph opt
ce in large-scale missions, where ten robots collaborate seamlessly in performing SLAM tasks. To benefit the community, we w

segraph optimization after assembling the segments to smooth the merged map globally. We demonstrate AutoMe

approach also enables voluntary waiting by introducing an idle task that the robots can select to wait. Experimental results have
given the predicted target trajectories with uncertainty measure in the road-networks; (ii) the nonlinear model predictive control (
human-in-the-loop, where a user can communicate and collaborate with a robot agent to complete tasks together. Project webs
tacle-avoiding control policies trained with end-to-end DRL that transfers zero-shot to real quadrotors. Our approach scales to 3
ow to quickly follow the existing collaboration and communication rules, and fast adapt to their roles in the team. The experimen
nce on a challenging problem for which the base policy cannot reach a terminal state. Despite the challenging experimental setu

on method, multi-agents using OWC with general FOV show comparable performance to those with radio wave like full commun
ount of time required to fulfill the task, results have shown to be comparable to the ideal scenario in terms of final configuration

em operates on real robots with limited onboard computational resources.

ur decentralized STGNN model based on the expert algorithm. We simulate the proposed STGNN approach in various settings
tensive simulation tests present prominent benefits of time synchronization on mutual localization. Moreover, real-world experim
autonomous robots. We show a communication-aware planning algorithm for a high-altitude aerial robot executing a collaborati
viding a complete system from the offline algorithm to the field execution.
work will allow the cells to freely roll/rotate about one another so that they may realize the full potential of their unique shape.

nd-and-decontaminate mission with a team of three heterogeneous robots and analyzed the utilization and overall mission time

objects. In contrast, the number of vertices and edges in graph representations scales exponentially with either.
namically uncertain environment.
presentative of most scheduling algorithms aboard Earth science missions today as well as an intractable upper bound. We sh
ts1, confirming its efficacy and potential for real-world applications.
aining task execution performance similar to an approach in which each robot always includes all other robots in its MPC comp
nd 70%-75% on inspection speed.

e different action selection policies, focusing on scenarios where targets outnumber agents. In simulation, we demonstrate that

output are used to construct a LogNormal distribution matrix from which positive bigraph weights can be drawn. The performan
insights on some of the challenges/opportunities from the perspective of a decade-long industry-university R&D partnership.
phs difficulties (load/unload locations, weak connections,), a predefined roadmap (constrained by the plant layout), vehicles err

Simulation results are reported to prove the effectiveness of this approach.

tion for singularity-free control maintenance is provided.
ng strategies towards the optimum. Complexity analysis indicates the polynomial computational complexity of EM-Patroller, and
w up to 55% improvement in median of fitness evaluations against an empirically chosen parameter set.
-robot collision, then 2) identifies ordering constraints between block placement and removal actions, and finally (3) computes c
and success rate (5%-10% improvement) over benchmark methods. The outcomes of this study have great significance in pro

monstrated successfully.
n various metrics.
of the expert’s actions. We develop both non-learning and visual behavior cloning-based controllers allowing HIRO Hand s

ehaviour within a mixed dataset and then utilize it to isolate the expert data. To further boost the BC performance, we take adva

are conducted to substantiate the effectiveness of the present multiple UAV-USV landing control law.

heme is experimentally validated with a 5G-enabled UAV in a real-world 5G stand-alone (SA) network.
e pushing interactions and approximately 4 million camera images is collected and serves as a proof of concept demonstrating

sumer hardware and not requiring the initial pose guess.

25/0.25 the average prediction precision (AP) of easy-level objects (cars, pedestrians and cyclists) reaches at competitive 52.2
pth denoising mechanism in the training phase. The entire model training is accomplished through an end-to-end manner. We p
hat are robust to weather corruption. Extensive experiments demonstrate that our method outperforms other methods for objec
ust towards inaccurate camera calibration and dynamic camera setups.
I 3D object detection benchmarks to similar methods that are not uncertainty-aware, making it suitable for real-time edge robot
tps://github.com/goldoak/DMSR.

that simultaneously provides pixel-wise anomaly scores and objectness scores. Our experimental results show that the objectn

raining. We evaluate our method on two widely used indoor detection benchmarks. By plugging our proposed modules into the
nts on motion model selection.
the attentive pillar aggregation module designed to selectively aggregate essential pillar features. Extensive experiments condu
comparing to baseline, our proposed model reduces the computational overhead by 70% while improving mIoU by 4%.
mance of Shift-SSD in both detection accuracy and runtime efficiency.

he-art approaches on several benchmarks, including a 4.7 mAP improvement when facing challenging class distribution shift.
ve experiments on nuScenes validate the effectiveness and efficiency of our proposed module, which achieves an increase of 1
mproves the overall detection performance. Our model can be used plug-and-play, showing consistent performance gain.
nd [email protected] by 9.38$%$ on VisDrone-2019-DET dataset.
d structure measurements by an average of 120% and can successfully assist real-world UIBVS control in an unstructured man
ements, with F1 score enhancements of 7%, 3%, and 1% on the FL-Drones, AOT, and NPS-Drones datasets, respectively. Fur
h a negligible additional latency of 9.66 ms and a small storage cost.

. We demonstrate the effectiveness of our proposed interactive perception pipeline in accurately segmenting cluttered scenes b
age segmentation model for detecting objects in the environment and a data association framework based on geometric consist

ack of clear interpretability and effective utilization of uncertainties for downstream tasks. We tackle these issues through the in
ur approach generalizes to perception tasks, including robotic hand manipulation and human-object pose estimation in the wild.

classify dynamic objects. Besides, we also propose a dataset of the low-resolution point clouds and manually annotate the gro
strate its generality on a novel data set with rare moving objects in complex environments. We make our efficient implementati
act performance, along with the underlying causes.
of-the-art results on nuScenes benchmark.

ures of objects in point clouds, while global equivariant features are extracted in both point clouds and images. Furthermore, the

ate-of-the-art methods.
ment local semantics and incorporate global long-range dependencies through multi-scale window attention. Compared with ex

n all evaluation tasks.

benchmark. The code will be made publicly available.

n (mIoU) gains of 4.55% and 6.38%, respectively, compared to the state-of-the-art active learning method Pixelpick on the City

of disc intersections. Finally we propose a semi-supervised learning approach to utilize all frames within our dataset and improv
addressing the cold start problem effectively. Results show that we are able to improve the performance of the initial model by a
acy in perturbed environments compared to other prediction algorithms. In addition, the proposed scheme effectively achieves d
on capability of the controller and simulation results highlight its effectiveness.

duction in the number of resulting linear system solver iterations.

s for time-optimal trajectory replanning and trajectory tracking.

recognition and generalization of constrained motions, such as during object manipulation in contact.
e that (1) we achieve better real- world robot performance by using quadratic objectives in place of the minimum-time objective,
 are used, each having 10 DoF, and the results confirm the smoothness of the generated solutions.

r proposed method on a double pendulum equipped with two of our newly proposed clutch-based Bi-Stiffness Actuators (BSA).
ents that the method can generate motion plans and control for robo
edictive control.

first time that a SOS-based time-invariant controller can swing up and stabilize a cart-pole, and push the planar slider to the de

at capabilities of CACTO in escaping local minima, while being more computationally efficient than the Deep Deterministic Polic

s of (i) an inverted pendulum aiming to stay upright, and (ii) a quadrotor aiming to fly to a goal location through an unknown clut

and input-to-output stability, it is guaranteed that the learned robust optimal control gain is stabilizing and that the solution of th
art-pole. Extensive simulations of off-nominal initial conditions demonstrate improved robustness, while real-system experiment
mpare our method in kinematic simulations with the state-of-the-art and demonstrate the applicability of an online exploration tas

to evaluate the penalty of black-box constraint.

essfully accomplishes the guidance task and notably reduces the leader's planning time to nearly half when contrasted with the
Problem (ASPP). We present a generalized permanent labeling algorithm (GPLA) to find an optimal solution for the ASPP. We a
on models leads to 40%, 51%, and 46% improvement in coverage performance as measured by the ergodic metric, and 15%,
google.com/view/samarl
scenarios with directed lanes.

ethods (two MAPF algorithms and a loosely-coupled coordination method based on precedence constraints). We evaluate each

n in mission completion time, and an average $23times$ reduction in per-agent computation time compared to SFC-based app

s show that our method is capable of solving challenging tasks with a higher success rate and lower cost than the existing state
ts at the whole system level. Our neural structure relies on a Graph Transformer architecture to allow agents to selectively com
n five 2D grid benchmark workspaces with 10 Quadcopters and 10 TurtleBots, respectively. In addition, to establish its practical

e it with two state-of-the-art MCPP planners on various instances, demonstrating a reduction in the coverage time by an averag

out the outdoor experiment, the feedback provided by the DVS is incorporated into a Nonlinear Model Predictive Control (NMPC
mparison with the state-of-the-art methods. Finally, the proposed approach is validated in a real-world experiment.
hat incorporating this synthetic data significantly reduces reliance on real data while maintaining high mean Average Precision (
eaches 94.6 mAP on the test set, outperforming the individual sensor networks by roughly 5%. We also described an implemen
ull- BA and ICE-BA show that Jetson-SLAM is the fastest available accurate and GPU-accelerated SLAM system (Fig. 1).
dition, we will open-source the implementation of Air Bumper on GitHub.
the functionality of our approach via a flight demonstration using the Crazyflie 2.1 micro-quadrotor. Without a priori obstacle kno
approach yields a superior tradeoff in map accuracy and size when compared to state-of-the-art mapping methodologies (both

putes over 10 times faster compared to state-of-the-art methods with shorter path and more complete coverage. The source cod

egy is shown to strike a great balance between cautious exploration of yet completely unknown regions and more aggressive e

.36% in real-world experiments, with gap orientations up to 60deg. To the best of our knowledge, this is the first paper that perf
. In contrast to existing perception-based safe control frameworks, we directly close the perception-action loop and demonstrat
wo different sources, like wind and a drone’s relative motion.
al and auxiliary learning-based module to enhance the obstacle detection range and dynamic obstacle identification. The propo
y improve state estimation accuracy, with root mean square error (RMSE) reduced by up to an order of magnitude. The source

method, we can determine the correct number of encountered updrafts with an accuracy of 99.99% instead of 79.50%, significan
ratures are maintained below 50â—¦C even on an edge- computing device, highlighting the practicality of our approach for long

olving scene geometry, resulting in a comprehensive and up-to-date point cloud representation. By circumventing the constrain

he robot's performance. by providing with the optimal end-effector pose while also reducing the stress on the motors and the ov
al environment, and superior results are achieved which demonstrate the effectiveness of our method.
s, and a Mixnet learns how to combine background and residual NeRFs. We contribute synthetic and real experiments that sug
essing step to reject false edges. We experimentally validate our framework using real-world wiry objects and demonstrate a m

8.9% and 91.2% detection accuracy respectively, which are superior to comparative methods. Besides, the effectiveness in our

mplemented our system on a real robot and performed several autonomous closed-loop experiments in which the robot had to
ure the 3D deformation of the surface via the movement of 82 colored markers. We trained a Convolutional Neural Network (CN
ts with symmetric or near-symmetric intercepts. SCAlign significantly outperforms the baselines on this challenging dataset. Ad

arm system. For dynamic grasping, we demonstrate catching swimming robot fish in the water.
to bridge the labor gap in agriculture.
tter and from shelves.
rent types of liquids (8 liquids + air), and 97.65% in estimating the liquid state. We investigate the effectiveness of our system w
of effective data throughput and success rate, and show that this approach leads to grasping policies that outperform policies tr

d to capitalize on the advantages of multi-view observations. Simulation results demonstrate that the efficiency of our system ou

ts the desired point to choose from the given point cloud and then the correct displacement to achieve a smoothed cloth. Exper
jectory data compression as an additional benefit. Experimental results show our method provides a valid and accurate means

sks, maintaining the leaderboards, and serving as a contact point for the community is also provided.

tics of a given scenario, and admissibility, by considering the structure of roads and kinematic compliancy, motivated by real-w
w benchmark dataset will facilitate advances in the field of camera-only 3D detection by providing more informative signals that

iment and a survey involving 12 participants practicing key Iaido techniques with guided training from Sawyer. Our results dem
ed on the muscle activation feedback. Finally, the Robo-Coach is implemented with a variable impedance controller to achieve l

te the performance of the proposed controller. The control results are also compared to state-of-the-art control methods, highlig

subjects performed 120 squats per session, followed by a maximum voluntary contraction (MVC) task. No adverse events, suc
hree groups such that the subjects used either predominantly their hip joint, knee joint, or a combination of both joints implying
nd robust than the rating scale method. Compared to RS, PC fosters authentic participant interests in the experiment by intuitiv

work. With our newly introduced test set, we show that our model can better scale to a large variety of unseen objects and hum
describe the details of the control framework, feature engineering and training process of the intent recognition. The proposed c
human modeling and robotic methods in physical human-robot interaction can lead to both improved human understanding and

n inadequate control strategy is deployed. Furthermore, through ANOVA, it is discovered that the influence of base compliance
man-robot interaction behavior. These findings highlight the advantages of multi-body simulations as a tool to perform modular,
y subjects exhibited a significant decrease in energy consumption and jerk when using WANDER with PBO parameters as well
adaptively based on the quality indicator of the sub-map merging process. Additionally, our approach extends the recent single

ared to state-of-the-art robust MPPI variants and stochastic MPC methods.

simulations.

w these bounds can both be used to certify any pruning heuristic in retrospect and propose a novel approach to determine whic
 ark domains demonstrate that the proposed shielding methods successfully guarantee safety (unlike the baseline POMCP with

CPOMDPs while non-hierarchical baselines cannot.

nt methods. Furthermore, we illustrate that our designed uncertainty field is ideally suited for real-world robotics tasks, such as
for each executed transition and avoid new transitions with context similar to previous unreliable ones. This is helpful for levera
ploration and collision avoidance.

nts. Our numerical investigation confirms that IBBT finds non-trivial motion plans and is faster compared with previous similar m
ency as well as robustness with respect to out-of-distribution human motion. With the parallelization of a sampling-based optimi

EAMS), outperforms other determinization-based approaches in terms of combined traversal time and collision cost. https://sites
tual maps in a simulated environment that reproduces environmental conditions common in robotics applications. Our results s
hods, highlighting the framework's effectiveness in stochastic, uncertain, time-varying domains.
ow that dynamically selecting SLAM front ends, even myopically, improves localization robustness compared to selecting a sta

ciently construct the database. We show that the developed algorithm speeds up database creation by over a factor of 100, ma
cenario-based planning approach in comparisons to state-of-the-art methods and validate our framework in an experiment with

20 % reduction of the probability of mobility hazards in real challenging terrains.

d into a second-order cone programming problem to achieve real-time performance. Through simulations and physical flight ex
25/3 nodes/agents, using a fraction of the time needed for J S G to solve such complex problems; Second, we learn a general

nches of the tree explore different homotopy classes of paths, and all the branches are locally shortest. Comparative experime

schedules to solve such problems, by casting the problem in the reinforcement learning framework. Our approach relies on an
at our algorithm is computationally efficient while improving average makespan compared to other baselines.
are unavoided in practical systems. We validate our effectiveness in simulations with over 1000 robots, what’s more, real e

ed actuator is sufficient to generate the joint torque needed for walking.

across all eight bone areas with the only exception of one channel in the ankle. This method enables the robust measurement o
approach, particularly in reducing the EMG levels of the user during different tasks by providing exoskeleton assistance as nee

progression and task. Fourier series coefficients are determined via convex optimization such that the controller best reproduce
ce the peak power of the motor by 24%.
namic forces, applied at the joint level can alter the inter-joint coordination, while task performance remains unaffected. The tes
versatile finger movements of the prosthesis hand. The concept of motor neuron discharge timings was evaluated through expe
up to 40% during lifting, without increasing the muscle activation during walking.
Furthermore, the proposed wrist can be utilized as a robotic wrist, affording two additional DOFs, the ability to lift loads more tha
nt of more flexible and adaptable ankle exoskeletons for walking assistance.
uracy in detecting transitions between activities. These promising results show the robustness and reliability of the method, rein
s employing the exoskeleton in zero-torque mode (without assistance) exhibit an alignment with the exoskeleton's recovery ass
t users of some devices, there may be a range of nearly equally effective torque profiles to reduce the metabolic cost of walkin
e ankle torques and joint trajectories despite not tracking explicit joint-level references in stance.
ation within a built-in structure of the robotic exosuit. The findings confirmed that the suggested solution notably minimized mus

tionRobotique/libPeR_base.
acy and applicability of RadCloud on commercially available UAVs and UGVs, with off-the-shelf radar platforms on-board.
0. We achieve accurate metric estimates comparable with state-of-the-art methods with almost half the representation size, spe

ry to the removal of the bias from the measurements, we demonstrate that the depth correction models help to reduce localizat

pability of being applied in the unseen object reconstruction tasks.

t navigates through an indoor obstacle course. Our complete implementation, called waverider, is made available as open sour

l input sequence can be approximated by a Gaussian distribution. To overcome this problem, we further apply SVGD to MPPI w

ng during femoroplasty. Results from these use cases demonstrate its potential to extend to a wider range of medical scenarios
ailable. To reproduce manipulation skills using the learned set of prioritized geometric constraints in novel scenes, we introduce

ser, enhancing its efficiency. These findings underline the importance of performing proper optimization while designing exoske
method provides a framework to identify and characterize the components that must be redesigned to lower exoskeleton weigh
matic pressure to generate force and compensated for kinetic friction using electromagnetic forces. The results showed that the
ans of Reversible Dynamic Movement Primitives. The system allows for self-paced rehabilitation exercises by back-and-forth m

ded to other control systems, propelling RL towards general uses in the real world.
uscle fatigue when the hybrid adaptive controller is used, compared to its non adaptive equivalent. This suggests that our hybrid
ngaged throughout a session and augment physical exercise with cognitive exercise. We conducted a user study for which peo
. Furthermore, we validate the practicality and effectiveness of RbSL by deploying it on a real Panda manipulator. Code is avail

mpirically demonstrate the efficacy of ASRL, and assess generalization and scalability to out-of-distribution scenarios.

ned tradeoff. The experimental results obtained in seven different robotic environments indicate a considerable reduction in the
k and the DeepMind Control Suite (DMC) respectively. These savings also enable large-scale perceptive DRL that previously re

ed on the OpenLane dataset demonstrate the superiority of Bi²Lane, achieving a notable F1 score of 63.8% using a simple Re
y defining the problem and then examine the effectiveness of traditional control methods,on-policy and off-policy reinforcement
ciency, which we demonstrate in various control learning environments.
rmation to potentially expose private data. To validate our approach, we conducted experiments on the AI2THOR simulator and
uous actions, which highlights the distinct strengths of each algorithm in different scenarios.
method can benefit more from both inaccurate simulator and limited offline datasets to achieve better performance than the sta
the first robust RL approach that uses adaptive adversaries to tackle uncertain disturbances in drone tasks.
st HDC-based framework deployed in real-world settings.
se insertion scenarios achieves an 85.6% success rate in simulation and 66.67% in real robot experiments. Please refer to our
query for the event detector. Simulations in a 2D grid world and image-input robotic inspection environments show that our me
noid robot interacts with 100 objects across 10 exploratory behaviors, we demonstrate the versatility of MOSAIC in two task fam
objects. The results show that our framework achieves the best performance among all presented methods and is, on average,
nd them with an adaptive blending operation. For each region of the scene, the blending operation gives larger blending weight

so discussing limitations to provide future direction to the research community on open challenges in utilizing neural fields in un
set of patches to learn local patterns, and a lightweight decoder is designed to leverage these visible patterns to infer the missin
representations for robotics, as well as prior object-centric representations that are typically trained from scratch.
rence time. Finally, our simulation-trained framework shows promising hanging results in the real world. For videos and supplem

ach slot conditioned on robot actions. We demonstrate the effectiveness of SlotTransport in learning object-centric features tha

iptors. We also evaluate our method using our real robot experiment. We notice an average of 49% reduction in the image sea

46% lower mean global SDF error across all test scenes than a state of the art continuous representation, and a 30% lower err
os: https://sites.google.com/view/robot-soft/

to generalize effectively across various uneven terrains while maintaining real-time performance, achieving frame rates exceed
d data is available at https://samsunglabs.github.io/RIC-project-page/.
nvironment with severe occlusions.

ent solutions. Our proposed modalities are cost-effective and highly adaptable, allowing users to experiment and implement UA
h its lightweight design (0.4 MB memory requirement), our method is extremely suitable for mobile and robotic applications. Ad

erent models, we designed an adaptive curriculum scheduler to automatically search for the best timing for course switching. In
idth, and vehicle sensing among connected autonomous vehicles. The experimental results show that our representation achie
M) and establish theoretical conditions under which the EM approach converges to the ground truth. We evaluate the approach

proprietary-distance-estimate baseline by an order of magnitude in most scenarios. We demonstrate a simple robotics applicatio
e classic dynamic point cloud backbone and propose an improved version of PSTNet, im-PSTNet, which can efficiently model p
complex reasoning over spatial and semantic concepts. To explore the full scope of our experiments and results, we encourage
 holds and can dynamically focus training on the hard classes. Our experiments on the large-scale ScanNet200 benchmark sho

ght as it tracks the sun. In this paper we discuss the design, manufacturing, and control of OriTrack. We then compare OriTrack

ce of 92.67 N (up to 135.1 times of its mass) in hanging mode. Finally, the gripper was integrated into a robot arm and successf
cribing each singularity type using Grassmann line geometry. Twelve different singularity configurations are identified and descr

hibits positive transfer and improves the capabilities of multiple robots by leveraging experience from other platforms.

t of in-domain data, and can generalize to tasks not seen during training. We show that our learned system can perform over 16

e learned FT-NCBF, we synthesize a control input and formally prove the safety guarantee provided by our approach. We dem

m to obtain the maneuver set.

onstraints as a constrained optimization problem, which can be effectively solved by mainstream gradient-based optimizers than
ting our approach's ability to perform online reachability calculations for an unmanned surface vehicle subject to currents and a
entum from straight punches by Olympic boxers (26.506 kgâ‹…m/s). Moreover, the case study highlights other anti-intuitive ob
xiing system that uses a vision-based controller for taxiing. Our results show the efficacy of the proposed approach in identifying
c safety controller exhibits less conservative behavior while retaining theoretical safety properties.

an control invariance. Safety is instead guaranteed under an even weaker assumption on the safe set, triggering a safe task-ab

xploiting the unique low inertia and clean dynamics properties of MR actuators.

man, and safety algorithms to continually choose a backup controller in real-time. We demonstrate our method's efficacy on a d
estimate scaling, we leverage scale information contained in the vector norm of the equivariant features. We evaluate the effect
entation to aid OOD detection by training a decoder to reconstruct input image patches. Third, we perform theoretical analysis o
nnecessary computation. The proposed SR-VINS is validated extensively through numerical studies, demonstrating that when
on Waymo dataset.

exhibit significantly improved compositional generalization at test time to unseen tasks. Furthermore, UVD-discovered subgoa
 nt cloud, generates an axis-aligned bounding box (AABB), and estimates the strawberry weight based on the bounding box and
omprehensive experiments are conducted to validate the efficacy and performance of our AV4GAInsp system, achieving an ave
ene, to a dynamic scene of plants undergoing growth. Another contribution is a thorough evaluation of our registration method

gmentation accuracy on sugar beet breeding plots planted under the supervision of the German federal office for plant varieties

ers the difficulty of correctly segmenting the points close to the stem, where all the leaves petiole overlap. The experiments pres
mplete robot autonomy in physical specimen sampling and transport in the field.

arn to distinguish between fruit still on the trees and fruit that have fallen on the ground, and they can effectively transfer the kn
e skeleton graphs in developing an algorithm to perform correspondence estimation among the skeleton nodes using a cosine s

rial manipulator using the proposed controller and two different existing controllers. The comparative results confirm that our co
free navigation. Second, we compare our custom CEM optimizer with conventional variants and show significantly reduced coll

mobile robot. We validate and evaluate the performance of Osiris using real-world data collected at several farms and show that

y, the robot and proposed method were verified in a plant factory. The experimental findings demonstrate that the developed ha
from 0. It was confirmed that inverse optimal control is superior to sliding-mode control and is more robust to environmental ch
results with the MSE value of 0.08, the SD value of 0.15, and with testing results from the field showing up to ± 0.3 deviation f
e on fruit trees and represents a useful base platform to be improved in the future.
mation on-the-go and in real-time by employing only onboard computational and sensing resources.
ollected by a real robotic system in the field.
cclusion and improve perception. Through simulation experiments, we showed that our planner can handle occlusions and imp
obot, without task-specific programming.
er Potato 2022 dataset consisting of over 16,800 images for object detection in potato fields. Extensive evaluations of our prop
ional manual labeling.
n conducted outdoors, on a commercial log loader retrofitted for autonomy. Our results demonstrate an overall 87% success ra

m or strong correlation to reduce prediction error by a factor of 1.4--3.4 within the first 7 samples, and poor hypotheses are quick
proof and empirical results collected from different simulators demonstrate that the proposed control enables the robots to con

optimal to match the sampling time of the RL algorithm to the delays of the wheel loader's hydraulic actuators.
mpling and domain parameter estimation with a sensitivity-aware importance weight. In the proposed adversarial domain samp
ing module to reconstruct underground targets and create a 3D map for better visualization. Field testing demonstrates that the
manipulation tasks specified by abstract language instructions. We also demonstrate successful generalization to the novel but s
robotic experiments. Our dataset and code are available at https://grasp-anything-2023.github.io.

ollow natural language commands and robustly recover from failures.

ment with a swarm of robots.

ation through two numerical simulations and corresponding real-world robot tasks, demonstrating significant performance impr

3 scale race car and a full-size car, and also through experiments with a physical F1/10 autonomous race car.
ed robot. These experiments demonstrate the ability of the approach to significantly improve performance in face of uncertainti

the physical safety of humanoid service mobile robots operating in dynamic human environments. It enables the digital twin to
r approach through extensive evaluation in simulation and showcased its effectiveness for a cutting skill in the context of assist
n has been proposed to accelerate preprocessing by a factor of 2. Finally, we experimentally demonstrate that our method can
mpared to an only keypoint data processed-based model, our multimodal approach made a 17.7% accuracy improvement. The e
tions. We demonstrate the real-time feasibility of our approach with human-in-the-loop simulations. Project page can be found a
in both predicting the human motion and achieving the robot's goal.
per frame. Additionally, the strategy lowers the CCP and ACP in a baseline model by 11.25% and 13.50%, respectively.

ves efficiency by at least 25% on adversarial training with the same performance. Finally, we evaluate our method on a physica
matic setpoint computation that preserves human body scale and can be used as input to an image-based visual servo control s

obot shapes. Robot experiments confirmed that the proposed method reduced the peak gain of the frequency response by appr
sponders can rely on a SAR cellular-enabled robotics framework when planning mission-critical operations to take informed dec
aph from traversable to not-traversable. This operation involved oriented sample space of a path segment in the global graph la

plications and limitations of the approach.

mbitions, it can imagine physico-realistic socio-physical scenes from texts, demonstrate understanding of these texts, and all th
ic simulator.
ge queries, making LLM-Grounder an effective approach for 3D vision-language tasks in robotics.
perceptual conditions and show that our method produces reliable estimations in various environments. In addition, the experim
herical voxelizations for hybrid voxel feature learning. To address the radial imbalance of object instances, we propose a novel
ng the way for learning complex visual recognition tasks leveraging foundation models. To illustrate its general applicability, we

sentation via Dynamic Grouped Representation Generation and Dynamic Mask Generation. Extensive experiments validate tha
es in objects with up to 91% accuracy.

ance of 3D convolutional networks on three real indoor datasets, while working in real-time. Moreover, it shows better cross-sen
We validate our exploration strategy and demonstrate the utility of structural semantics in exploration in two complex indoor env
nes the strengths of region-based VLFM and TSDF, facilitating real-time 3D scene comprehension that includes object concep
ess 6-DOF bounding box parameters from spatio-temporal instance queries, which is used as an auxiliary task to foster spatial
y decodes the same instance over time, carries its ID and the tracking is performed implicitly. This enables us to jointly optimize
he experimental results in the Habitat simulator demonstrate that our HSP brings a significant boost over the baselines. Further
e proposed method. We establish the novel concept of textit{Belief Scene Graphs} (BSG), as a core component to integrate ex

tion. The unsupervised deep stereo serves as an auxiliary functionality, capable of generating accurate disparity maps that can
sults demonstrate improvement over the compared baseline methods, illustrating the robustness of our method in few-shot 3D
d provides model-agnostic modules to enhance scene interpretation. The benchmark is based on the RadarScenes dataset and

es obtained from onboard sensors and satellite maps through an alignment module. The experimental results on the augmented
ree RGB-T semantic segmentation benchmarks. Our source code is available at https://github.com/UkcheolShin/CRM_RGBTS
nd their assignment to certain map sections, other vehicles observed by the ego agents, and the pose graph from SLAM includ

d real-world environments, demonstrating its superiority over classical and state-of-the-art methods. The open-source code is a
MP achieves competitive performance compared with the SOTA approach training by LiDAR's data. In addition, we use the enh
entative algorithms on nuScenes dataset validate the effectiveness of our surrogate metric, demonstrating that sensor configur
ions, we propose a novel Symmetric RGB-Event Fusion (SREF) module to effectively fuse RGB-Event features without relying
ot required during inference anymore since the 2D knowledge branch provides 2D information according to the 3D LIDAR input
multaneously fusing the high-level semantic context obtained from optical (EO) and IR modalities to improve the feature genera

tate-of-the-art baseline methods with large margins on all intervals. Additionally, we apply the generated HD map to a downstre

ed for flexibility, making it easy to form any kind of integration algorithm. For illustration, RTK, PPP, and four Real-Time Kinema
he time image. In temporal reasoning stage, we construct the events into a point cloud along timestamp, and use RANSAC algo
improvement of 7.40% in AP value demonstrates its efficacy in comparison to other existing fusion-based methods. Our code
hallenging nuScenes dataset and our self-collected ZJU-4DRadarCam dataset, respectively. Our code and dataset will be relea
itionally, an algorithm that leverages spatial anticipation for dynamic object localization is presented, trained, and evaluated on

of events and frames as the initial masks and learns the consensus regions to guide the inter-modal feature fusion. The fused f

robust to mapping failures, which occurred in 37.5 % of the experiments without our method. We also propose the Tumbling-In
the robot platform. Experiments show that our method achieves state-of-the-art performance on both KITTI/Waymo datasets an
 -world experiments confirm these features, which are useful for many robots performing contact tasks.

affordability and productivity of MagicTac can be enhanced with a minimum manufacturing cost of £4.76 and a minimum manu
mputed by our robust-TDoA model, which begins with parameters acquired during the first training step. During this process, o

hile achieving integrity in unknown characters. To access our code and supplementary materials, please visit our project page:

ng this, our work extends the applicability of legged robots to a broader range of scenarios.

hod in speed (maximum 40fps along with observer), precision (

vidually. Ultimately, ARIS 1.0 serves telepresence and novel regional language capabilities, specifically Sinhala-based self-com

t solves various perception tasks in the context of mobile manipulation actions in a household domain. Code and additional mat
t deployments. The statistical conclusions on experimental results validate that the proposed system can synthesize a morphol

diate state faithfulness, trajectory faithfulness, and adherence to global trajectory patterns.
of the simulator, by deploying the learned policy for dynamic navigation with up to six robots in numerous of real-world experim
ontact and a robust second order optimizer to ensure stable and penetration-free solutions to tight tolerances. It is a general pu
than those arising from simplified travelling-wavy undulations. Our study underscores the significance of integrating real fish ki
erage precision than the model trained on real-world data, demonstrating our simulator's photorealism. The code is fully open-s

y tuned to replicate complex system behaviors, such as sliding effects.

dataset demonstrate that KDD-LOAM significantly surpasses LOAM and shows competitive performance in odometry.
a university campus that provide data to support research into a variety of autonomous driving and parking tasks. Long trajecto

e datasets presented: https://github.com/ntnu-arl/lidar_degeneracy_datasets.

n for typical robotic applications. Finally, we benchmark the impact of the critical design parameters - the number of cameras an

age: https://hero-slam.github.io.
on and reconstruction.
ample of how our backend works by concatenating it to a monocular VIO frontend. In simulation and real-world dataset, our spa
e perform extensive experiments to evaluate our approach against other state-of-the-art approaches and show the benefits of o
aintained local geometric information, we devise a robust and accurate hierarchical data association scheme where point-to-sur
gher similarity rates than the raw unfiltered map when benchmarked against a corresponding LiDAR map.
h it is simple and relatively easy to implement, IPC competes with or outperforms the other tested methods in handling outliers
eo matching. Taking into account the practical applicability of our method in real-world robotics applications, we also propose a
equently update the map. In the visual object odometry backend, the updated map is employed to further optimize the camera
s. Built upon the novel feature sharing framework, Lite-SVO is able to incorporate the fine-grained feature sharing representatio

for future research in this field.

datasets, including simulated and real data of indoor environments from varying construction sites, and on a real public dataset
ustrate to prospective practitioners and researchers how VPR is implemented and evaluated.
aging radar configuration to mitigate the challenge of discrepancy in radar data. To attain high-precision 3D state estimation, we

that our method outperforms state-of-the-art LiDAR odometry in accuracy and robustness.
e maps competitive with existing real-time implicit mapping methods that use groundtruth poses.
f our implementation is publicly available (https://github.com/YibinWu/LIO-EKF).

strate the high accuracy and efficiency of our method compared to other state-of-the-art methods (see our attached video). By
paper develops an efficient second-order BA solver. Besides estimating the lidar
DORF across various types of highly dynamic datasets.

hat sensors placed with OASIS outperform benchmarks in terms of mean squared error of visual SLAM estimates.
e the system on EuRoC-MAV and TUM-RGBD benchmarks. The experimental results show that our method consistently surpas
algorithm, we present the improved results of camera-object relative pose accuracy and localization and mapping accuracy in t
f the maximum degree of the measurement graph. It removes the redundant information concentrated on a few nodes, and imp

F) SLAM is used to solve the SLAM problem at the back-end. Experiments were conducted and demonstrated that the propose

omparable to that of other state-of-the-art SLAM methods.

d notable generalization capabilities.
mensions and outperforms the state-of-the-art (SOTA). For place recognition and trajectory estimation tasks we achieve equival
ns compatibility with traditional PGO solvers, its efficacy benefits from a custom-tailored optimization scheme. To this end, we in

ow that our method can plan multi-task views and reconstruct a scene with high quality.

C dataset, where it achieves 7.9·10-3 and 5.9·10-3 RMS relative error on two independent sequences, outperforming the ex
or significant error and unexpected state changes during, e.g., long-duration missions. We evaluate our approach with a multitud
the effectiveness of the method.
mprove the robustness of such navigation policy by retraining it in the sense of asymmetric self-play. We evaluate the retrained
urfaces that need to be cleaned. We additionally illustrate our approach with a robot on 2 carefully designed surfaces. We find a

decisions. Compared to previous state-of-the-art methods, our method demonstrates superior robustness in terms of safety, ef

ces after receiving 33 rounds of feedback from a user on average, which is only 2% of the number of states covered by the use
lities of reconfigurable electromagnetic systems.
s of the proposed planning approach. This method is versatile and can be adapted and deployed for various types of continuum
so be captured. This work broadens the frontiers of in-flight capturing design, and we envision broader promising applications.
s can increase more than 12-fold to resist the external load. By pulling the threads, the actuator could be restored to the initial s

a shelf life assessment on seven measured traits reported greatly improved shelf life of between 30 and 150%, across all traits
eliable structure and control strategy to enhance the multifunctionalit
move in smaller incremental steps or through narrow passages (80 mm wide). The maximum speed was attained with a boundin

sorting, which proves its robustness and versatility. The proposed WEE-based E-skin can be easily applied to other soft robots
to estimate the extent of task execution. We establish an experimental setup to simulate the process of unlocking. The results

w routes for the potential applications of soft robots in obstacle-filled scenarios, including search and rescue, exploration, and in
m in the harvesting sequence were accordingly adjusted and finalized for following in-lab experiments, in which the gripper was
tage to drive the motion of the robotic fish. The masses of our tethered and untethered fish are 1.9g and 5.1g respectively. The

eatability of a single actuator, and demonstrate the array achieving diverse antenna configurations. This work advances the pre
plore more soft-rigid tactile robotic hand designs with greater capabilities in the future.

sponge. This sponge-jamming gripper shows great potential for developing next-generation robotic grippers for the manipulatio

he robot's capacity to accomplish diversified manipulation tasks. The proposed liquid robot indicates flexibility and provides nov
n optimal actuation capability under design constraints, a multi-objective optimization function is formulated to balance the actu

eformation caused by external loads. Additionally, tracking experiments with various trajectories in 3D space verify the accurac
 e are developed, and their rotation is realized by the proposed SFTTA actuator.
unts, steering mechanisms, and other application- specific soft growing robots. This research offers an analysis and experimen
enables a steering method inspired by plants in which the circumference of the robot is inextensible, but the sides can stretch t
ess can increase up to 95% in a controllable way, and the output voltage change of the springs can be employed for position se
ment of larger scale continuum designs, demonstrating that flexibility and tip stability under loading can co-exist without compro
the proposed whisker system and model. The results show that our localization can achieve excellent performance, with an inf

good fit and high output force, can effectively assist training and assist grasping, and can perceive the contact force.
d and physical testing validates its high-speed and impact force delivering capabilities.
ed a stiffness increase of up to 93.8%. These results underscore the potential applicability of the soft sliding structure in ankle s

uld provide a new approach to the interaction of prostheses.

odeling approach can be integrated into soft manipulation platforms to enable features such as identifying shape and material p
with different body-schema.
he EERLSMC can achieve precise tracking control; verifying the generality of our framework.
, a pneumatically actuated soft robot developed for minimally invasive surgery. Our vision-based force estimation model demon
of freedom (i.e., bending on two orthogonal planes) and is controlled by two tendons. When connected in series, multiple actua
oldering fires. We envision using similar robots to find hot spots in hard- to-access environments, allowing us to put out potentia
of external forces including the gravity and the interaction force at the end-effector is also taken int
the valve at an ambient temperature of 50°C. We also evaluated the pressure generation capacity of the reactor, which can r
f the sensing solution, without sacrificing the richness of the sensing information in the field of soft fluidic sensors.
ral concepts as well as first experiments on an industrial robot testbed.
scope of this paper, we embed our tooling suite into the overall context of streamlined robotics deployment and compare it to al

ared in six warehouse settings of increasing complexity: E. Coli, dung beetle, and a random walker. Our results demonstrate GS
cooking recipes. We demonstrate that our system significantly decreases LLM API calls (-51%), latency (-59%), and cost (-42%
eal-world project.

se areas that would otherwise be avoided by the Nav2. Our approach has been implemented and validated on a Clearpath Hus
a products to maximize their value for science and exploration of the Moon and other celestial
ents in space exploration.
ontroller that outputs the MU's thrust vector and is informed by noisy sensor feedback (for realism) of its relative location. Syste

h recordings of real-world robot grasping, is available at Fanuc_Robotiq_Grasp.

t and simulated event data using the same camera intrinsics. Furthermore, we introduce an image-based event representation
oasting improvements in run time, memory usage, precision, and resolvable obstacle size compared to existing grid-based map
ts are conducted.

r and data transfer.The obtained results will be helpful in the development of a multi-functional interface suitable for lunar applic
approach angle misalignment of up to 33.0 degrees.
to carry workloads such as manipulator arms, cameras, IMU etc. The VWDER robot can climb steps or doorsill using its two ad
(iii) Required Torques, and (iv) Computational Effort.

of the method to occlusions and various materials. Compared to state-of-the-art linear and data-driven methods, our approach

oject information is available at href{https://medcvr.utm.utoronto.ca/ral2023-rollblock.html}{https://medcvr.utm.utoronto.ca/ral202

cal instruments. Among the improvements, we developed an automated data generation pipeline and an improved surgical sce
learning where both the physics simulation and reinforcement learning occur directly on the GPU. We demonstrate between 10
d with data from diverse tasks. We have conducted extensive experiments on 10 tasks using the surgical robot learning simula
the extensible bronchoscope in an isotropic manner. Phantom experiments confirm the feasibility and effectiveness of the exten

from Hamlyn, SCARED, and Kidney boundary datasets. The experimental results show that Ada-Tracker achieves superior acc
es. The effectiveness of the proposed mechanism is evaluated in the manipulation of both a thin-shell and volumetric tissue, rep
ories than the comparison methods, and the resulting occupancy map achieves high accuracy.

he training of models that can generalize better to unseen tasks. Also, higher MP recognition accuracy can be achieved by trai

or future vascular intervention robot applications.

8.87 mm or 26.61 +/- 17.00 % robot length. We showcase that the proposed domain randomization approach reduces errors b
model in design optimization, planning, and control of CPPRs.
± 0.20 mm depending on the chosen spatial resolution, achieving the submillimeter accuracy required for interventional proce
stems, where fluoroscopic imaging is only required during catheter steering at bifurcations in the vasculatures. The proposed m
O can also estimate loads on tip with an average accuracy of 3.3%. The presented system can lead to better
tal results show that the mean puncture error is 2.48 mm, which can meet the requirements in a wide of clinical scenarios.

e solution to tackle such tasks in constrained GNSS-denied indoor environments successfully.

n and advancement in multi-swarm interaction studies.
terns via physical interactions and optical communications.

d in macro- and micro- length scale robot swarms.

oordination between the two groups.
ide a case study of decentralized model predictive control (MPC) that is run on a ground station to command a fleet of sixteen
 in real world experiments.

opological graph search. We validate the proposed framework through simulations and demonstrate the system with real-world
mes a special case. Experiments confirm the robustness and efficiency of the method on problems including stochastic bandits
ment pauses and leveraging all the data gathered during the robot motion. Thirdly, we further expand the capabilities of the con
g time, path length and number of steps. We validate our framework on the physical Toyota HSR robot and report comparisons

ethods that neglect the historical data in terms of action planning capability, enabling the generation of well-informed action plan

showing better performance.

ovide a set of injuries that may also be expected for human subjects under certain robot mass-velocity constellations in collision

ined using other state-of-the-art methods.

e these problems, we propose an adaptive RRT tree growth strategy for frontier point detection, which can adjust the step size a
ing the teleoperation performance and the user experience with and without haptic feedback. The conducted experiments cons
he proposed approach to enhance the effectiveness of robot learning from remote demonstrations.
ht of the arms (1.0 kg LDA, 2.5 kg FDA) and the human-size and human-like kinematics facilitate their use in a wide variety of a
witch between autonomous and teleoperation modes. A real-world data set is collected from our teleoperation testbed to train b
n DTMR and a quantitative assessment of the newly developed SPTA. The results demonstrate an enhancement in both the hu
p to 44 percent while using less than half the number of model updates when compared to JND-based MMT. Another experime

limited workspace at the expense of the amount of haptic information from simple vibrotactile feedback. User study to compare

ess of the control modality or robot used, demonstrating the usability and user-friendliness of TELESIM.

references. Code, video, and details are available at: https://joonhyung-lee.github.io/spots/

lot adjusts the robot pose to apply larger contact forces and move a heavy box at a constant velocity of 0.2 m/s.
ct experiments. Autocomplete outperformed traditional method on all criteria with at least 30% improvement in all objective mea
ough extensive experiments. Experimental results demonstrate that DA has competitive performances against full fine-tuning a
odal-specific prior is then shared across modalities to achieve better rare object segmentation. Extensive experiments show tha
oach has been tested through experiments conducted on three locomotion tasks and two robotic manipulation tasks. The empi

BEV feature spaces. It then transfers the sufficient domain knowledge of multiple spaces to the student model. In order to joint

ched, where each level of the network is learned by performing RL for a given simulation resolution. The proposed framework is
ks and peg-in-hole tasks conducted on PR2 and UR5 robotic platforms.
ncers can practice dancing in the VR system and then perform the dance in real-life settings. Different types of mobile robots ar
GRF. Furthermore, the reduced GRF decreased the ZMP, improving the stability of the humanoid robot's walk.

ning tasks while adsorption-based moving simultaneously.

ulator. Our experiments confirmed that manipulability affects the success of robot demonstration even with accurate object perc
advanced sensors and technologies will further improve its performance. We believe that this approach not only proves the sys
Infrared sensors, etc. It has a clear privacy-preserving benefit over cameras for scenarios such as health care and rehabilitatio
turning point in environmental conservation, leading to a significant reduction in pollution and the restoration of ecosystems. Ou

h a well-designed robust monocular image-based multi-person mesh detector. Locally, it preserves body aspect ratios and refin
ned efficiency and frictionless processes. This work serves as a proof of concept for integrating three Neural Networks in a hum

can be integrated to provide long-term localization on a daily basis. To demonstrate the robustness, we evaluate the proposed

with an RMSE of 0.23m and velocity and pose within the error of 0.022m/s and 16.59 degrees, respectively.

n, spoken instructions, and familiar music. Overall, the MUSE system seeks to bring the benefits of music intervention to those w
Key technological advancements encompass real-time computational model and sensors, echo image processing, augmented

c human features and generated robot trajectories through LfD and RL with DMP. This experiment resulted in a stable robot pe
e optimization technique reminiscent of flywheel steering in satellites. Unlike high-speed rotating wheels, it relies on precise rob
m (MRS) for distributed mapping in In-Situ Resource Utilization (ISRU) scenarios. The MRS system optimizes global efficiency
open hand movement from the ANN. This proposed method may lead to the further development of advanced prosthetic hands
t continues its task automatically
s. EEWOC has demonstrated fast and reliable free climbing on large steel structures, as well as dynamic swinging over gaps w
egiving companies such as Caritas facilitate the integration of these technologies into the community. Ultimately, the robotic sy
otic gripper with resistive bending sensors (Flex Sensor) in a laboratory environment. We use the feedback of LED lights to dete
nsors enables real-time state estimation, facilitating autonomous operation. The robot's advanced control system employs a rec
an generate bending motions to assist in turning the human body over. This is a device designed to support body weight while
he buoyancy regulator has been applied to small deep-sea robots, helping them achieve buoyancy control in extremely high-pre
wo motors at the same time. It greatly reduces the burden of the motor compared to traditional ways. Finally, the proposed RTS
and surface electromyography was used to measure muscle activities of the biceps brachii and triceps brachii, and participants

rmal set. For that, we propose a novel convex formulation and a globally optimal solution method. Experiments on photo-realist
obot's joint angle and pose while also providing state uncertainty which could be used later on for robust control. We demonstra

in a table grape vineyard, particularly challenging due to strong self-similarity and density of the instances. Both the code and t
e available and discarding them when their ranging measurements are outliers. These ideas improve the efficiency and robustn
constraint. Extensive evaluations on the EuRoC dataset illustrate that our method excels in accuracy and robustness. It outpe

and XGX performs over two-fold better than the best baseline from simulation benchmarking.
This paper makes an important step towards long-term autonomous T&R navigation with ensured safety guarantees.

erogeneous devices is applied to achieve real-time performance. This visual perception pipeline is used in closed-loop control to
n goal attainment speed and managing model uncertainty influenced by velocity. In the conducted studies, our approach excels
pirical results demonstrate that CMN achieves high navigation success rates in unseen environments, significantly outperformin
types of 2.5, 5, and 10 cm square acrylic 2D markers, each equipped with a single LED blinking at 1 kHz, enabling recognition
using a Jetson Orin NX embedded system on the MPI-Sintel and KITTI public benchmarks show that RAPIDFlow outperforms

ational efficiency.
AI2-THOR simulator and conduct a series of experiments to demonstrate the effectiveness and efficiency of our navigator. Code
arious agents in object goal navigation.
M-based planner to utilize the commonsense knowledge between objects and rooms to select the best waypoint. Evaluations ac

an RGB colorspace that can effectively encode the multiple levels of traversability. We show that the planner can adapt to diver
approach is more efficient in gathering information for OOIs while avoiding obstacles. The results show that UIVNav chooses to
ce and performs a nearest-neighbor search in this space to estimate operator preferences over novel terrains. Through physica
ructure. Extensive experiments performed on simulated and real-world datasets demonstrate the efficacy of our proposed meth
ustness to domain shift and viewpoint change.
reducing the dependence on costly 3D occupancy annotations. Extensive experiments demonstrate that RenderOcc achieves
ised methods on the KITTI and ScanNet datasets. Our source code will be made public.

associated with model selection and tuning. All the code used in this paper can be found in https://github.com/longkukuhi/Robo

TTI dataset, KITTI-360 dataset and the nuScenes dataset demonstrate the potential of FSNet. More visualizations are presente
h quantified metrics at all stages of the pipeline establish our method unquestionably as the current state-of-the-art (SOTA). The
ally-realistic cluttered bin-picking dataset SynPick and the YCB-Video dataset and demonstrate improved pose estimation accu

sons and ablation experiments on these datasets demonstrate the significant outperformance of our method over existing state
nsive strategy, our method successfully overcomes the aforementioned limitations and outperforms existing self-supervised 6D
der extreme conditions. The experimental results confirm the superior performance of our method over both existing frame-base
anced controllable video generation models that can handle complex and dynamic scenes.
he targeted objects in the fused images are as salient as in thermal images. Our method effectively preserves the thermal radia
e accuracy and efficiency well and dramatically reduces the labeling effort for accurate cuboid annotation. Extensive experimen

nto aligned multi-view semantic segmentation maps by learning the shape correspondence between the BEV and perspective v

at https://github.com/Zerory1/Ev3D.
performance of more expensive LiDAR systems. The integration of single-beam LiDAR with our system allows for dynamic adj

ccuracy and benefits of our method in comparison with the state-of-the-art.

eal-world tasks, confirming its consistent advantages over the standard diffusion-based policy and other baselines.
ce code are available at https://sites.google.com/view/bi-kvil.

aset generation of arbitrary 3D scenes, and 3D reconstruction. Through experiments which pair a BoV-based reinforcement lea
lts illustrate that CopperTag exhibits superior robustness and accuracy in detection when compared to other state-of-the-art fid
oth real-world and simulated datasets. The results show that, the ~emph{FreeAlign} empowered robust collaborative perception

ded area. Experimental results on representative datasets prove that our method outperforms state-of-the-art methods, and imp
proposed model outperforms others on the occluded datasets. These comparative experiments demonstrate the effectiveness o
-art methods leads to a reduction in depth estimation errors by up to 4.2% and 10.8% on KITTI and CityScapes benchmarks, re
uced to improve feature matching. To further enhance dehazing and generalizability, the decoder's output undergoes a sequenc
ow that our semi-supervised framework can effectively improve prediction accuracy. To the best of our knowledge, this is the fir
red to manual annotation while providing accurate open-vocabulary auto-annotating results.
th only limited tunable parameters incremental. Capitalizing on the high dynamic range and temporal resolution afforded by eve

g a spike-oriented uncertainty scheme. Then, we propose a Self-Correcting Teacher-Student (SCTS) mechanism to screen ou
nt upsampling methods in quantitative and qualitative measures. We also confirm that when PillarGen is incorporated into birdâ
addition, the deep Q-learning algorithm is utilised to find the maximum FoV coverage and avoid collision. Three experiments ar
ion and regression network training, respectively. We evaluate our method on the AASCE challenge dataset and achieve supe
asets to investigate the influence of optical/imaging effects on the classification ability.
en as geometric prior that is used to iteratively solve for the optimal pose. Experiments indicate that the proposed approach for r
plinary applications, especially, for in vivo acoustic therapies.

Furthermore, we demonstrate the successful integration of our approach with real robot random bin picking, resulting in signific

rposed version of the SCARED dataset. In addition we perform explicit studies showing its robustness to large rotations. Our co
CL framework that leverages a multimodal LLM as the additional teacher. The strong generalization ability of the LLM can bridg
8.97% in recognizing surgical processes. This study paves the way for leveraging domain knowledge and image segmentation
 at this controller achieves convergence to a target, even with the uncertainties introduced by walking to a new location. We beli
nd. Furthermore, to prevent the object from dropping while the robotic arm is in motion, the distance between the centroid of th

e experiments in simulated and real robot scenarios. We show that our method provides better dynamic properties in both imag

able objects of various shapes and materials (paper, rubber, plastic, foam).
s. The unified end-to-end learning of enhancement and tracking in DCPT enables a more trainable system. The darkness clue p
ds to more versatile and adaptable operations.
MDA, an Adaptive Feature Synchronization (AFS) technique is presented to update the object features seen from different angle

h the box regression module of the visual detector, and prompt an open-world segmenter with the refined box to segment the o
the initial input sparse pillars and the output of each stage to all subsequent stages and the target localization network, which c
e pre-trained model to estimate motion in a video, and then select the subset of motion estimates which we can verify with cycl

rive MOT dataset show that the combination of our method and informative uncertainty estimates reduces the number of ID sw

Orin NX edge device. Moreover, our LiteTrack-B9 reaches competitive 72.2% AO on GOT-10k and 82.4% AUC on TrackingNe

en improve tracking accuracy.

e ratio of the inverse distance of keyframes, which enables more convergent map points aiding in robust tracking in regions lack
ameter degradation. The proposed SRI-GVINS is extensively evaluated on our own collected UAV datasets and the results dem
method outperforms SoTA methods in terms of depth prediction and 3D reconstruction. In addition, we develop a demo that ca

aggressive flight and large-scale outdoor experiments. Videos of the evaluations can be found on our website: https://youtu.be

Jetson Orin NX. Through outdoor experiments, it demonstrated adaptability in different sports scenarios, assisting users in achi
ced the surface myoelectricity of the lower limbs during STS transfer. The VGFL could demonstrate high support performance in

echatronic design, mathematical model and includes a pilot study with an able-bodied subject performing sit-to-stand tasks. The

g process continues. The robot tracks the individualized trajectory under a variable impedance model to realize the assistance.
future back exoskeleton designs.

is will promote the research that concerns utilizing HELIE for exoskeleton control parameters optimizations in the future. Finally
use of inherently soft materials. A spherical scissor linkage mechanism (SSLM) was utilized in conjunction with a customized gu
w the execution of trajectories for traversing steps, driving up ramps, jumping, standing up, and driving up entire stairways. To th

ed PTCM is demonstrated and validated by simulations and outdoor experiments.

d experimentally validated on a prototype robot. Smooth transformations of the multimodal motions are also validated by the pr
put is designed by the backstepping approach. Furthermore, the global convergence of the state trajectory to the reference one
ge, this is the first time that a quadruped robot can successfully recover from falls with horizontal velocities up to 3 m/s in simula
cting with heavy objects. Experimental results with Unitree Aliengo, equipped with a custom-made robotic arm, showcase its ab

: https://sites.google.com/view/lrdf/home

mance in familiar environments. Conveying the robot’s intention through haptics showed further improvements since the op
The results demonstrate the effectiveness of our method and show that we can achieve the new state-of-the-art performance on
ses. In contrast, the MOT had an average of 4 tracking errors for each video.
scale datasets to demonstrate that SpectralGV outperforms other state-of-the-art re-ranking methods and show that it consiste
skeleton zero-shot method by 27.0% and 19.7% on NTU-60 and NTU-120, respectively. Additionally, in the context of RGB vid
ound at https://github.com/Yavinr/LPS-Net.

ell-suited to achieve a notable improvement over two state-of-the-art novel object 6D pose estimation methods, without the need
alyzing the flow and clustering to distinguish scenes. The experimental results reveal that the proposed model outperforms stat

locity at low final times and 2) in terms of the movement strategy itself: The clutch mechanism allows the BSA to generate cons

g task that requires force amplification and as such is not realizable using conventional springs.
mooth, and feasible trajectories for the marsupial system.
imization model that accounts for communication constraints, obstacle and collision avoidance, and the performance constrain

GWT GAP9 System-on-Chip. Addressing the domain shift problem via on-device learning aboard nano-drones not only marks a

enerate complex images and strokes using various configurations.

h learns to capture the evolving structure of latent disturbances over time (in addition to standard meta-learning on the managea
the integration of PPM and the probability searching module yields an efficient hybrid vision system capable of achieving 120 f
navigation system, extensive experiments demonstrate that our proposed framework has achieved superior performance in term
% accuracy.
thermore, our method showcases enhanced robustness to network initialization compared to the gradient descent counterpart.
detection capability.
ation. Building upon this unified description, a kinematic-aware planner model is proposed to generate precise 3D manipulation
al results show that ISR-LLM is able to achieve markedly higher success rates in sequential task planning compared to state-of-
ying time-invariant latent features.

oal orientation without additional guiding direction approaches or policies. To the best of our knowledge, this is the first framewo

an approach provably returns an identical-quality solution while dramatically accelerating the online decision making. Our univer

ed simulation environment, and finally, deploying the policies on the real robot.
demonstrate that SG-Bot outperforms competitors by a large margin.
hybrid frameworks in 3D simulations where the vision-to-text problem and dynamical errors are considered. Please see project

ern, we spotlight the automated Foosball table as an invaluable tool, possessing numerous desirable attributes, to serve as a d
es intricate tasks that demand intent recognition and reasoning.
cal observations: (i)including both instructions and examples improves performance, (ii)interleaving state predictions in the exa

rform on par or even better than the current best-performing strategies in multiple situations regarding navigation robustness an

reinforcement learning with non-temporally decomposed reward functions fails. Gen2Sim provides a viable path for scaling up r

ovements in task completion. We show that our approach outperforms several methods using LLMs as planners in complex tas
SC is robust to hand-off errors, changes in the environment layout, dynamic obstacles (e.g., people), and unexpected disturban
s and robot state data. Experimental results show that our forgetting techniques significantly reduce the space requirements of

are available at https:// collaborating.tuhh.de/ilt/airfisheye-dataset

mulation and physical experiments prove that our planner can efficiently and safely explore dynamic environments. Our softwa

s representing the bistability of the underlying dynamics, i.e., detecting when the controller results in a desired versus an undes
osed approach in both parameter estimation and trajectory tracking compared to start-of-the-art approaches.

ns. Our method demonstrates real-time performance robustness to different scenarios, including unseen trajectories and high ro

cted experiments in VirtualHome environment and on a real robot. The experimental results show that our system strictly adhe
of performance and robustness.

k in a simulated multi-robot
ers of agents and tasks. The numerical results reveal the superior convergence and optimality of DMCHBA over all other selec

mation gain and, therefore, can quickly improve their knowledge about their surroundings.

d ablation experiments on the Waymo Open Motion Dataset (WOMD) demonstrate HHGNN's effectiveness within the motion pr
stly determines motion and position states even in radar-challenging scenes, such as environments with few radar targets or w
strate that nighttime navigation facilitated by Long-Wave Infrared (LWIR) thermal cameras can rival the performance of daytime

ble setting on the Waymo Open Motion Dataset.

localization tasks.
ether using a novel training regime. DiPA achieves new state-of-the-art performance on the INTERACTION and NGSIM datase
e experimental results demonstrate superior performance to other strong baselines in various driving environments.
ne-changing cost function from demonstration and performing scenario-adaptive lane-changing trajectory planning.

brid paradigm through a proof-of-concept experiment, in which we develop a hybrid planner that switches between geometric a

d-loop CARLA urban driving simulator and provide extensive ablation studies. The experimental results demonstrate that our ap
rios. Experiment results show that SCALE outperforms the previous state-of-the-art methods for open-world navigation with a u

to any number that we want. Finally, our IPF is applied to large-scale synthetic dataset and real dataset, and its validity is demo
 Car Innovation Challenge 2022, is solidly compared in two challenging autonomous driving scenarios against state-of-the-art (S

f 74 fps. Field robotics experiments were also conducted and showed that MultiTCNN can generate MTraMap of the surroundin
ves comparable performance with significantly reduced system costs, highlighting its potential for large-scale deployment.
active gas source searching in unknown environments. Simulation results underscore the adaptability of our method to diverse

pen Dataset (WOD), WOMD-LiDAR dataset contains 100x more scenes. Furthermore, we integrate the LiDAR data into the mo

en situations while being more sample efficient. See https://robopen.github.io/ for video results and appendix.

how that our method can incorporate data from an existing robot dataset collected at a different institution and improve on a tar
ffective method for improving accuracy of predictive models for robotic systems with limited compute budgets.
uture research to better detect, categorize, and address meaningful physical unit mismatches.
ots using a designable method, and demonstrates the superior performance potential of flapping-wing robots compared to conv

on and the roll rate of the ornithopter is identified. Even though the desired radius is not reached, the FWUAV is capable of main

bot mobility beyond simple substrates towards a wider range of complex, heterogeneous terrains.
ching') by flow visualization through PIV (particle image velocimetry). Our results demonstrate the effectiveness of our design a

of transport, respectively. As a result, TSCP increases the swimming speed by up to 35.5% and reduces the Strouhal number b

ariety of scenarios. Our results show that the CPG, explicit interoscillator couplings, and memory-enabled policy representation

e what would otherwise be a technically challenging problem.

and jump-turn over 2 rad.

ions in robotics.

f sample points in proximity to the nest results in a more consistent return. Code and data are available at https://mavlab.tudelf
ential to provide a highly biomimetic platform to validate the flight mechanism studying of Odonata as well as the relative on-bo
strations with payloads evidenced the microrobot can stably fly at an inherent 16 mg weight while carrying an additional 72 mg l
 tigate how information is processed and transmitted in mBNNs, and also to examine the influence of local and global chemical
spects of continuum robots.
verse narrow passageways, which exhibit superior mobility and environmental adaptability.
ently, as evidenced by a reduced cost of transport, and 2) navigate through terrain heterogeneities, all achieved in an open-loo
m based on a geometric metric for the underactuated hybrid tail-wheel robot. Within the proposed algorithm, the physical limitat
urations. The process of self-assembly necessitated the application of perturbation, which was realized by imparting magnetic t

sed algorithm. The comprehensive evaluation results demonstrate its superior performance in comparison to the baseline algor

orms existing works. We will share the code publicly to benefit the community (after review stages).
Ds inside opaque spheres instead of a printed or displayed pattern. Our method provides a faster, easier-to-use extrinsic calibra
em encountered during training, demonstrating the robustness of the model. The calibration process executes in 93ms on a sin
PINN and the non-linear grey-box state-space estimation algorithm when confronted with non-linearity and dynamic coupling. M
gathered observations.

hand-eye calibration. Applying GBEC to repeated calibrations, we obtain transformations with standard deviations of 0.37mm,
curate estimation capability. The proposed approach has been demonstrated on a 4 degree of freedom (DOF) robot arm.
n. Our method achieved a mean translation error of 0.8751cm and a mean rotation error of 0.0562â—¦ on the KITTI dataset, su

et and extract the target center from a very cluttered point cloud, as we observed some important characteristics of 4D radar. E

mics Spot mobile manipulation platform. We deploy VLFM on Spot and demonstrate its capability to efficiently navigate to targe
n a real-world mobile robot platform, show effective navigation in unseen environments in comparison with five alternative meth

lementations demonstrate the potential of learning to navigate in dynamic environments using raw sensor observations and sim
emonstrate that the proposed method not only improves interpretability and stability compared to existing learning-based metho
 challenges, we propose PathRL, a novel DRL method that trains the policy to generate the navigation path for the robot. Speci
a gradient-based optimization framework. Numerical examples show how ZAPP can produce safer trajectories compared to bas

y length, and up to a 22.6% increase in the F-score compared to current navigation methods using other sensor modalities.
nd supplementary materials, including a video demonstrating experimental results.

y show its batch prediction and auto-differentiation capabilities for trajectory optimization in both simulated and real-world scena
sociated with raw LOS checks, making it a valuable tool for both researchers and practitioners in the field. In order to facilitate f

ironments. We evaluate our algorithm's navigation performance using a Spot robot in real-world outdoor environments. We obs

ompared to baselines that do not leverage physically grounded VLMs. We additionally illustrate the benefits of our physically gr
bility of our approach in a real-world experiment.
mited view and movement budgets. We further demonstrate the applicability of our approach in a real-world object reconstructio
with any combination of sensor modalities as input, demonstrating its modularity.
ity in real-world scenarios has been duly validated.
MF-MOS has been released at https://github.com/SCNU-RISLAB/MF-MOS.

fficiency. Our results show over 85% success rates in grasping in cluttered scenes with only 40 seconds of training time.

odels publicly available at http://amodal-flow.cs.uni-freiburg.de.

T3D_H) and KITTI_H datasets, respectively.

sed CVFormer obviously outperforms existing methods for occupancy prediction.
ootprint and running 40% faster on the NVidia Jetson Xavier NX. Furthermore, the "TID" scheme is even more efficient offering
ection performance from 29.89% to 45.15% in terms of [email protected] with about 50% fewer queries, showcasing the effecti
s in the completion task on the KITTI-360 dataset. Our project page can be found at url{https://kazuto1011.github.io/r2dm}.
utperforms various baseline models. Moreover, an in-depth ablation study substantiates the efficacy of the methodologies integ
mentation tasks, and it achieves state-of-the-art performance on both tasks.

manipulator's approach angle is not aligned with the surface normals, which is ideal for unknown non-flat surfaces.

distinction to avoid the long-term deep learning process. The second experiment shows the shape reconstruction capability of

erential-drive robots, Husky and Boxer, under different friction conditions.

okes that span a larger range of Fn-v space provide better friction estimates. These results can be applied to grasp control to e
ed skills can be reused to accelerate learning in new tasks domains and transfer to a physical robot platform.
enarios, especially on dynamic domain shifts. We demonstrate the generality of proposals via extensive experiments including im

y metrics. Our method is validated in the continual reinforcement learning simulation platform and real scene with a UR5e robo

Empirical evaluation in a simulated tabletop environment shows that LRLL outperforms end-to-end and vanilla LLM approaches

ied by mid-air separation experiments. The proposed system is promising to be applied in scenarios containing both complex p

sults show that WiBot can overcome the limitations of existing window cleaning solutions. Finally, several promising research di

n with 7 UGVs and real-world experiments with 3 UGVs.

’s applicability.
ls on automatically labelled data.

hes from related areas of Human-Object Interaction Classification, Visual Relationship Classification, and Scene-Graph Gener-
oses for each experiment, a feature not found in the current state-of-the-art datasets. Our results demonstrate that RTSs provid

working dynamic environment, i.e. complex crowded dynamic scenes.

anipulation data which we open-source. We release our code and datasets at https://portal-cornell.github.io/interact/.
Find the nuPlan dataset and code at nuplan.org.
cher information when compared to prior datasets with human-verified labels, and supports new research opportunities.
rate a single video-conditioned model named RoboVQA-VideoCoCa trained on our dataset that is capable of performing a varie
to effectively create critical situations for the baselines.
scription for each robot sequence. We have invested significant efforts in calibrating all the sensors and ensuring a high-quality
mobile robot interacts with human bystanders. Second, we utilize this dataset to train policies that minimize counterfactual pert
eving a society with zero collisions. To facilitate this, we have made our dataset and benchmark toolkit publicly at https://hcis-la

driving behavior in complex traffic situations.

al traffic. Furthermore, we evaluate and analyze the performance using various baseline detectors and demonstrated its poten
based method. Our dataset will be made publicly available soon.
d models for offroad perception, and additional documentation can be found at https://goose-dataset.de/.
ntext-aware grasp on objects with loosely defined RoI (Region of Interest). Our dataset and the grasp detection model are avail

confidence. For this novel task, we construct several datasets and verify the effectiveness of the proposed methods through a s

acceptance (see supplementary material for sample).

porated into the child points to guide the feature extraction process, facilitating local structure modeling and multiscale informat
pire more HRI use cases involving egocentric vision. All code and data are open-sourced and can be found on the project webs
nsitivity to outputs with substantial distribution shifts, effectively mitigating the problem of error accumulation. In contrast, TRP a
blation studies confirm the effectiveness of the spatio-temporal sequences fusion mechanism.
alidated on the KIT Whole-Body Human Motion (WBHM) dataset.

od in real-world robotic applications. Our code and dataset are publicly available at https://3dapnet.github.io.

isual servoing scheme. We obtained successful control results with this fully model-free control strategy, utilizing natural robot f

single image. Finally, we evaluate our approach on the task of extracting the optical flow caused by schlieren from both a static
ordinate and from 18 to 55% on the vertical image coordinate, on a real-world dataset of 30k images. Finally, our in-field tests s
n two benchmark datasets across multiple labeled data ratios with a more compact network compared to the existing method.

ce the Normalized Camera Coordinate Space (NCCS), ensuring alignment of estimated 2D keypoints across varying camera int
mental results indicate that LiLaDet outperforms existing camera- and LiDAR-based approaches in the 3D lane detection task o

ple way to define new mobile manipulation tasks. We demonstrate the capabilities of our proposed approach in extensive simul
omated annotations.
ulation and real-world robotic experiments.
-art baselines in terms of map compliance.

ecognition.
ements compared to electrical synapses, and that NCPs lead to better results in both synaptic models.
 al features. And incorporate classification and segmentation losses from downstream tasks into the Rate-Distortion (RD) optim
://jstmn.github.io/cppflow-website
demonstrate great superiority of SafeDPA in both safety and task performance, over state-of-the-art baselines. Particularly, Saf

m the hitting movement in robotic Air Hockey.

ventional online learning methods which incorporate only naive exploration strategies. Video: https://youtu.be/fPKOrjC2QrU

of data ultimately needed for this approach to succeed.

riments with novel objects, OS-TOG recognizes 69.4% of detected objects correctly and predicts suitable task-oriented grasps

asks that require both geometrical and spatial reasoning. Videos, data, and code for our simulator and evaluation are available
s and show that it achieves promising results with strong generalization across different types of objects and multi-agent setting
with a single-view policy alone. Furthermore, the student policy demonstrates zero-shot transfer capability, where it can succe

periments show that our policy can dexterously pick up tangled-prone objects with success rates of 90%. We further demonstra

p baseline Contact-Graspnet in both success rate and inference time. It is demonstrated that with integrated multi-view input an

m a constrained table-top workspace. Our method transfers to a real robot and is able to successfully complete the object pickin
fine-tuning in the real world. The robustness of our approach was demonstrated in simulation, and experimentally validated in
GAMMA significantly outperforms SOTA articulation modeling and manipulation algorithms in unseen and cross-category artic
ross various object categories achieving an object clearing rate of around 90% in simulation and real-world experiments. We als
pulation tasks. Finally, we deploy the trained policy from our method to a real Franka robot for a pick-and-place task. We obser
n surpasses the performance upper bound set by perfect visible point cloud inputs. Project page: https://pku-epic.github.io/ASG

ime. To address this trade-off, we employ a reinforcement learning policy to synchronize the manipulation strategy with active p
g the adaptability and flexibility of robotic manipulation systems.
ss rate of 94% and a clutter completion rate of 100%.

ttps://youtu.be/bwPf8Imvook

the two students. Second, as the ViT student has more uniform representations for all layers, we propose consistency-aware cr
t the first benchmark dataset, LCSD, for low-light crack segmentation. LCSD consists of 102 well-illuminated crack images and
e propose a novel binarization mechanism called dynamic-static hybridization. Specifically, our approach combines static binari

observational input. Robot videos are best viewed on our project website: https://see-to-touch.github.io/.

e source code for public access.

ntiates the effectiveness of our approach across diverse settings. Our approach achieves a success rate exceeding 80% in both
oposed system can provide accurate in-hand manipulation of sub-millimeter precision and that the inverse Jacobian-based cont
stream tasks. Experimental results demonstrate that our pre-trained model significantly aids in learning manipulation skills. Com

th a Shadow Hand have been conducted for validations.

nalyze the results on 7 hand configurations and 15 scenarios. We demonstrate 8 tasks on two robot platforms.

ent paths for R-UAV in the presence of obstacles. The experiment video is available at https://youtu.be/mVm0qCvB7HY.
OS2-LS also enables robust and continuous target following and can recover from network failures.

retain the quality guarantee. The proposed methods are validated via large-scale simulations with 170 robots and hardware exp
ons in high-fidelity environments, demonstrating significant improvements in task success rate, coverage ratio, and task executi
results reveal that the average error of each direction is 0.0027 m, 0.0031 m, and 0.0368 m in the x-axis, y-axis, and z-axis, res

han 15% of the time of the existing methods to find a feasible solution.

omparing the effective mass properties and the simulation results.

g environmental changes, and plan beyond the speed at which observations from sensors are available. Experimental results an
s show that our method, trained in simulation on a generic square peg-and-hole task, can generalize to a variety of real-world in

on of 0.45 degrees. Consequently, our proposed method can be seamlessly integrated into robotic manipulation platforms to su

obabilistic map matching framework and the Particle Swarm Optimization strategy to associate the magnetic data of multiple se

oblem instances. We evaluate and compare the proposed techniques in simulation and show that our methods in most cases p

th impact-aware control to extend the classical collision event pipeline. As a proof of concept, we show the effectiveness of the

multimodal sensor fusion based classifier network. Proposed network accurately detects and classifies failures from raw senso

UAVs in real subterranean environments, demonstrating the efficiency and speed of our method.
pers force application was analysed and the UAVs powerless robust perch is demonstrated by total disconnect of power while p
eks to accelerate innovation and broaden adoption of these techniques.
operator control and state-of-the-art methods show advantages in terms of time and energy efficiency, stability, and smoothne

ate successful teleoperated object grasping and manipulation in real-world conditions, with collection times ranging from a few
e rates of 10 Hz on a RaspberryPi 4B.

resent static obstacles based on an RGB-D camera. After inspection, the Structure-from-Motion (SfM) pipeline is applied to gen

0%) is consumed by the electronic components of the robot control box. Using this knowledge, the approaches to further optim
ational responsibility on the hub, ensuring the UAV remains unaltered. This ensures broad compatibility with standard UAVs, o
alternating exploration-exploitation like behaviour along with bench-marking its performance in contrast to the traditional sampli
 ling, while preserving the quality of estimates. Through simulation experiments using realistic spatio-temporal datasets, we dem

e such predictive capabilities are essential.

99.72%, surpassing TacTip, which achieved 94.60%. It also exhibited superior performance in both pose and force estimation

96% reduction in power consumption on the neuromorphic platform. This system out-performed previous work by the authors a
compliant robot arm. These results suggest that simple analytical calculations based on dipole moments can sufficiently extrac
implicity of our training environment and effective real-world performance highlights the value of rich tactile information for fine

e that the proposed L3 F-TOUCH sensor compromises GelSight’s limitation in force sensing and is more practical compared

abricate, and calibrate camera-based tactile sensors to best fit the needs of their robotic systems.
t of up to 84% over the single-task training.

ns of vision-based tactile sensors, we open-source both the hardware and software of 9DTact, along with a comprehensive vide

wing for monitoring of the contact during manipulation.

nd its divergence to estimate the CASR. Our model-based approach can accurately discriminate the compliance levels of the s
robustness of this system by experimenting with different unseen objects with applied perturbations in the real world. These tas

e real-time monitoring of dynamic contact force during vegetable harvesting in vertical farming scenarios. By reflecting contact

ain, for instance, reinforcement learning policies of manipulation tasks. The proposed model is verified in extensive experiments
rom normal and point cloud registration on the multiple frames of images obtained by rolling the cylindrical sensor, resulting in l
We evaluated and compared the positional accuracy of the markers and image similarity metrics of the images generated via o

ive visual-tactile model predictive control over single sensing modality and baseline model learning techniques. The proposed m
cate that MViTac facilitates the development of improved modality encoders, yielding more robust representations as evidenced
ccurate contact angle estimation as well as excellent tracking performance and strong robustness in force control.
fficulty of machining is also proposed. In addition, the validity of the proposed method is demonstrated using experiments. Con
ess of RUSS under dynamic conditions. The results demonstrated that the system exhibits comparable manipulation capabilitie

amming approach for the direct translation between task-level constraints and robot control parameter design is shown to be a

us vehicles obey traffic rules more robustly than human drivers in the dataset.
te best practices when employing Riemannian geometry within robot learning applications.

The effectiveness of our method is demonstrated through simulated case studies where obtaining the optimal solution would ot
nments and transmits rich LiDAR sensory data over congested WiFi links.
ness. We show the scalability of our planning algorithms in simulations of robotic tasks.
g of the Transformer module resulting in an improved representation of LTL. The simulation results demonstrate the effectivene
utational limitations and improve the scalability by two orders of magnitude by a new way of constructing models for PRISM-Gam
easible planning that satisfies both global and local tasks. Experiments demonstrate the effectiveness of the proposed decision

osed new design is that one pair of fingers can be used to switch the stiffness of another pair if a dual arm robot is used.
to unseen object categories, as shown in both simulation and real-world experiments.

for different reference sources. Model trained with the proposed learning pipeline shows robust performance over a range of si
perimental results on the Visual Manipulation Relationship Dataset (VMRD) and the large-scale relational grasp dataset named

on the Franka Research 3 arm.

production error of 3.9% of the maximum finger joint angle and are able to generate successful grasps on a simulated human h

ating it against state-of-the-art methods on synthetic datasets, indicating superior performance for grasping and reconstruction.
uations preemptively. We train MoMo within an interactive imitation learning framework, where it continually updates the model
ss rate nearly 20% higher than methods using the global features of the entire image from pre-trained models. In addition, our
aging demonstrations in model learning while respecting real-world safety considerations -- exploration centering, agency hando
s. By integrating these advanced techniques, we present a promising solution for enhancing robotic grasping capabilities in rea

efficiency and reliability of robotic grasping in recycling applications.

pens up new possibilities for enhancing robotic applications that require precise grasping and manipulation skills of real-world r

same time.
ping irregularly shaped objects and is able to carry a maximum payload of 6.76 kg.

esent a promising solution for enhancing adaptability, safety, and performance in human-robot interaction applications.

dvantageous in the context of automated food serving

dimensional manipulation systems feasible.

rse kinematics models of human fingers. Furthermore, our proposed method is compared with a baseline method, in which the

ments demonstrated that point-wise vibration pattern production on our tactile board achieved an average level of about 0.9 in t

l interfaces capable of providing effective haptic feedback.

collected data outperform their randomly collected counterparts in real-time training tests. Additionally, we find that the resulting
ontour following in haptic exploration.

ationary single-degree-of-freedom knee exoskeleton test platform in a proof-of-concept study. Preliminary results indicate that c
s diverse fields.
ere able to successfully perform online control of multiple grasping motions of a real wearable robot hand, reporting for high reli
meet the individual preferences of most subjects involved in the experiments, improve the human perceived comfort, and foster
es lacking labels are augmented with valid grasp bounding box information, whereas images afflicted by catastrophic labeling e

ed by day 15.
how that our approach outperforms the other policies on all tested scenarios.
ut requiring additional human intervention. Lastly, to confirm the practical applicability of our approach, we conduct experiments
akes, significantly improving the odds of task completion. Incorporating part-time interaction yields an increase in success of 0.3

mental results show that our method achieves a performance gain of at least 4.02% in the net revenue and offers more robust p
reach 6.0 m/s for a novice drone pilot in a real racing scene. We will open source our code later.
wered exoskeleton case. Results show that the average compensation/release time (payload) among subjects is measured as 0
arantee for safe manipulation. Physics experiments can be viewed in the attached video.

ely 86.46% when predicting pedestrian trajectories. Thus, we show that FlowMNO can seamlessly integrate into robot navigatio
hile reducing the number of collisions by 82% and reducing the proportion of time spent in the pedestrians' personal space by u
questions to gauge interest and preferences of navigational Explainable AI (XAI) in robots. A large majority also showed intere

ation score. Through BSON, users can evaluate their socially-aware navigation methods and compare them to baselines.
ng agents and leverages deep-learned prediction models as prediction priors that the join trajectory optimization tries to stay clo
s demonstrate that the proposed GraphicTEB achieves spatial completeness, higher scene pass rate, and fastest computing e

ablation study demonstrates a significant improvement in trust recognition performance of the spatial representation. The sourc
t that our classifier obtains this performance with minimal data processing in the standard algorithm setting reinforces the powe

he proposed framework’s effectiveness and adaptability to diverse environments.

hannel dropout. The proposed framework maintained high performance under a sensor dropout reliability study conducted. Res

imental results demonstrate the importance of the ToM-based robot policy for human-robot trust and the effectiveness of our ro

ds, respectively, and there is no significant difference exist between the two speeds. In addition, we guide users in realistic runn

ccurately predict future positions in a 200 ms window with errors of 3.1 ± 4.8 mm averaged over the prediction window with inf
lable. Experimental results on path instructions generated in a real environment demonstrate that generating explicit maps ach
dance plans then current operational methods.
 te that sEMG encodes significant potential for real-time estimation of EoP in the design of next-generation pHRI controllers sup
Our results show that the proposed method exhibits better performance across different metrics and successfully generalizes to
up to 40%. Participant feedback through a questionnaire, in terms of perceived effort, walking naturalness, and stability, also fav

uman following and obstacle avoidance without intruding into the intimate zone of the target person.

oduct belief MDP is constructed between the probabilistically labelled belief MDP and the automaton translation of the scLDTL

me. Our system is implemented in both a lab and a real industrial warehouse setup. The results show a significant improvement
tasks and different LSTM-based classification networks to supplement real data.

RK to assess its usability and robustness. The results show that, compared to traditional manual teleoperation, our method can

mework improves the robustness against such changes.

ew task or object instances. DROC outperforms baseline Code as Policies by using only half of the total number of corrections n

adjusted online according to human preferences and ergonomics.

ses underscore users' perception of the method's efficacy in enhancing their performance.
uiring dexterous telemanipulation. This paper's contributions include the seamless integration of FIC into multi-arm systems, the
on to HRC plan generation, eliminating the need for a manual and laborious definition of restricted planning problems and agen

by a quadratic program (QP) based on different control sets. Experiments were conducted to verify the proposed framework a

performances on the representative VLN task R2R show that our method surpasses the leading zero-shot VLN model by a larg

e also use the total collected feedback to train a robot policy for socially compliant trajectories in a simulated social navigation e
canvas and one with human created work. More generally, our fine-tuning procedure successfully encodes the robot's constrain
g93.github.io/publications/MATERobot/MATE Robot.html.
in a real-world scenario.

ncurring with prior NASA studies. When the SuperLimbs provided assistance, the astronaut's energy consumption and deviatio

ures at the midpoint), asking participants to rate robot competence in between every video as well at the very end of the experim
presents predictable motion) and turbulent flow components (that reflect more unpredictable and arbitrary motion), LaCE-LHMP
n a laboratory environment over two months. Our results demonstrate the effectiveness of our architecture to allow a physical ro
addition, our portrait drawing inpainting method significantly outperforms previous advanced methods. The code and dataset ha
so carried out active (online) tests across all participants in the same virtual training environment, with a varying level of externa

3 standard models and a user-defined model as examples. The results show that our framework is able to draw visually pleasin

the robot showcases an emerging ability to transit naturally between distinct locomotion patterns as desired speed changes.
motion on non-flat terrain in simulation and on the real humanoid robot BHR7P.
d temporal summation effect in humans. The improved bio-plausibility and the capability to reproduce the human-like features m

an models. An energy-based method is proposed to predict and control the coupling deformation of the layered structure. A phy
m to 2.5 Nm, even in the presence of external contacts.
nt cloud-based 3D object detection.
3D object detection dataset.
datasets, significantly outperforming previous early-fusion and late-fusion methods with comparable transmission costs.

entation methods.
OPV2V-HPL demonstrate that HPL-ViT achieves SOTA (state-of-the-art) performance in all settings and exhibits outstanding g
D voxel features interpolated from the original image features. Experiments on the Occ3D-nuScenes benchmark demonstrate t

With these advantages, our method yields state-of-the-art performance on the camera-based SemanticKITTI Scene Completio
ark with 5 metrics, including the most common tracking accuracy metric called AMOTA, to demonstrate how ShaSTA may impa
sed 3D object detection, presenting it as an invaluable augmentation to existing multi-agent perception systems. The code is av
attacks substantially undermine the performance of state-of-the-art methods, showcasing remarkable transferability across diff
d-to-end methods for the first time. Second, our in-depth study uncovers a substantial performance gap when transitioning from

rspective, the proposed auxiliary constraints facilitate cooperative-view independence mining, resulting in enlarged perception r
that can better utilize the theoretical property of Kalman Filter-based tracking methods. The experiment results show that our a
regardless of the choice of the model architecture. Our results predict the trend that in the future we will see larger and larger m
ution (OOD) tasks.
policy, providing more information about the environment for the policy to make decisions. Empirical results on the Mujoco dat

ating when task progress is being made. This ranking function can be integrated into an adversarial imitation learning scheme r

at it exhibits generalization.
ation alone has also been measured. This speed-up allows us to successfully run DL robot-control applications even in limited c
n scenarios. Two real-world experiments also show that the computed normalized escape energy estimates appear to correlate

ut to show the proposed robotic hand can perform simultaneous adaptive grasping and in-hand manipulation.
Allegro robot hand in a real-world in-hand manipulation experiment. The results show that our methods have faster training and
bang-bang-like RL policy actions that are still safe for real robots, simplify reward engineering, and help sequence real-world, h

a smart city. Experimental results show our method outperforms both traditional scheduling algorithms and Reinforcement Lea
ntial of SCRNet as an effective solution for enhancing low-light images.

and the base in order to keep the arm in configurations with good dexterity. The autonomous controller, on the other hand, take

demonstrate how the behavior of our planner leads to qualitatively distinct trajectories, leading to improvements in both succes

ptimal solutions. Simulation results show that our method outperforms SOTA learning based method MPNet and advanced sam

several of the no longer available dynamic quantities of interest in real-time, from the limited set of measurements that are ava
ompare robot solutions with a human performer. In the best case, robots performed 9x faster than humans in specialized tasks
exible manipulator with respect to the ground truth camera-based motion capture system. Model transformation is also numeric
garment production. This innovative system is expected to play a critical role in realizing the vision of smart garment manufact
work may be of independent interest to a broader array of autonomous control challenges.

s to demonstrate the effectiveness and efficiency of the proposed approach.

s based on the input courtesy values. We examine SCBG on the WOMD and show that we are able to control the SCBG model
warnings only when necessary. We conducted human and simulated experiments in a virtual driving environment, and our resu
s with our safety-adapted loss function mitigates the misdetection of safety-critical pedestrians with robust performance for the g
POSS benchmarks, achieving a 2.8% and 4.0% improvement, respectively.

ubstantial reduction in communication costs by 23.10% compared to the non-judger scheme. Additionally, we achieve a significa

ensed drivers. Experiments show it is possible to discriminate the two behaviors by analyzing video clips with an accuracy of ∼

ystems, and traffic monitoring systems; and 3) images only require communication to equipment. In this work, we show robot ve
.21%, coupled with an impressive route completion rate of 91.34% in the offline evaluation conducted via CARLA. These perfor
erforms the conventional single-stage planning approach.
s. Our results show that the proposed approach outperforms state-of-the-art baselines and provides more accurate and scene-
as an extension to WOMD and the robustness benchmarks to aid the community in building more reliable and safe deep-learni

taneous assignment (IA) for re-balancing taxis across sectors and a sector-wide multiagent rollout algorithm that is executed in
al scenarios based on these failure probabilities, CLIC tailors individualized curricula for downstream training, aligning them wit

demonstrate that our framework is resilient to imperfect tracklet inputs. Furthermore, we show that by combining with our propo

p-down Leaf Link Aggregation (LLA) (obtaining the probability of future behaviors of vehicles for certain instances). We validate
sed POMDP. The simulation experiments demonstrate the superiority of the proposed solution to the existing warning generati

ine intention estimation and the real-time control are performed, and the estimation angle error and time delay are controlled at
eving a competitive score under the MEGC2019 standard, compared with recent complex recognition networks.
m/ykung83/Driver-Intent-Prediction.

uch as high ego-speed, compared to the past arts.

xations when the human operator is attentive and mitigates its influence when the operator loses focus. A real-time implementa

ed architecture can serve as a basis for complex musculoskeletal joint arrangements.

(4 - 12 Nm/rad) for plantarflexion not to hinder natural ankle motion in the swing phase.

he merits of each approach. Collectively, this preliminary study signifies advancements in kinematic optimization, with potentia

ckly traverse the reference paths and compare the required traversing time and the resulting path accuracy with other state-of-t

experiments. A visual representation of the performance index is also presented.

their invariance.
eal-time FEM framework SOFA. Experiments show that both our method and ABAQUS have approximately 10% error relative t

1% compared with the active-joint space method and Cartesian space method respectively. The average and peak jerks of the

DQI7CD8w.

nce while maintaining safety. Real-robot experiments have shown that the proposed algorithm indeed reduce conservativeness

tools can be transferred to the real environment.

nowledge, including probable sequences of human actions and activities. We demonstrate how these localized activity predictio
te (≒90%).

to predict and prevent action failures by up to 50%.

nd second order, and are used to control a 7DoF robot manipulator in its end effector space and joint space. More details regar
ector to avoid hitting complex obstacles. The proposed method is evaluated on three tasks and deployed to a real-world robot.

he task. To explore what influences mutation, we varied several control and environment conditions such as the maximum stiffn
bot arm. We also validate that the haptic representation encoder, pre-trained in simulation, captures the properties of real objec
resources in finding comparable solutions over the state-of-the-art methods in the seen and unseen challenging environments
o other classical and learning-based methods such as imitation learning, reinforcement learning, and model predictive control.
red to the state-of-the-art techniques. The code, dataset, and videos are hosted at https://rule-learning-rlap.github.io/.
sured to evaluate the proposed method.

the proposed framework can successfully align the generated robot behaviour with subjects across varying levels of proficienc

babilistic model. We empirically validate our framework both in simulation and on real robots, which include a fixed-base manipu

ge margin under a range of domain shifts, including variations in lighting, color, texture, and camera pose.

data collected by a single RGB camera. The reduction of the required accuracy of the original information resources can exten
enchmark tasks. For instance, the offline RL algorithms AWAC, TD3-BC, DT, and CQL all get zero or near-zero normalized eva
ery approach to acquire safe and efficient motor primitives in a hierarchical reinforcement learning fashion and leverage them th
A key advantage of our approach over similar works is that no prior knowledge about the material is required.

n with minimal variance and effectively guides the Teacher Ecoder to update and acquire a more efficient representation. RLTS

s efficient and effective sim-to-real transfer for vision-based model-based RL tasks. Experiments in simulated and real robotics

e snake robot tasks, including goal reaching, wall climbing, shape formation, tube crossing, and block pushing. COMPOSER ac
, we propose a novel domain-adaptation strategy to align feature embeddings originating from diverse underwater environment

d outperforms the state-of-the-art LiDAR-based geo-localization methods.

we investigate the transfer of trained models to new robots, enhancing their traversability estimation and extending the applica

ate the capability of MATRIX to serve as data augmentation for imitation-based motion planning.

high-dimensional observations in a self-supervised manner. A simple REachability discrimination Module (REM) is also designe

evasive objects, outperforming monolithic reinforcement learning and other techniques. We also demonstrate the efficacy of ou
solve some tasks with success rates of 70% or higher using under 30 minutes of data. All videos and codes are available at ht

ation studies further accentuate the pivotal role of two algorithmic components: the skill transition model and tactile ensemble po

quivariant architectures or do not use our contrastive pre-training procedure. We also show real robot experiments where the ro
asping task trained in simulation via TBRR transfers sim2real with 70% success rate subject to sensor impairment and physical
lengths or materials.

ction for the high-level policy, so that the subgoal selection process takes the performance of the low-level policy into considera

on. Moreover, we demonstrate that our approach can effectively generalize to different cities and sensor setups than the ones u
Extensive experiments demonstrate that our method can achieve accurate camera tracking and produce high-quality reconstr
navigate in an unknown environment without fine-tuning. The proposed PIRL is evaluated using a wireless digital twin (WDT) b

proposed adaptive reward function.

d samples in an on-policy setting. Based on this, we show that the corresponding approach achieves faster convergence and im
e search-based methods to this specific setting and compare our method with them. Experimental result shows that our approa
L algorithms. Experiments in both benchmark safe RL tasks and real-world driving tasks based on the Waymo Open Motion Da

per employs a visual encoder and distance-based heuristics to construct a graph containing main nodes and their correspondin
and is extensively tested on the quadruped robot ANYmal. We show the qualitative benefits of SPF in natural environments for
cut the failure rate up to 43.6% during high-speed turning and up to 31.8% for quadruped under external forces compared to us

eal transfer is achieved on the real-world SOLO8 robot, countering both electrical and mechanical failures.
while walking through it. The proposed CCRL framework leads to policies with consistent styles and lower joint torques, which w

chieving impressive results: a peak turn rate of 8.5 rad/s, a peak running speed of 3 m/s, and steering at a speed of 2.5 m/s. Re
erent physical properties. Parkour videos at https://extreme-parkour.github.io/.
s and code are available at https://sites.google.com/leggedrobotics.com/risk-aware-locomotion.

PPO comparisons while maintaining comparable energy consumption levels. Furthermore, EIPO exhibits superior task perform
to follow these waypoints by generating low-level motion commands. This approach allows the policy to explore its own locomo

ive and negative work throughout the stride.

nts conducted on a rat robot.Compared to both a non-spine based trot gait controller and a trot gait controller with lateral spinal

rhythmic output (Rhythm Generation), which is then mapped to a Pattern Formation (PF) layer. Across different robots, the onl

a collection period, and generalizes well to new speeds and paths.

tal trials on a large-size quadruped robot that involves carrying a 20 kg un-modeled payload (84% of body weight).
n that the distributed MPC has a performance similar to that of the centralized MPC, while the computation time is reduced sign
for robot-specific adjustments in the reward function and hyperparameters.

ies recover more effectively from perturbations, are more robust to large disturbances, and use less feedback control effort tha

method allows the balancing robot Ascento to climb 15cm stairs in the real world, a task that was previously impossible for this

obustness than current solutions to serious domain shifts away from the training data distribution, such as to indoor environmen

show its versatility and effectiveness. Compared with the state-of- the-art methods, MBFusion achieves 3.6% and 4.1% absolu
nsive experiments, the proposed residual-quantized VIO (rQVIO) achieves remarkably competing performance even when usin

n based formulation, NF-Atlas shows competitive performance in terms of accuracy, efficiency and memory usage on both simu

daptable to evolving environments through continuous update.

more than 60% on average. Our method is open-sourced on GitHub

e number of occupancy grids, than both OctoMap and SuperRay.

r generalization to new environments. Videos and additional material can be found here: https://sites.google.com/view/visual-tra
tions of existing methods, which require laborious human supervision and lack scalability. Extensive experiments in diverse con

entation of our method has been released as open source at https://github.com/shenyehui/TSCM.

nd an off-the-shelf image segmentation model by fusing outputs from both. We evaluate 3DIML on sequences from the Replica
 even though the map scale reaches more than 6 kilometers, while efficient and accurate localization is still guaranteed on NCL

detail levels across various areas of interest. To facilitate further research in this domain, we have released Colmap-PCD, an op
cally improved robustness against geometric degeneracy in environments where all compared baseline approaches fail.

ve performance in similar viewpoint cases.

tantially reducing accumulated errors. Our experimental results demonstrate that our global localization method significantly ou

85.3% Recall@1 on the MSLS_val dataset and 59.1% Recall@1 on the Nordland dataset when using ResNet18 as the backbo

within 3m in large indoor spaces. Project Site: https://rllab-snu.github.io/projects/WayIL/

mples that bring the most information gain and promote the performance with the highest efficiency.

e of different methods in different environments outperforming the state-of-the-art continual learning method. The implementatio
xS28/OptiState

ation allows for the active inclusion of additional learning constraints to minimize 3D alignment errors between the two 3D scen

ctor space. A transfer learning mechanism is also proposed to improve the accuracy and the generality of the aggregation netw
d that VPRTempo’s accuracy is comparable to prior SNNs and the popular NetVLAD place recognition algorithm, while bein

ate datasets from similar environments. Our detailed experiments and analysis lay a foundation for building VPR solutions that m
 rch areas such as channel estimation, image classification, etc. Our dataset is available at: https://github.com/ilaydayaman/LuV

guration do not come at the expense of performance.

world data from the Ardupilot community to demonstrate the performance of the proposed filter on known cases where existing
down to 1Hz.

ations. The multiple update PF exhibits faster convergence with fewer particles when compared to the conventional PF. Moreov
ed point to line ICP pose tracking perform very well, even when compared to localization and SLAM algorithms that use the curr
that utilized an adaptive extended Kalman filter.

s. These results are useful not least for applications of navigation with resource-constrained robot platforms in feature-sparse a

etypes, we not only affirm the consistency of our covariance estimates but underscore its superior computational efficiency.

igation method.
gorithm is demonstrated through real-time obstacle avoidance experiments executed using both simulation platform and a real-

performs the baseline in both environments (a 62% improvement for in-domain ropes and a 15% improvement for out-of-distrib

C algorithm by up to 27% with fewer samples and an end-to-end policy trained with MFRL by 19%. Moreover, because DMPO r
on and in the real world that this approach enables safer behaviors upon initialization and becomes less conservative (i.e. faste
ps surprisingly, we find that our implementation can achieve very efficient learning, acquiring policies for PCB board assembly,

eneralize broadly. On several manipulation tasks on a real WidowX robot and in simulated settings, our framework produces po
tegies such as differential wheel spinning and pivoting around a single sweeping wheel. We then used these insights to further
he environment using safe actions only. Finally, experiments of continuous control tasks demonstrate that our approach has be

uccess rate of 95.2% in simulation. Overall, we present a promising approach to address the reorientation challenge in manipu
https://force-sight.github.io/.
eryday objects.
manipulate a cable in a scene with obstacles using a dual-arm robot.
oud to propose fine-grained final actions. The proposed method is thoroughly evaluated both using large-scale simulated noisy
xtensive simulation study and show previously unachievable results on a real-world PR2 robot.

ial improvements over state-of-the-art methods. Our code is available at https://github.com/TRAILab/UncertainShapePose.

sed map representation can be used to i) generate navigation plans in the form of `hops over segments' and ii) search for targe

al estimation leads to normals that are more robust in areas with high curvature which leads to maps of higher quality. We also
nstruct various scenarios with accuracy and stability. The code will be released at https://github.com/liubonan123/OmniColor/.

n real-world environments. The source code is released to the community: https://github.com/YZH-bot/RH-Map.

getic conditions at a wave basin. To support continued research and development in underwater robotics, we make the design o

capabilities have excellent practical prospects for open seas and contribute significantly to developing agile robotic fishes.
ing, lateral swimming to the left or right, zero-radius turning, no zeroradius turning. The passing and collision experiments prese
mong roboticists and scientists to develop novel solutions and increase the sampling capabilities of end-users.

mance even in environments with obstacles that may move orders of magnitude faster than the AUV itself. Simulation experime
nd 7 FPS on a single Intel i9-9900K CPU core, making it suitable for direct deployment on embedded GPU systems. The implem
anipulator in a reaching task.
straints and sonar-based place recognition. Code as well as simulated and real-world datasets are made public on https://githu

derwater imagery, outperforming existing approaches in terms of color consistency.

n and mapping.
s make ReefGlider ideal for tasks such as long-term visual and acoustic monitoring of coral reefs. We present the overall desig
reases by 7.78% and 4.20%, respectively, these results clearly demonstrate the enhanced resilience of MPC-CBF to disturban

trated the performance of our proposed models in an example stereo-based visual odometry application. The low computationa

are also analyzed.

mparisons against a full belief propagation approach and a standard shortest-distance approach are conducted.

g water at 1.2 m/s, which satisfies the requirements of delicate repair operations.
nd truth motion, and the flooded basement of a WW-II submarine bunker. Among others, it is shown that there is an increased a

a crucial source of fresh water and providing valuable historical records. Experimental results from underwater caves demonstr

h motion-only and manipulation tasks.

recognition, which further escalates to 100% in post-contact analysis. To the best of our knowledge, this is the first work to inc
on for outdoor flight without human assistance.
. We further show effectiveness of the proposed transition mode in the switching controller by ablation studies where large over
espect to real-world settings. The experiments also show the simulator's benefit for the rapid prototyping and transitioning of ae

e the induced velocity and energy consumption during vertical descending. Then, the theoretic model is verified to be consisten

on to only 22% and 1.5%, respectively, compared to a drone disentangling maneuver. We also calculated the minimum idle tim

. Peak thrust enhancement ranging from 300 to 600% is found for certain configurations operated in close proximity (0.2 mm) to
kov-Dubins problem reduces to the standard CSC and CCC paths and the corresponding degenerate paths when μL and μR
 r to reliably perch on moving inclined surfaces (up to 1.07m/s and 90â—¦) with a success rate of 70% or higher. The efficacy of

lled axes. Additionally, these conditions serve as hardware design and control guidelines for effectively integrating the proposed

strategies to address the reality gap or exploring offline reinforcement learning with real flight data.

ntrol strategy has superior performance adapts to instantaneously drastic disturbances.

o a benchmark controller commonly used in aerial robotics, as the PFM used to quantify the cost of flight is reduced by 60.30 %

€™s power to save energy. Extensive real-world experiments demonstrate that the proposed method can adapt to various robo
erall, we provide a framework for collaborative transportation with aerial drone swarms, that uses only as many drones as nece

omotion phenomenon not yet described in previous studies. The results presented were supported with data from numerous ex
e results confirm excellent scalability and manufacturability of the proposed design, providing a versatile mechanism for fast pro

different actuation systems. Finally, performance results for PULSE115- 60 are presented, demonstrating a torque bandwidth

al symmetric arrangements.

d improves its performance.

cluttered unstructured environment. Experimental results are presented to show the tracking and navigation performance of the

ments show promising results towards real-time trocar pose estimation and tracking, achieving an average error of 3 degrees i
nd power also allowed the user to volitionally and continuously achieve activities beyond level walking, such as ascending ramp
trument-aligned B-scan. To dynamically model task-specific uncertainty, we introduce a weighting scheme to assign an importa
llows tube that contains silicone oil and a speed of 5.2 mm/s in an ex-vivo porcine intestinal segment. Impressively, the TSER b
which both existing and prospective users should be aware of before embarking onto dVRK-related research. The hope is that

monstrate that our proposed pipeline facilitates robust automated robotic acquisition, mitigating artifacts from breathing motion,

cartilage point clouds from different patients. The results demonstrate that the proposed graph-based registration is more effect
al long-axis view. The result showed that the median detection rate of the observation points was 63.3 ± 5.3%, which implies t
al-time segmentation results and automatically optimize the probe pose for enhanced Doppler images. This quality-aware mod

g (MRI), and accurately identifies finer drilling stages without relying on other sensors.
EE) of our ophthalmic robot of just 8 μm, resulting in a 64% increase in motion accuracy compared to our previous work where
h level of autonomy as it performs the skull milling task with minimal human involvement throughout the entire experiment. The
ngertip and wrist motions with a scaling factor of 3x, 0.5 mm diameter circles could be traced with a mean bidirectional precisio
proposed haptic feedback mechanism enhances the safety of drilling around critical structures compared to systems lacking ha

rent force models compared to the standard setup, respectively) between the instruments and the delicate blood vessel during
safety. The proposed method eliminates the need to switch control modes, enabling simultaneous visual tracking and tactile int
ments of the two pairs of wires are both 3.0 mm, the errors of the two models are 0.42 mm and 5.79 mm (4.2% and 57.9%), res
of the designed haptic interface. It can generate dynamic and accurate force feedback and realize leader-follower grasping forc
could be easily attached and detached from the drive units. A prototype was fabricated and experimentally validated. The result

CR is theoretically investigated. A design methodology for MRMs is then proposed and the fabrication process of RMCR is des

the precision upon existing techniques for FBG-based shape estimation but also incorporates a mechanism to quantify the mo

morphometrics, in particular the sharp turnings at the segment junctions.

ervised classification, both simulation and experimental data showed that the lesions are best clustered into two categories rep
ng with a convolutional ray-casting ultrasound simulator to produce synthetic ultrasound images of endovascular interventions.
 a force control algorithm allows the robot to dynamically minimize the tool-sclera contact force. Moreover, for the first time, we
r robot operation is reduced.

ment prevalent in real-world gastric applications. The research herein represents a significant stride in leveraging technology fo
tively. Finally, in the experiment of carotid artery catheterization, EFPF can control the catheter force within 393 mN, which is fa
ape-force sensing accuracy demonstrate the effectiveness and superiority of the proposed methods.
ow. The system successfully performs six consecutive supervised autonomous pseudotumor resections on porcine samples. Th

sk of radiation exposure for both patients and medical staff present in the operating room during the procedure.

n vivo applications. These findings offer a significant microrobot design paradigm by exploiting natural physical intelligence at th

was demonstrated as expected theoretically. This concept has the potential to be used as microrobots for future medical applica

based on the ``palpation" framework was derived. Consequently, we apply this strategy to automated tasks and achieve superio

s of the method. This deep learning network pipeline can be seamlessly integrated with the robotic micromanipulation tasks to

cale robot and the singularity of the mechanism is found and solved here. The experimental results prove that the refined S2wo
re insufficient for operations in deep-buried regions of large animals and human (e.g., implant therapy, biopsy, and tissue remov
ite their small size. Experimental results indicate that the CSG network surpassed the state-of-the-art models designed for sma

e elastic objects and cell membrane. QCR force sensor achieved quite small force measurement and feedback control at the c
on long-range AFM by adding a long-range coarse resolution stage to a preexisting AFM system all without needing to actuate
tem achieved a success rate of 93.5% and a time cost of 5.5 s for automated sperm immobilization.
ccurately reconstruct tail endpoints. Experimental results demonstrate that the proposed network outperformed the state-of-the
y-point scanning systems.
tact, and it has the potential to be universally applied in many cellular studies.
success rate of 76.6% when moving along the intended pathways, with a mean absolute position error of 1.6 ± 2.1 mm.

hen apply the formulation to successfully solve a task on a mobile robot.

s across embodied AI and real-world use cases.

t avoids a thrown ball while balancing a tall bottle.
AMPC approach on several tasks such as carrying, pushing, and door opening.
ce under various grasping metrics. Furthermore, we discover that the incorporation of temporal information from grasping poses

nd without active perception objectives. Finally, we demonstrate the transfer of our mobile grasping strategy to the real world, in

during the race. The MPCC (formulation without the NN) is deployed on a real autonomous formula student racing car showing

ely 10 cycles per link, which for the nine-link Purcell swimmer provides a factor of ten improvement in optimization speed over th

d with the composite model for DLON manipulation, which completes DLON installation tasks, both in simulation and with a phy
presented and corroborate the feasibility of the approach. The proposed method outperforms both the first principles physical m
in the attached video, we perform a separate set of experiments on a physical robot to demonstrate the robustness of the prop

sp width and depth of the object determined, dynamic movement primitives (DMPs) are employed to mimic human grasp action

etween the accuracy and complexity of our method. Our code is publicly available at https://github.com/ChrisChenYh/EfficientH
an power frequency, and errors to experimental data. Our algorithm provided consistent estimates of muscle forces and kinema

rates within complex, highly cluttered environments which existing algorithms cannot account for.

er the continuum's centerline and an average cross-section error of 2.11mm. The real-world experiments of tracking and estima
complexity and easier control implementation. Effectiveness and advantages of the proposed methods are validated through p
n optimization-based non-real-time controller. Our study contributes to the foundations for bringing soft robots into everyday life

quare trajectory and a circular trajectory, respectively. The autonomous steering framework is showcased within a 3D-printed m
r by 23.1% compared to the state-of-the-art end-to-end soft robot proprioception models and achieved 4.91 mm average Cham
vely refining models, which is applied to optimize the input waveform for the hydrogel crawler. This transfer to realistic environm

erturbation map) we assume the body velocity (perturbation) of the base to be a joint-defined linear mapping of joint velocity and

uct experiments on six tasks, where our method with the backup policy outperforms the classical methods (MPC, STL-solver), m

llows fast identification of homotopically distinct paths captured in the dense roadmap. We show, that compared to the existing

ed methods.

time of achieving a specific cost by 22.1% on average in simulations and 11.2% in the real-world robotic arm experiments. The
Euclidean space planners in terms of success rate and computational efficiency.

robot and demonstrate real world navigation around static obstacles and dynamic humans. Code is available at https://github.c

aring it to an Oracle value function obtained through an Oracle trajectory optimization algorithm. Our analytical findings demons

gorithm, it iteratively refines the solution's quality within the allocated time budget. This enables our approach to strike a balanc
11.2% compared to existing end-to-end offline RL and other outdoor navigation methods.
ufficient in capturing diversity across scenes, we use an ensemble of costs to guide the diffusion process, significantly improvin

solution. The algorithms were validated in simulation and results are provided.
o validate the superiority of the proposed method. Quantitative performance indexes show that APP outperforms the existing m

We deploy the SMUG planner on the quadruped robot ANYmal and show its capability to guide the robot in multi-goal missions
vironment. All code is available at https://github.com/KumarRobotics/kr_mp_design.

nd real-world experiments are conducted, which prove the flying ball player can hit arriving balls from different directions and dis

ach has been experimentally verified, supporting its feasibility in real-world monitoring tasks.

cal gaps from simulation to reality. Experimental emulation results, supported by critical analysis and discussion, lay the founda
trategy are verified through comparative experiments on the D-Arm.

dvantage of four-wheel steering. We also show that JetRacer-4WS can be used as a validation platform for testing advanced c
ss and advantages of the proposed method.

ng between different levels within the column tray structure.

nder step-and-scan trajectories commonly used in wafer scanners. Despite its sim- plicity, the herein proposed design procedur

are reliant on future states. Furthermore, a force feedback mechanism is proposed to help human understand the intent of auto

mulations, we demonstrate the high accuracy and practicality of this control method in various scenarios.

oopman bilinear model captures the dynamics with unknown parameters and has good prediction performance. Closed-loop tra
high data efficiency for learning object shaping by grinding and also provide generalization capability for initial and target shapes
0 Hz and successfully manipulates previously unknown objects with non-smooth surface geometries. Accompanying media can
conditions. Our code and dataset are both available online at this link: https://github.com/norlab-ulaval/DRIVE.
We show in simulations that we can control the system without the need for projectors using these coordinates, and we valida
gh success rates and finds paths in sub-seconds, which is many times faster than the state-of-the-art CMP methods.

re research directions. Overall, the paper shows that physical adversarial attacks on motion planning algorithms pose a serious
ergodic trajectory, we introduce the energy-aware (eware) filter, which iteratively validates the ergodic trajectory to ensure that
ight on effective approaches to address motion planning difficulties in intricate 3D puzzle-like scenarios.
m repeatedly deforms the pre-defined path for easier tracking, whilst generating a suitable velocity profile for robot execution. C
RT (CBF-INC-RRT) can increase the success rate by 14% while reducing the number of nodes explored by 30%, compared wi

eriments, showing that the proposed system can enhance autonomous driving of the UGV and help it effectively avoid high-risk

stinct locations on the robot with different fields of view. Multi-sensor fusion approach achieved a handover success rate above
l world demonstrating that LIM2N has superior user needs understanding, alongside an enhanced interactive experience.

lti-view images, enhancing interactions among vertices of dual hands within global and local contexts. We conduct extensive ex

unsafe interactions. We also employ a feature reduction technique that reduces the number of inputs to five to achieve a more

performance.
e at https://github.com/gicheonkang/prograsp.
ncorporating physical knowledge with deep learning.
eal-world human-robot interaction experiments.

ever, the user command interface obtained favourable results over the standard interface in terms of perspicuity and efficiency.

tialization, pose graph optimization, and map fusion. Extensive experiments demonstrate that our system can efficiently and sta
nefit the community, we will open-source our code at https://github.com/PengYu-team/Co-LRIO.

strate AutoMe

Experimental results have shown that our approach well addresses heterogeneous collaborative scheduling with dynamic subte
model predictive control (NMPC) for optimizing the robot trajectories under uncertainty and safety constraints. It is shown that
 ks together. Project website: https://project-roco.github.io
. Our approach scales to 32 robots with 80% obstacle density in simulation and 8 robots with 20% obstacle density in physical d
n the team. The experiments on both the Webots simulator and the Starcraft II platform indicate that our methods have better pe
llenging experimental setup, our algorithm reached a terminal state in 86% of all experiments, compared to 31% for state-of-the

adio wave like full communication. Additionally, it has been demonstrated that this method can improve performance in various
rms of final configuration achieved by the system.

proach in various settings, demonstrating its decentralized capacity to emulate the global expert algorithm. Further, we implem
oreover, real-world experiments are conducted to show the practicality and robustness.
bot executing a collaborative task while maximizing the amount of information shared with ground robots.

l of their unique shape.

n and overall mission times as metrics. Our results show that our system can improve the overall effectiveness of a team while

with either.

able upper bound. We show that our algorithms have great potential to improve science return from Earth science missions.

er robots in its MPC computations.

ion, we demonstrate that DecSTER is robust to unreliable inter-agent communication and outperforms information-greedy base

n be drawn. The performance of this new bigraph matching approach augmented with a GRL-derived incentive is found to be at
ersity R&D partnership.
plant layout), vehicles errors, dynamical obstacles, etc. In this paper we propose a flexible and robust methodology for multi-AG

lexity of EM-Patroller, and we also show that EM-Patroller has additional benefits of catering to miscellaneous user-defined join
 and finally (3) computes collision-free paths for multiple robots to perform said actions. Compared to an optimization approach
e great significance in promoting the motion planning of robot hand-arm systems and upper-limb prostheses.

ers allowing HIRO Hand successfully achieves grasping and in-hand manipulation ability.

erformance, we take advantage of the geometric symmetry of the arena to augment the training dataset through mathematical

of concept demonstrating data collection capabilities. A project website with more information is available at https://cloudgripper

aches at competitive 52.2%, 61.1%, 86.3%, respectively, while mean IoU of free space is 87.8%. Especially, the occupancy ma
end-to-end manner. We propose a baseline detector and evaluate the effectiveness of our proposal with +2.2%/+2.7% NDS/m
s other methods for object detection in adverse weather conditions.

e for real-time edge robotics.

sults show that the objectness scores play an important role in improving the detection performance. Based on these results, w

roposed modules into the state-of-the-art transformer-based 3D detectors, we improve the previous best results on both bench

ensive experiments conducted on the nuScenes dataset show that our FG-PFE not only requires less computational power but
ving mIoU by 4%.

g class distribution shift.

 achieves an increase of 1.3% in mAP compared to the baseline model.
nt performance gain.

rol in an unstructured manipulation environment.

datasets, respectively. Furthermore, we showcase its real-time processing capability by deploying our model on an edge-compu

menting cluttered scenes by achieving an average object segmentation accuracy rate of 80.7%, an increase of 28.2% when com
ased on geometric consistency for map alignment. Our approach accurately aligns frames with only 0.18 m and 2.7 degrees of

hese issues through the introduction of an uncertainty-aware IoU-based loss, an evidence-aware multi-task loss, and the implem
ose estimation in the wild.

manually annotate the ground truth, which contains abundant samples of cars, pedestrians, and motorcycles. Through testing a
our efficient implementation and the novel data set available as open-source.

images. Furthermore, the network learns diverse rotation-equivariant features and mitigates a significant amount of detection e

tention. Compared with existing methods on the publicly available dataset OPV2V, our approach exhibits superior performance

thod Pixelpick on the Cityscapes semantic segmentation benchmark. These improvements are achieved while maintaining an e

in our dataset and improve performance.

ce of the initial model by a large margin. Combining warm start and size-balanced sampling with established information measu
eme effectively achieves dynamic tracking of the Unmanned Aerial Vehicle (UAV) by the UGV under multi-disturbed conditions

e minimum-time objective, and (2) our implementation is comparable or faster than state-of-the-art retiming algorithms.
Stiffness Actuators (BSA).

the planar slider to the desired pose.

Deep Deterministic Policy Gradient (DDPG) and Proximal Policy Optimization (PPO) RL algorithms.

n through an unknown cluttered environment.

and that the solution of the closed-loop system is uniformly ultimately bounded despite the existence of dynamic uncertainties
le real-system experiments show increased insensitivity to torque disturbances.
of an online exploration task with a 7-axis Franka Emika robot.

when contrasted with the model-based baseline method.

olution for the ASPP. We also introduce additional modifications to the labeling algorithm to significantly improve the computatio
ergodic metric, and 15%, 22%, and 20% improvement in time to find all targets in the search region, for the three methods resp

traints). We evaluate each environment for different human densities to investigate its impact on each coordination method. We

mpared to SFC-based approaches. We also experimentally validated our approach using a Crazyflie quadrotor swarm of up to 1

cost than the existing state-of-the-art.

agents to selectively combine these different sources of information and reason about their inter-dependencies at different spa
n, to establish its practical feasibility, we conduct one indoor experiment with two real TurtleBot2 robots and one outdoor experi

overage time by an average of 42.42% and 39.16% over them, respectively.

Predictive Control (NMPC) scheme which drives an octorotor towards reference setpoint positions while minimizing the cable a
 experiment.
mean Average Precision (mAP) scores. Compared to the baseline Mask R-CNN, the proposed approach with Sim2Real adapta
so described an implementation of the overall algorithm on a UAV platform and validated it in real-world experiments.
LAM system (Fig. 1).

ithout a priori obstacle knowledge, the quadrotor follows a dynamic flight path while simultaneously calculating and tracking psi_
ping methodologies (both GMM- and not GMM-based). Evaluations are conducted using both simulated and real-world data. T

coverage. The source code will be made publicly available to benefit the community3. Project page: https: //hkust-aerial-robotic

ns and more aggressive exploration of smaller areas of unknown space. This results in full coverage of forest areas in multi-UA

is the first paper that performs the learning-based variable-tilted narrow gap traversal flight in the real world, without prior know
ction loop and demonstrate the full capability of safe control in combination with real-time VI-SLAM without any external infrastru

e identification. The proposed method is implemented in a small quadcopter, and the results show that our method can achieve
of magnitude. The source code will be released to benefit the community.

stead of 79.50%, significantly improve the precision of strength as well as spread estimates, and reduce the computational dem
y of our approach for long-duration flying object detection and monitor tasks.

rcumventing the constraints posed by occlusion, our method enhances the overall grasp planning process and empowers state

on the motors and the overall power consumption. The derived results, clearly demonstrate the MIGHTY and GEM-C schema'

real experiments that suggest Residual-NeRF improves depth perception of transparent objects. The results on synthetic data
ects and demonstrate a manipulation task using the reconstruction. The proposed framework exhibits superior performance ove

s, the effectiveness in our method has also been validated in real-world low-light imaging scenarios.

in which the robot had to open a door with unknown joint while estimating the articulation online. Our system achieved an 80%
utional Neural Network (CNN) to estimate the safety margin from the tactile images. Because it gives a distance to slip, the safe
is challenging dataset. Additionally, we show that set correspondences can be leveraged to distinguish positive and negative m
ctiveness of our system with a series of ablation studies. These findings highlight our work as a promising solution for enhancin
that outperform policies trained with human supervision alone, while requiring significantly less human intervention.

efficiency of our system outperforms existing single-view systems. The effectiveness of our system is further validated in a phys

e a smoothed cloth. Experimentally, we first assess our results in a simulation environment by comparing them with an existing
valid and accurate means of comparing state estimation systems, which is also easy to interpret and report.

ancy, motivated by real-world driving constraints; 3) Using the proposed benchmark, we conduct extensive experimentation on
e informative signals that can better indicate the system-level performance.

Sawyer. Our results demonstrated improvement in all participants' Iaido strike skill after training with Sawyer, and they respond
nce controller to achieve load-adjustable resistance training with the trainee. The biceps curl exercises were conducted and the

rt control methods, highlighting the excellence of the designed controller.

k. No adverse events, such as fatigue and dizziness, were reported during the study. Our results show a higher significant p
on of both joints implying selective joint strategy adaptations for different external dynamics conditions for the same interventio
the experiment by intuitive question design and reducing the experimental duration. Besides, the accuracy and reliability of PC

f unseen objects and human motions compared to the baselines.

ecognition. The proposed controller was implemented on a UR10e robot (Universal Robots) and evaluated through human studi
human understanding and functional assistance.

ence of base compliance can be minimized and system excitation can be avoided by sufficient human operator training and the
a tool to perform modular, inexpensive, and predictive investigations in physical human-robot interaction, without affecting the r
h PBO parameters as well as improved user performance and comfort. The great interpersonal variability in the optimized param
h extends the recent single-agent hierarchical exploration strategy to multiple agents cooperatively by planning for agents with m

pproach to determine which hypotheses to prune in order to ensure a predefined limit on the loss. We evaluate our approach in
the baseline POMCP without shielding) on large POMDPs, with negligible impact on the runtime for online planning.

d robotics tasks, such as next-best-view selection.

. This is helpful for leveraging online information even if the simulated transitions are far (in the state-and-action space) from th

ed with previous similar methods.

f a sampling-based optimization technique, our method operates in real-time, demonstrating successful and safe navigation in a

collision cost. https://sites.google.com/cs.washington.edu/dreams/

applications. Our results show that, by accounting for environmental effects within our perceptual maps, we improve the state e

mpared to selecting a static front end, and that using a POMDP policy provides even greater improvement.

by over a factor of 100, making the process computationally tractable. We demonstrate the effectiveness of the proposed frame
work in an experiment with a human hand as obstacle.

ons and physical flight experiments, we demonstrate that the proposed approach effectively reduces trajectory execution time
cond, we learn a general RL policy for any N-node EGs to produce efficient supporting behaviors. We present extensive exper

st. Comparative experiments for 𝑘-SNPP are conducted in challenging grid-based simulated environments to validate the pe

Our approach relies on an attention-based neural network, allowing agents to reason about the current state of the system to se

ts, what’s more, real experiments are provided.

the robust measurement of lower extremity bone positions from 1D raw ultrasound signals. It shows great potential to apply A-m
skeleton assistance as needed.

e controller best reproduces the joint torques and kinematics in a reference able-bodied dataset. Experiments with a powered kn

mains unaffected. The tested force fields induced varying degrees of changes in joint contributions and synchronization. This s
as evaluated through experiments on one amputee participant and six able-bodied participants. The results show that the neuro

ability to lift loads more than 20 times its weight, and variable joint stiffness.
iability of the method, reinforcing its potential for integration into practical applications.
xoskeleton's recovery assistance mode, that shows the consistency of the control framework with human intention. To the best
e metabolic cost of walking. The torque timing was remarkably similar across the four torque profiles while the torque magnitud

on notably minimized muscle strain during dynamic activities, exhibiting a consistent correlation with the mass of the object bein

r platforms on-board.
e representation size, specifically 1.33 kB on average.

ls help to reduce localization drift. Additionally, we release a data set that contains point cloud scans captured in an indoor corr

de available as open source.

her apply SVGD to MPPI with the spline interpolation. SVGD is based on the optimal transportation algorithm and has a propert

ange of medical scenarios.

ovel scenes, we introduce a novel keypoint-based admittance controller. We evaluate our approach in several real-world applic

on while designing exoskeletons, as well as providing a significant improvement to this specific robotic design.
o lower exoskeleton weight and maximize performance.
e results showed that the compensation by the electromagnetic force reduced the root mean square error of the torque to less
cises by back-and-forth movements along the trajectory according to the patient's reaction, and it has been successfully applie

is suggests that our hybrid adaptive controller may be able to adapt to the behaviour of the user to provide assistance as neede
user study for which people with PD (n=10) performed 6 exercises with the system. Participants perceived the SWS to be use
manipulator. Code is available at https://github.com/Sunlighted/RbSL.git.

ution scenarios.

nsiderable reduction in the number of safety violations along with higher, or competitive, policy returns, in comparison to six diffe
ive DRL that previously required paging between flash and RAM to be run entirely in RAM, improving the latency of DMC tasks

f 63.8% using a simple ResNet50 backbone, surpassing the performance of existing state-of-the-art methods.
d off-policy reinforcement learning as potential solutions. Finally,we validate our approach through experiments, demonstrating
he AI2THOR simulator and evaluated our algorithm on active perception, a prevalent task in embodied AI. The experimental res

performance than the state-of-the-art methods.

ments. Please refer to our project page: https://lmeee.github.io/DLOInsert/

onments show that our method successfully learns actions to follow LTL instructions even with uncertain event detectors.
of MOSAIC in two task families: object categorization and object-fetching tasks. Our results underscore the efficacy of MOSAIC
ethods and is, on average, two orders of magnitude computationally faster than the best-performing baselines.
ves larger blending weights to those advantaged MPIs with stronger local representation abilities while giving lower weights to th

utilizing neural fields in unconstrained environments.

patterns to infer the missing geometries via self-attention. We find that occluding a very high proportion of the input point cloud
om scratch.
d. For videos and supplementary materials, please visit our project webpage: https://hcis-lab.github.io/SKTHang/.

object-centric features that accurately encode both visual and positional information. Further, we highlight the accuracy of SlotG

eduction in the image search space without compromising the trajectory estimation accuracy. Our method reduces the executio

ation, and a 30% lower error than a discrete representation at the same resolution as our coarse global SDF grid. Videos and c

eving frame rates exceeding 35 FPS.

riment and implement UAV threat assessments. MMAUD's dataset collection process follows a methodical strategy, selecting i
d robotic applications. Additional DINO features can increase the performance with a higher memory requirement. The dataset

ng for course switching. In the experiment, the proposed solution is proven to be easily incorporated into various architectures
t our representation achieves over 100x reduction in the shared data size that meets the requirements of communication bandw
We evaluate the approach in simulated and real datasets ranging from table-top scenes to self-driving scenarios and demonstra

a simple robotics application which uses our method to sense the distance to and slope of a planar surface from a sensor moun
ich can efficiently model point cloud videos. Experiments demonstrate that our method achieves state-of-the-art performance fo
and results, we encourage readers to visit our project webpage.
 anNet200 benchmark show that our method achieves improved performance, surpassing methods that rely on re-sampling, re-

We then compare OriTrack to a flat reference panel, the solar energy solution commonly used on roofs today, and find that OriT

a robot arm and successfully grasped different objects, and the maximum grasping weight is 221.6 g in the grasping mode.
ns are identified and described with seven relatively simple geometric conditions. Our approach is general and can be applied to

other platforms.

ystem can perform over 16 manipulation skills that generalize to 40 objects, encompassing 100 real-world tasks for table-top m

by our approach. We demonstrate our proposed approach using two case studies: obstacle avoidance problem for an autonom

ent-based optimizers thanks to the end-to-end differentiability of our framework. Finally, we assess the framework with various
subject to currents and actuator failures.
ghts other anti-intuitive observations, demonstrations, and implications that, to the best of the authors' knowledge, are first-of-it
sed approach in identifying and handling system-level anomalies, outperforming methods such as prediction error-based detect

, triggering a safe task-abortion strategy whenever a risk of constraint violation is detected. We evaluated our approach on a sim

r method's efficacy on a dual-track robot in obstacle avoidance scenarios. Our framework guarantees robot safety while adheri
es. We evaluate the effectiveness of HEGN through experiments with the synthetic ModelNet40 dataset and the real-world Sca
form theoretical analysis of the loss function used in the evidential uncertainty framework and propose a principled objective th
demonstrating that when the state-of-the-art (SOTA) filters have numerical difficulties, our SR-VINS has superior numerical sta

UVD-discovered subgoals can be used to construct goal-based reward shaping that jump-starts temporally extended explorati
d on the bounding box and placement plane as constraints. Through comparison with different methods, we achieved a maximu
system, achieving an average F1-score of 98.4% and excelling at inspection efficiency by over 20x speedup. Kappa statistic te
f our registration method that integrates intra-day crop reconstruction with earlier-day scans in a seven degree-of-freedom align

ral office for plant varieties. Furthermore, we show that our neural network helps in filtering out leaves with lower segmentation

lap. The experiments presented in this paper suggest that our approach boosts the performance over all the investigated scena

effectively transfer the knowledge to the target fruit dataset.

on nodes using a cosine similarity based approach. We demonstrate the effectiveness of the proposed skeletonization techniq

esults confirm that our controller enables the aerial manipulator to maintain its stability after the flight mode transition and show
w significantly reduced collision and occlusion rates. Finally, our CVAE initialization allows our optimizer to operate with smaller b

veral farms and show that this system is able to accurately get to the underlying structure of these agricultural environments wh

rate that the developed harvesting robot can complete robust detection of fruit peduncles and non-destructive picking of sweet
obust to environmental changes.
ng up to ± 0.3 deviation from the GT value during prediction, which is sufficient to comply with agricultural standards.

andle occlusions and improve the 3D reconstruction and position estimation of nodes equally well as a sampling-based NBV pl

ve evaluations of our proposed INoD pretraining strategy for the tasks of object detection, semantic segmentation, and instance

an overall 87% success rate of the log loading operation, with primary failure cases due to log segmentation errors and deficien

poor hypotheses are quickly identified and rejected eventually having no adverse effect.
enables the robots to converge around the boundary of the animals and guide them toward the designated goal.

adversarial domain sampling technique, the domain parameters leading to low expected Q-values are actively sampled during
ting demonstrates that the robotic system significantly increases the data collection speed, and the correlation of findings from
alization to the novel but similar natural language instructions.

nificant performance improvements and augmented response speed (up to 90%) resulting from the seamless integration of our

ance in face of uncertainties.

enables the digital twin to preemptively identify potential safety hazards and formulate safe intervention actions to ensure the ro
kill in the context of assisted teleoperation.
rate that our method can navigate a 6-DoF robot earlier than a geometric-primitives-based distance checker in a dynamic and c
uracy improvement. The experiments highlight each pipeline step's significance, showing important results even when hands a
oject page can be found at https://uw-ctrl.github.io/phri/.

.50%, respectively.

our method on a physical robotic arm system with object detection, generating adversarial patches as our attack, and adopting
sed visual servo control scheme. We show that our setpoint computations tend to agree both quantitatively and qualitatively wi

equency response by approximately 3 dB, which increased the extra injection range of the nozzle by approximately 16 %.
tions to take informed decisions with limited resources, and, ii) illustrate the number of robots versus explored area and respon
ment in the global graph layer, while removing the respective edges from connected nodes of the global graph in cases of obstr

g of these texts, and all these with no data- and resource-intensive learning.

ts. In addition, the experimental results validate that our method outperforms other self-supervised traversability estimation met
nces, we propose a novel data augmentation technique termed as RadialMix that uses radial sample duplication to increase the
 general applicability, we further deploy SPINO on real-world robotic vision systems for both outdoor and indoor environments.

e experiments validate that our model achieves the best 3D semantic segmentation performance on SemanticKITTI and NuSce

, it shows better cross-sensor generalization abilities compared to 3D CNNs, enabling training and inference with different dept
in two complex indoor environments by utilizing a Turtlebot3 as the robotic agent. Compared to traditional frontier-based metho
at includes object concepts and open-world semantics. We encourage the readers to view the demos on our project page: url{h
iliary task to foster spatially compact predictions. Mask4Former achieves a new state-of-the-art on SemanticKITTI test with a sc
ables us to jointly optimize segmentation and tracking and directly supervise for 4D panoptic segmentation. We plan to provide
ver the baselines. Furthermore, we take an essential step from simulation to reality by bridging the gap from Habitat to ROS. Th
component to integrate expectations into abstract representations. This new concept is an evolution of the classical 3D scene g

te disparity maps that can be used by other perception tasks that require depth-related data. Comprehensive experimental resu
ur method in few-shot 3D scene segmentation against noisy samples.
RadarScenes dataset and is available at https://doi.org/10.5281/zenodo.10203864.

l results on the augmented nuScenes showcase the seamless integration of our module into three existing HD map constructio
kcheolShin/CRM_RGBTSeg.
e graph from SLAM including 3D panoptic point clouds. We extensively evaluate CURB-SG in urban scenarios using a photorea

he open-source code is available at https://github.com/jmwang0117/AGRNav.

n addition, we use the enhanced point cloud to perform object detection, localization, and mapping to demonstrate mmEMP's e
ating that sensor configurations significantly impact point-cloud-image fusion based detection models, which contribute up to 30
nt features without relying on a specific modality, thus ensuring a balanced and adaptive fusion for all-day detection.In addition,
ng to the 3D LIDAR input. We show that our CMDFusion achieves the best performance among all fusion-based methods on S
mprove the feature generation capacity when dealing with only IR images during inference. SK-Net outperforms state-of-the-art

ted HD map to a downstream path planning task, demonstrating that the long-range HD maps predicted by our method can lea

nd four Real-Time Kinematic (RTK)-based algorithms from GICI-LIB are evaluated using our dataset and public datasets. Resul
mp, and use RANSAC algorithm to extract the columnar shape in the cloud for peeling off the background. Finally, we fuse the r
ased methods. Our code is publicly available at https://github.com/sanjay-810/AYDIV2.
e and dataset will be released at url{https://github.com/MMOCKING/RadarCam-Depth}.
rained, and evaluated on the dataset.

eature fusion. The fused feature are then fed back to enhance the frame and event feature learning. Meanwhile, it utilizes an o

o propose the Tumbling-Induced Gyroscope Saturation (TIGS) dataset, which consists of outdoor experiments recording the m
KITTI/Waymo datasets and a quadruped robot for various indoor and outdoor scenes.
4.76 and a minimum manufacturing time of 24.6 minutes.
ep. During this process, our approach can be trained to accommodate geometry information of the target microphone array, wh

ase visit our project page: https://luoprojectpage.github.io/callirewrite/.

ly Sinhala-based self-communication features. This enables inter-party communication among doctors, medical students, and p

. Code and additional material are available at https://robokudo.ai.uni-bremen.de/rkop.

can synthesize a morphology that significantly improves the area coverage performance in an environment of interest.

rous of real-world experiments.

erances. It is a general purpose framework, not tied to a particular use case such as grasping or learning, it is written in C++ an
e of integrating real fish kinematics into robotic fish design and hydrodynamic studies in schooling fish.
m. The code is fully open-source, accessible here: https://github.com/AntoineRichard/LunarSim, and comes with demonstrations

ance in odometry.
rking tasks. Long trajectories (average 10~min) and many loops make the dataset ideal for the evaluation of SLAM, odometry a

he number of cameras and the overlap between their FoV that define the camera configuration for SLAM. All our software and

eal-world dataset, our sparsified backend is shown to be accurate and efficient. We open-source our backend, along with the V
and show the benefits of our proposed approach.
cheme where point-to-surfel association is prioritized over point-toplane. Extensive experiments on diverse public datasets dem

thods in handling outliers while providing online performances. We release with this paper an open-source implementation of th
ations, we also propose a novel patch descriptor distillation strategy to further reduce the computational complexity of correspon
ther optimize the camera pose and the objects. Meanwhile, local bundle adjustment is performed utilizing the objects and point
ture sharing representation to further optimize the performance. Specifically, the proposed cross-feature sharing module allevia

d on a real public dataset of several indoor office areas. On average over our datasets, S-Graphs+ outperforms the accuracy o

on 3D state estimation, we introduce a strategy that seamlessly integrates radar data with a consumer-grade IMU sensor using

e our attached video). By the way, our implementation VoxelMap++ is open-sourced on GitHub which is applicable for both non

AM estimates.
method consistently surpasses the baseline at comparable speed.
nd mapping accuracy in the several sequences of YCB-video dataset.
d on a few nodes, and improves the estimation accuracy of the sparsified graph. To validate the performance of the proposed t

onstrated that the proposed UWB-based radar SLAM is able to map natural point landmarks inside an indoor environment while

tasks we achieve equivalent performance to the SOTA, all while utilizing the same pre-trained model. Lastly, we demonstrate
scheme. To this end, we introduce an iterative primal-dual algorithm, capable of handling large graphs. Empirical benchmarks,

ces, outperforming the existing publicly available event-based stereo VO pipeline by two and four times, respectively.
ur approach with a multitude of different experiments using three different prominent real-world datasets.

We evaluate the retrained navigation policy in various crowd environments with diverse pedestrians. The experimental results sh
signed surfaces. We find an average 12.9% improvement on the MESSYSURFACES benchmark and an average 15% improve

ness in terms of safety, efficiency, and generalization in various challenging scenarios.

states covered by the user preferences and outperforms the two baseline methods.

arious types of continuum robots through parameter adjustments.

er promising applications.
d be restored to the initial state with an error of less than 3% of the actuator length after an operation cycle. The soft gripper bas

and 150%, across all traits for gripper harvested blackberries. Our study demonstrates the potential of soft grippers for delicate

as attained with a bounding locomotion gait at 19.6 mm/s (0.15 body lengths per second, or BL/s). Rolling locomotion attained b

pplied to other soft robots with facile integration and further expedites advanced functionalization in robot-object interaction.
of unlocking. The results validate that the ECR successfully accesses the gap (65mm) and accomplishes the unlocking task. T

escue, exploration, and inspection.

s, in which the gripper was also outfitted with ripeness sensing. Out of 40 berries, 62.5% (25) were successfully picked, with 0%
nd 5.1g respectively. The swimming speed of these two types of fish can reach 42.5mm/s and 17.0 mm/s. And this design prin

his work advances the precision and scalability of photothermal actuation not only for use in intelligent electromagnetic surfaces

ippers for the manipulation of textured and discontinuous surfaces.

lexibility and provides novel solutions for small-scale untethered manipulation.

ulated to balance the actuation distance and the actuator size, where the Pareto front demonstrates the trade-off between the a

D space verify the accuracy of the kinematic model, which indicates that the soft manipulator possesses a large workspace and
 n analysis and experimental examination of the operating principles and behaviors of the proposed mechanism. It also present
but the sides can stretch to allow turns. This steering method and body pressure do not work against each other, allowing the r
e employed for position sensing. This design offers a highly compact, lightweight, and low-cost solution exploitable in a wide ran
an co-exist without compromise.
t performance, with an inference time of 1.4 ms and a mean error of only 9.18 mm in Euclidean distance within 3D space, layin

e contact force.

sliding structure in ankle support applications.

fying shape and material properties of an object being grasped. Such manipulators can benefit from the softness and complian

e estimation model demonstrated a sensing precision of 0.05 N in the XY plane during testing, with data collection and training
d in series, multiple actuators form a soft robotic manipulator (i.e., arm), physically compliant or capable of delivering a relativel
wing us to put out potentially long- burning fires faster.

of the reactor, which can reach up to 0.22 bar, and demonstrated the thermo-responsive behavior of the reactor to trigger a bioc
dic sensors.

ment and compare it to alternative solutions. We release our tools as open-source software at https://github.com/ika-rwth-aach

Our results demonstrate GSL-Bench's ability to provide valuable insights into algorithm performance.
ncy (-59%), and cost (-42%) compared to a baseline that queries the LLM for every newly encountered action at runtime.

dated on a Clearpath Husky and we demonstrate that with this extension the robot is able to perform navigation tasks that woul

its relative location. System performance is assessed in terms of capture success and overall fuel consumption by the MUs. R

ased event representation that performs better than existing representations. In addition, we propose an effective data filtering m
o existing grid-based mapping algorithms. We validate our approach through rigorous simulation and real-world testing. This te
ce suitable for lunar applications and missions having similar challenging environmental conditions

or doorsill using its two adjacent shell-shaped legs. This paper introduces the design of the VWDER and analyzes the kinematic

n methods, our approach is more precise by 46.5% and 15.9%, and saves 55.2% and 25.7% manipulation time, respectively.

cvr.utm.utoronto.ca/ral2023-rollblock.html}.
d an improved surgical scene. To show the applicability of our system, we generated a dataset of 7.5k images with pose annota
e demonstrate between 100-5000x faster training times compared with previous surgical learning platforms. The code is availab
gical robot learning simulator SurRoL. The results show that our new approach achieves promising performance and task versa
effectiveness of the extensible bronchoscope.

cker achieves superior accuracy and performs more robustly against prior works. Code is available at https://github.com/wrld/Ad
and volumetric tissue, representative of most tissue scenarios. This work contributes to the advancement of simulation-based

y can be achieved by training separate models for the left and right robot arms. For task-generalization, MP recognition models

pproach reduces errors by 56% in mean errors as compared to no domain randomization. Furthermore, we demonstrate path f

ed for interventional procedures.

ulatures. The proposed methodology is tested with an MSRC prototype, where the magnet location estimation algorithm is imp

of clinical scenarios.

mmand a fleet of sixteen drones to follow a specified trajectory. We first run the controller on the simulator interface to verify pe
the system with real-world autonomous trolley collection tasks.
cluding stochastic bandits and a suite of TAMP problems, and compare our approach to a round robin baseline. An appendix c
the capabilities of the continuous sampling by introducing a novel informative path-planning strategy, which takes into account
ot and report comparisons on a series of mobile manipulation tasks of increasing difficulty.

f well-informed action plans for the next step. Moreover, we provided qualitative evidence highlighting the significance of levera

y constellations in collisions. To extend this work, testing on human samples and a collaborative effort from research institutes w

can adjust the step size according to the known map information and thus improve the efficiency and accuracy of detection; an
nducted experiments consisted in a loco-manipulation mission with the CENTAURO robot, a hybrid leg-wheel quadruped with a

r use in a wide variety of applications. The performance of the system is evaluated using an industrial task board for benchmar
peration testbed to train both user intention recognition and DRL algorithms. Our results show that the proposed framework can
nhancement in both the human-robot interaction experience and teleoperation performance.
d MMT. Another experiment with a human operator interacting with a virtual environment showed that using an offline estimated

ck. User study to compare the performance of POsticks and conventional interface reveals that POstick-KF and VF showed sup
vement in all objective measures.
s against full fine-tuning and other efficient fine-tuning approaches. The code is publicly available at https://github.com/JaronTH
sive experiments show that our method achieves state-of-the-art performance on the challenging MM-UDA benchmark. Code w
nipulation tasks. The empirical results demonstrate that our method not only achieves better performance of the transferred poli

ent model. In order to jointly alleviate the domain shift, GAS projects multi-geometric space features to a shared geometric emb

he proposed framework is evaluated by conducting policy learning experiments on virtual aerial (2D) and underwater (3D) robo

t types of mobile robots are utilized by the dancers to achieve their dance performances.

n with accurate object perception in manipulation tasks.

ch not only proves the system's capability to perform a variety of cooking tasks but also sets the stage for the development of m
ealth care and rehabilitation. This also clearly appeals as a robot-mounted sensor module to occupancy-aware HVAC energy sa
oration of ecosystems. Our vision for environmental robotics is demonstrated through the successful deployment of our prototyp

ody aspect ratios and refines person-wise features. Globally, it learns dense-depth-guided features for accurate depth estimatio
Neural Networks in a human-robot collaborative activity.

we evaluate the proposed method on both human and car localization. The results show that the proposed method is able to ac

usic intervention to those with dementia while also alleviating caregiver burden.
ge processing, augmented reality user interface, novel actuation hardware, and AI based catheter control. The project aims to e

sulted in a stable robot performance while preserving a high degree of human-likeness, as good as the best heuristic tuning.
els, it relies on precise robot posture control to spin around a pivot foot while the other foot is lifted. This unique dynamic config
optimizes global efficiency by fostering coopetitive (cooperative and competitive) interactions among robots, ultimately streamlin
dvanced prosthetic hands and reduce the burden of continuous grip force modulation on human users.

mic swinging over gaps with its limb. EEWOC measures 26×30×30 cm and weighs 2.1 kg, much smaller and lighter than o
Ultimately, the robotic systems developed in our lab prove invaluable to senior citizens, enabling them to lead more fulfilling liv
back of LED lights to determine whether the target object was securely gripped. This may be a possible way for human-machin
trol system employs a receding horizon planner, allowing it to autonomously follow prescribed trajectories with precision. This w
upport body weight while reducing wrinkles in clothing. It is expected that this will help prevent bedsores.
ntrol in extremely high-pressure environments.
 Finally, the proposed RTSRAs has been successfully applied in multiple typical scenarios including load operation, overhead s
s brachii, and participants reported their perceived exertion after each trial. The study revealed that medium or high stiffness le

periments on photo-realistic synthetic data and real household and medical data show that the proposed method outperforms e
ust control. We demonstrate the efficacy of our approach on a snake-like robot in both stationary and non-stationary (moving) c

nces. Both the code and the dataset are available for future comparisons. Both the code and the dataset are available for future
the efficiency and robustness of the fusion about UWB ranging measurements with the visual-inertial system, as well as the ea
y and robustness. It outperforms ORB-SLAM3, the current leading stereo visual-inertial initialization method, in terms of absolut

ety guarantees.

d in closed-loop control to steer the mobile furniture from initial to a desired location demonstrated in experiments on real hardw
dies, our approach excels over the existing baseline by consistently achieving higher metric values in predicting future events a
, significantly outperforming baselines, and is robust to different coarse maps.
kHz, enabling recognition with an event camera. Our experiments revealed the effects of marker size, LED light quantity, recogn
t RAPIDFlow outperforms previous approaches by significant margins at faster speeds.

ncy of our navigator. Code available: https://github.com/nuoxu/AKGVP.

t waypoint. Evaluations across both photorealistic indoor simulators and real- world environments validate the effectiveness of o

planner can adapt to diverse real-world environments without requiring any real-world training. In fact, the planner is trained pur
w that UIVNav chooses to visit the areas with larger area sizes of oysters or rocks with no prior information about the environme
terrains. Through physical robot experiments in outdoor environments, we assess PATERN’s capability to extrapolate pref
cacy of our proposed method in restoring low-light blurry images with increased visual perception compared to state-of-the-art m

that RenderOcc achieves comparable performance to models fully supervised with 3D labels, underscoring the significance of t

hub.com/longkukuhi/RoboLLM.

sualizations are presented in url{https://sites.google.com/view/fsnet/home}

ate-of-the-art (SOTA). The code can be found at bit.ly/v2ce.
oved pose estimation accuracy as well as better object detection accuracy.

method over existing state-of-the-art approaches. Our proposed solution can effectively address the challenges of real-world su
 xisting self-supervised 6D pose and size estimation baselines on the NOCS test-set with a 16.4% absolute improvement in mA
er both existing frame-based and event-based methods. Our code, pre-trained models, and dataset are available at url{https://g

eserves the thermal radiation of salient objects in infrared images while incorporating the textures of visible images.
tion. Extensive experiments on the public Waymo and NuScenes datasets show that our method can produce human-level cub

he BEV and perspective views. Subsequently, the Street Image Generation phase utilizes these segmentations as a condition t

em allows for dynamic adjustment of scale uncertainty in depth maps generated by monocular camera systems within SLAM. C

er baselines.

V-based reinforcement learning model with VPT, we demonstrate the efficacy of our model in reducing the number of required
o other state-of-the-art fiducial markers, even in scenarios with heavy occlusion. Moreover, CopperTag maintains an average p
st collaborative perception system perform comparably to systems relying on precise localization and clock devices. We will re

f-the-art methods, and improves inference efficiency by 90x, achieving real-time prediction at 6 FPS.
nstrate the effectiveness of the study in addressing the challenge of action recognition in occlusion situations. Our code is publi
ityScapes benchmarks, respectively.
tput undergoes a sequence of gamma-correction operations and generates a sequence of multi-exposure images that are com
r knowledge, this is the first work that explores improving visual BEV semantic segmentation performance using unlabeled data

resolution afforded by event cameras, our proposed multi-modal Event-RGB learning architecture effectively augments the per

mechanism to screen out reliable pixel-wise pseudo labels and ease the domain shift of the student model, which avoids error
n is incorporated into bird’s eye view object detection, a significant improvement in detection accuracy is achieved.
on. Three experiments are implemented to verify the application feasibility of the proposed deep camera position optimisation f
dataset and achieve superior performance with the SMAPE of 7.28% and the MAE of 3.18°, indicating a strong competitivene

e proposed approach for reflective texture-less metal parts has a significant improvement over the state-of-the-art approaches.

picking, resulting in significantly improved picking success rates.

s to large rotations. Our comparison against recent learning-based and classical approaches shows that RIDE sets a new state
ability of the LLM can bridge the knowledge gap when domain shifts and data imbalances occur. We then put forth a novel data
and image segmentation to improve robot-assisted manipulation of soft tissues in ophthalmology.
to a new location. We believe that this algorithm is a step towards enabling intuitive interfaces that allow a user to initiate a gras
between the centroid of the hand and the object is calculated. The experiments were conducted using twelve different food item

mic properties in both image and Cartesian spaces, superior tracking performance, and less sensitivity to noise compared to vel

stem. The darkness clue prompting efficiently injects anti-dark knowledge without extra modules. Code is available at https://git

s seen from different angles. Additionally, a Dual Motion-based Prediction (DMP) method is employed to forecast the object pos

ned box to segment the objects. We decide the termination of an object track based on the objectness score of the propagated
alization network, which converts pillars to bird’s eye view (BEV) feature maps and predicts the state of the target with a sm
ch we can verify with cycle-consistency. This produces a sparse but accurate pseudo-labelling of the video. In the second stag

uces the number of ID switches by around 19% and improves mMOTA by 2-3%.

2.4% AUC on TrackingNet, and operates at 171 fps on an NVIDIA 2080Ti GPU. The code and demo materials will be available

ust tracking in regions lacking texture and during rapid rotation. We evaluate the performance of VI-HSO on public datasets and
tasets and the results demonstrate that the proposed method is able to suppress VIO drift in real-time and also show the effect
we develop a demo that can run on a mobile to demonstrate the real-time capabilities of our method.

website: https://youtu.be/KnWZ4anBMK4.

os, assisting users in achieving free movement of 400 meter at an average speed of 1.34 m/s, meeting the level of normal peop
gh support performance in STS with the CoG tracking strategy.

ing sit-to-stand tasks. The results show that the exoskeleton can recover energy while assisting the subject and reducing mean

to realize the assistance. Finally, the experimental results are presented in this paper to validate the proposed control scheme.

ations in the future. Finally, experiments demonstrate the proposed LIET framework could speed up the exploration process and
ction with a customized guide rail to adjust the center of rotation of the SSLM. This configuration facilitates knee flexion/extensio
 up entire stairways. To the authors’ best knowledge, enabling the latter by means of trajectory optimization is a novelty for

e also validated by the prototype.

ctory to the reference one is strictly proved by Barbalat's Lemma. Finally, simulated and actual experiments on a car-like robot d
cities up to 3 m/s in simulation. Experiments prove that the used platform, Go1, can successfully attain a stable standing config
botic arm, showcase its ability to successfully lift and carry an 8kg payload and manipulate doors.

mprovements since the operator can utilize the feedback for reactive short-distance planning and visual feedback for long-dista
-of-the-art performance on point cloud domain adaptation.

and show that it consistently improves the recall and pose estimation of 3 state-of-the-art metric localization architectures while
in the context of RGB videos, we surpass the state-of-the-art method on Kinetics-400 by 2.1%.

methods, without the need for task-specific fine-tuning. Experiments are performed on LMO, YCBV, and TLESS. In comparison
ed model outperforms state-of-the-art methods in action recognition accuracy.

the BSA to generate consistent launch sequences while the SEA has to rely on lengthy and possibly dangerous oscillatory swi

he performance constraints of heterogeneous UAVs to plan the trajectories of the mission UAV and relay UAV in order to maint

o-drones not only marks a novel result for hardware-limited robots but lays the ground for more general advancements for the

a-learning on the manageable disturbances). Experimental results demonstrate that HMAC exhibits more precise and rapid ada
capable of achieving 120 fps multi-object search and detection. Extensive experiments are conducted to verify the system's effe
uperior performance in terms of both accuracy and efficiency in robot scene perception and autonomous robot navigation. The r

dient descent counterpart.

e precise 3D manipulation waypoints via a designed kinematic-aware chain-of-thoughts prompting method. Our evaluation span
ning compared to state-of-the-art LLM-based planners. Moreover, it also preserves the broad applicability and generalizability o
ge, this is the first framework that controls humanoid robots using a single learning policy network and LLM as a planner. Exten

cision making. Our universal framework applies to any belief-dependent constrained continuous POMDP with parametric belief

dered. Please see project website for prompts, videos, and codes.

attributes, to serve as a demanding learning environment for advancing robotics and automation research.

tate predictions in the examples helps reasoning,(iii)instruction-based prompting benefits from human feedback. Our observatio

g navigation robustness and efficiency.

viable path for scaling up reinforcement learning for robot manipulators in simulation, both by diversifying and expanding task an

as planners in complex task domains. See our project website for prompts, videos, and code.
and unexpected disturbances. Supplementary videos at adaptiveskillcoordination.github.io.
he space requirements of a robot’s LTM while maintaining its capacity to successfully perform tasks relying on LTM informa

environments. Our software is available on GitHub with the experiment video.

a desired versus an undesired behavior. The evaluation on high-dimensional robotic datasets indicates data efficiency in RoA e
 en trajectories and high route completion. Our project page can be found at https://changyaozhou.github.io/Autonomous-Vehic

t our system strictly adheres to the safety constraints and scales well with complex temporal constraints, highlighting its potenti

CHBA over all other selected algorithms.

eness within the motion prediction task.

with few radar targets or with heavy metal structures. Our method keeps up with common approaches such as wheel speed sen
he performance of daytime navigation systems using RGB cameras. Our work paves the way for scene perception and trajector

TION and NGSIM datasets, and improves over the baseline (MFP) when both closest-mode and probabilistic evaluations are u
environments.
ory planning.

hes between geometric and learning based planning. Our experiments on both SCAND and two physical robots show that the

s demonstrate that our approach significantly outperforms the state-of-the-art methods.

n-world navigation with a unique capability of localization recovery, significantly reducing the need for human intervention. Code

set, and its validity is demonstrated by comparing with the previous methods.
 against state-of-the-art (SOTA) LfI baselines, and results of both simulation and real-world experiment confirm the superiority o

MTraMap of the surrounding 20 m × 20 m environment at a rate of 9.39 fps, with a slight reduction of 0.61 fps compared to th
ge-scale deployment.
of our method to diverse unknown environments, along with its superior gas source searching capabilities compared to conven

e LiDAR data into the motion forecasting model training and provide a strong baseline. Experiments show that the LiDAR data

tion and improve on a target task within as little as 3 hours of autonomous real-world experience. We also demonstrate in simu

robots compared to conventional UAVs.

FWUAV is capable of maintaining a closed turning maneuver for several laps. Therefore, the insight provided by the model prov

ctiveness of our design and highlight the potential of RoboTwin for future applications in exploring further hydrodynamic interac

es the Strouhal number by up to 21.9% at high frequencies without extra power.

bled policy representations are all beneficial for energy efficiency, robustness to noise and sensory delays of 90 ms, and tracki

le at https://mavlab.tudelft.nl/learning_to_home.
well as the relative on-board applications such as bio-inspired vision.
ying an additional 72 mg load, achieving a record 5.5 thrust-to-weight ratio. These results open possibilities to incorporate micro
 local and global chemical modulation on within-network signal transmission.

ll achieved in an open-loop manner, without the need for environmental awareness.

orithm, the physical limitations on tail and wheel velocities are automatically accommodated. Numerical comparisons among wh
d by imparting magnetic torque originating from an external magnetic field onto the magnets embedded in the modules.

rison to the baseline algorithm. Our source code, demo video, and supplement are publicly available at mias.group/StereoCalib

sier-to-use extrinsic calibration approach that maintains high accuracy for both event- and frame-based cameras.
executes in 93ms on a single GPU, meeting real-time requirements. Despite the modest 9M model parameters, LCCRAFT achi
and dynamic coupling. Moreover, the PINN exhibits the potential for extension to various dynamic systems.

rd deviations of 0.37mm, 0.65mm, and 0.40mm (translation) along x, y, and z axes of the end-effector, respectively. The tool al
m (DOF) robot arm.
—¦ on the KITTI dataset, surpassing existing state-of-the-art methods and demonstrating strong robustness, accuracy, and gene

racteristics of 4D radar. Experiments with two 4D radars in real environments with four configurations demonstrate our method

efficiently navigate to target objects within an office building in the real world, without any prior knowledge of the environment. T
n with five alternative methods, and demonstrate significant improvements in performance and lower collision rates, despite utili

nsor observations and sim-to-real transfer.

ting learning-based methods but also upholds performance in complex and various collision scenarios.
n path for the robot. Specifically, we employ specific action space discretization techniques and tailored state space representa
jectories compared to baselines in interactive scenes.

her sensor modalities.

ated and real-world scenarios.

ield. In order to facilitate further research within the community, we have made the source code of the proposed algorithm publ

oor environments. We observe an increase of 10-30% in terms of navigation success rate and up to 15% increase in AU-ROC c

enefits of our physically grounded VLM on a real robot, where it improves task success rates. We release our dataset and prov

world object reconstruction scenario.

nds of training time.

en more efficient offering an additional 5x faster inference speed and 8 ms ultra-low latency at the cost of only a 9% performan
es, showcasing the effectiveness of ActFormer in multi-agent collaborative 3D object detection.
1011.github.io/r2dm}.
of the methodologies integrated into our framework.

at surfaces.

construction capability of TacShade for a large panel with ridged patterns based on the location of robots and image splicing te
plied to grasp control to enable a quantitative trade-off between the likelihood of slipping vs. grasp force levels, and to sliding m

ve experiments including image classification on ImageNet-C and semantic segmentation on GTA5, C-driving, and corrupted Ci

al scene with a UR5e robot arm. Experimental results indicate that our method achieves higher success and lower buffer size re

d vanilla LLM approaches in the lifelong setup, while learning skills that are transferable to the real world. Project material will b

containing both complex perturbations and confined spaces, while also having the potential to improve exploration efficiency thr

eral promising research directions are mentioned involving the proposed reconfigurable robot architecture in cleaning operation

and Scene-Graph Gener- ation. Additionally, we utilize our dataset to examine the effectiveness of OOI and HOI-based feature
onstrate that RTSs provide measurements that are 22 times more stable than GNSS in various environmental settings, making

ub.io/interact/.

arch opportunities.
pable of performing a variety of grounded high-level reasoning tasks in broad realistic settings with a cognitive intervention rate

nd ensuring a high-quality dataset. The dataset is made publicly available on our website: https://rh20t.github.io.
nimize counterfactual perturbation. We provide supplementary videos and make publicly available the visual navigation dataset
t publicly at https://hcis-lab.github.io/RiskBench/

nd demonstrated its potential applications. Project page: https://opendriving.github.io/lidar-cs.

 detection model are available at https://github.com/SonYeongGwang/CoAS-Net.git.

osed methods through a series of experiments.

ng and multiscale information integration. Consequently, TrA enables the model to focus on critical local structures and extract r
found on the project website.
ulation. In contrast, TRP are allocated to positions that are responsive to outputs with minor distribution shifts, which are fine-tun

gy, utilizing natural robot features in the runtime and not requiring camera calibration or robot models in any stage of this proce

schlieren from both a static and moving camera on previously unseen flow patterns and background images.
Finally, our in-field tests show a reduction of the average tracking error of 37% compared to a previous SoA work and an endu
d to the existing method.

across varying camera intrinsics. Experimental results demonstrate that the proposed method outperforms the state-of-the-art
e 3D lane detection task on the K-Lane dataset and our LiSV-3DLane.

proach in extensive simulation and real-world experiments on multiple kinematically diverse mobile manipulators.
Rate-Distortion (RD) optimization. Our approach outperforms existing learning-based lossy point cloud coding methods through

aselines. Particularly, SafeDPA demonstrates notable generalizability, achieving a 300% increase in safety rate compared to th

outu.be/fPKOrjC2QrU

able task-oriented grasps with 82.3% accuracy, having a physical grasp success rate of 82.3%.

d evaluation are available at https://sites.google.com/view/geometric-peg-in-hole.

cts and multi-agent settings, including human-robot collaboration.
ability, where it can successfully grasp and lift objects in real-world scenarios for unseen visual configurations.

0%. We further demonstrate the generalization of our policy by picking a set of unseen objects. Supplementary material, code,

grated multi-view input and sim-to-real training design, our method improves real-world applications of autonomous grasp.

complete the object picking task in 98% of experimental trials.

xperimentally validated in the real world for specific manipulation tasks, achieving a 94.2% mean zero-shot overlap success rat
n and cross-category articulated objects. Images, videos and codes are published on the project website at: http://sites.google.c
world experiments. We also outperform state of the art approaches. Moreover, the proposed approach exhibits its usability in re
and-place task. We observe that the robot can solve the task successfully. The videos and code are publicly available at: https:
s://pku-epic.github.io/ASGrasp

ation strategy with active perception, achieving a balance between 6D pose accuracy and manipulation efficiency. Our experim
pose consistency-aware cross distillation (CCD) to transfer knowledge between the pixel-wise predictions from the cohort. We v
minated crack images and 41 low-light crack images. The dataset and code are available at https://github.com/zy1296/CrackNe
ach combines static binarization of the overall network model with fine granularity dynamic binarization of data-sensitive compo

ate exceeding 80% in both simulated and real-world scenarios. A video showcasing our experiments is available online at: https
erse Jacobian-based controller can support real-time manipulation control of up to 900Hz.
g manipulation skills. Compared to methods without pre-training, our approach achieves a success rate increase of over 60%.

e/mVm0qCvB7HY.

0 robots and hardware experiments of 13 robots, against several strong baselines such as GreedyNE and FastMaxSum.
age ratio, and task execution time compared with others.
 xis, y-axis, and z-axis, respectively, and 0.8502 degrees, 2.1941 degrees, 0.2408 degrees corresponding to three attitude angl

e. Experimental results and comparisons with baseline methods establish that our proposed method outperforms other approa
to a variety of real-world insertion tasks involving narrow or even negative clearances, all without requiring any fine-tuning.

anipulation platforms to successfully execute diverse assembly tasks.

agnetic data of multiple sessions. Evaluations on both simulated and real world experiments demonstrate the greatly improved u

methods in most cases produce better patrolling strategies when compared to classic solutions. Moreover, for small problem in

w the effectiveness of the approach through numerical simulations as well as with preliminary experimental results.

es failures from raw sensory data without any prior knowledge. In this work, we use our extended FAILURE dataset [1] with 99

sconnect of power while perched. These results suggest that such a system could be a valuable tool for the upkeep of electricit

y, stability, and smoothness of motion.

times ranging from a few minutes to around ten minutes. Overall, this research contributes to advancing the capabilities of ULR

) pipeline is applied to generate the 3D shape of the target. To our best knowledge, this is the first time autonomous inspection

proaches to further optimize their energy consumption need to shift towards efficient robot electronic design instead of efficient
ity with standard UAVs, only necessitating an API for attitude setpoints. Experimental results underscore the efficiency of our m
st to the traditional sampling path planners and various reward functions.
emporal datasets, we demonstrate the benefits of our approach over traditional deployment strategies, including significant sav

pose and force estimation tasks with the minimum error of 0.08mm and 0.03N, respectively, in contrast to ViTac's 0.12mm and

ous work by the authors and the state of art, both in terms of accuracy and classification speed.
ents can sufficiently extract physical quantities from visuotactile sensors.
actile information for fine manipulation.

s more practical compared with equipping commercial three-axis force sensors. Thus, the L3 F-TOUCH could further empower

with a comprehensive video tutorial, all of which are available at https://linchangyi1.github.io/9DTact.

compliance levels of the specimens, both when the sensor probed the surface perpendicularly and with different inclinations. Fi
n the real world. These tasks and our system can also serve as a benchmark for bimanual tactile manipulation. Code will be ope

rios. By reflecting contact status, this sensor provides a promising glimpse into the future of agricultural automation, enhancing

d in extensive experiments with test data collected from real sensors and also shown to maintain embedded force information w
drical sensor, resulting in large surface reconstruction. We compare our method with the representative lookup table method in
e images generated via our method with those from prior methods. The generated tactile images from the proposed model sho

chniques. The proposed method enables a robot to unfold back-opening hospital gowns and perform upper-body dressing.
resentations as evidenced by linear probing assessments.
orce control.
d using experiments. Consequently, a force sensor with the same level of performance was developed using the proposed met
ble manipulation capabilities to human operators with only force sensing, indicating a high level of adaptiveness.

r design is shown to be a viable solution for increased robotic deployment in flexible manufacturing lines.

e optimal solution would otherwise be computationally expensive. Our approach efficiently explores the solution space, generati

monstrate the effectiveness of the T2TL framework.

ng models for PRISM-Games.
s of the proposed decision-making and motion planning.

arm robot is used.

rmance over a range of simulation experiments.

onal grasp dataset named REGRAD demonstrate that our method significantly improves the accuracy of relationship detection

s on a simulated human hand and a real prosthetic hand.

asping and reconstruction. Additionally, we demonstrate real-world applicability by decluttering scenes with varying numbers of
nually updates the model from the experiences of the human-robot team collected from trustworthy deployments. Consequently
d models. In addition, our proposed representation method demonstrates promising performance when applied to a different rob
n centering, agency handover, and actor-critic ensembles. We empirically demonstrate the contribution of these ingredients in f
rasping capabilities in real-world scenarios.

lation skills of real-world reconstructed objects.

tion applications.

eline method, in which the joint angles obtained by a motion capture system are served as ground-truth. The results demonstrat

age level of about 0.9 in the Structural Similarity Index Measure (SSIM) evaluation, when compared to the ideal single-point-foc

y, we find that the resulting network entropy maps can be used to identify high salience portions of a tactile scene.

nary results indicate that combining adaptive oscillators with interaction force sensing can improve transparency compared to th

and, reporting for high reliability over repeated grasp-transportation-release tasks with different objects. Statistical support is pr
ceived comfort, and foster trust in the robotic partner.
by catastrophic labeling errors are completely removed. The open-source tool Labelbee was employed for 53,026 iterations of

h, we conduct experiments involving actual human users in a simulated social navigation setting. Videos of the experiments can
increase in success of 0.38 with only a 0.3 expert interaction rate within the Habitat simulation environment using a simulated h

e and offers more robust performance in worst-case scenarios compared to state-of-the-art baselines.

subjects is measured as 0.98/0.91 seconds (0.5 kg), and 1/0.86 seconds (1 kg). The average effort reduction among the subje

egrate into robot navigation systems, enhancing their ability to navigate crowded areas smoothly.
rians' personal space by up to 19 percentage points for the most difficult cases. We also describe how to apply these modificat
ajority also showed interest in understanding robot's intention (path plans and decisions) through various modalities like speech

them to baselines.
ptimization tries to stay close to. Furthermore, by leveraging free-end homotopy classes—a novel concept we introduce in this
, and fastest computing efficiency. Experiments are implemented in long corridor and broad room scenarios, where the robot g

representation. The source code and EEGTrust dataset are available at https://github.com/CaiyueXu/EEGTrust.
etting reinforces the powerful nature of kernel methods for learning to recognize complex patterns.

ility study conducted. Results show conventional models' performance significantly degrades with dropout and is recovered usin

the effectiveness of our robot ToM-based robot policy in multiagent interaction settings.

uide users in realistic running scenarios on sports tracks. Subjects in this experiment wore an eye mask to simulate the visually

prediction window with inference times of 0.26 ± 0.44 ms. Performance was higher for short range predictions with errors ove
nerating explicit maps achieves significantly higher accuracy than storing implicit maps in the LLMs.
ation pHRI controllers supporting concurrent transparency and stability.
successfully generalizes to real-world complex scenarios.
ess, and stability, also favored the aforementioned controller.

translation of the scLDTL formula. Finally, we show that the optimal policy can be obtained by leveraging existing point-based v

a significant improvement in the map quality and mapping duration.

peration, our method can achieve faster and more consistent peg transfers. An analysis of the participants' effort also reveals a

al number of corrections needed in the first round and requires little to no corrections after 2 iterations.

nto multi-arm systems, the ability to perform robust interactions in highly variable environments, and the provision of a compreh
nning problems and agent models.

he proposed framework and compare the proposed framework with existing methods. An application example is presented to s

-shot VLN model by a large margin on all metrics. Additionally, real-robot experiments display the obvious advantages of our m

mulated social navigation environment. We release code and video examples at https://sites.google.com/view/rl-polite
odes the robot's constraints and abilities into a pre-trained text-to-image model, showcasing promising results as an effective m

consumption and deviation in their tracking as they performed a post-fall recovery was reduced considerably.

he very end of the experiment. We found that while the average between-video rating of robot competence remained stable ac
rary motion), LaCE-LHMP extracts the laminar patterns in human dynamics and uses them for human motion prediction. We d
cture to allow a physical robot to continually adapt to the changes in the environment from limited data provided by the users (e
. The code and dataset have been released at: https://github.com/fei-aiart/HRICA.
h a varying level of external stimuli, and found a significant improvement in the performance of participants in both VR groups c

ble to draw visually pleasing oriental ink paintings with expressive strokes.

desired speed changes.

the human-like features make the proposed method suitable for devices that provide perceptual feedback to human users and

he layered structure. A physical robotic shoulder prototype has been built to verify the effectiveness of our method.

ansmission costs.

and exhibits outstanding generalization capabilities.

benchmark demonstrate that our FastOcc achieves state-of-the-art results with a fast inference speed.

ticKITTI Scene Completion benchmark. Codes and models can be accessed at at https://github.com/ucaszyp/MonoOcc.
e how ShaSTA may impact the ultimate goal of an autonomous mobile agent. ShaSTA achieves 1st place amongst LiDAR-only
n systems. The code is available at https://github.com/jinlong17/BDS-V2V.
e transferability across different point cloud based 3D detection systems. This alarming revelation underscores the pressing nee
p when transitioning from curated to perception-based data. In particular, we show that this gap (1) stems not only from differen

g in enlarged perception range within the limited bandwidth. Extensive experiments on the DAIR-V2X dataset demonstrate the
nt results show that our algorithm improves the tracking accuracy by 17% with only 0.037x communication costs compared with
will see larger and larger models being used in offline-training-online-deployment imitation learning tasks in robotic applications.

results on the Mujoco dataset show the effectiveness of the proposed approach.

mitation learning scheme resulting in an algorithm that can learn behaviors without exploiting the learned reward function. We de

plications even in limited computing hardware. All the work presented here is open-source, including models, weights, assessm
mates appear to correlate strongly with the probability of escape of an object under randomized pose perturbation.

s have faster training and higher task performance than baselines and can solve real-world tasks when selected tactile features
lp sequence real-world, high-performance policies. We describe the framework more generally and create a specific instantiatio

s and Reinforcement Learning navigation algorithms in metric terms. Additionally, it shows remarkable scalability and generaliz

er, on the other hand, takes responsibility for following the unknown contour on the manipulated surface by only taking observa

rovements in both success rate and speed by up to 60%.

MPNet and advanced sampling based methods IRRT* and BIT*.

easurements that are available in an operational setting. This modelling paradigm is known as virtual sensing.
mans in specialized tasks but achieved only 16% of human speed across the full set of tasks. Finally, we present the design an
sformation is also numerically implemented for the proposed nonlinear endpoint control method based on the original PDE mod
f smart garment manufacturing.
o control the SCBG model to generate realistic behaviors with desired courtesy levels. SCBG is able to identify different motion
environment, and our results demonstrate that our proposed CDAS has a similar perception of risky behaviors compared to hu
bust performance for the general case, i.e., safety-irrelevant pedestrians.

ally, we achieve a significant improvement on the average precision of Intersection over Union (AP@IoU) by 7.15% compared w

ps with an accuracy of ∼ 80%, which improves up to ∼93% when the target’s state information is available. Lastly, we

his work, we show robot vehicles using image observations can achieve competitive performance to using precise information o
via CARLA. These performance metrics are comparable to the most advanced driving models.

more accurate and scene-consistent predictions. The demonstration video is available at: https://youtu.be/Wjla071BPBA
able and safe deep-learning models for motion forecasting.

orithm that is executed in parallel for each sector. We provide two main theoretical results: 1) characterize the number of taxis
training, aligning them with the evaluated capability of AV. Accordingly, CLIC not only maximizes the utilization of the vast pre-c

combining with our proposed strategies, classical models outperform their baselines, further validating the versatility of our fram

n instances). We validate eMem-NDT on BLVD and LOKI datasets, and the results show that our model can obtain a superior
existing warning generation methods.

me delay are controlled at approximately 10°and 300 ms, proving the feasibility of the subject-independent estimation method
us. A real-time implementation in a construction HRC scenario (collaborative tiling) showcases the intuitive nature and potential

optimization, with potential applications spanning both biomechanics and robotics.

curacy with other state-of-the-art approaches.

mately 10% error relative to experimentally measured displacements, as well as to each other. And our method outperforms SO

rage and peak jerks of the moving platform are decreased by 14.6% and 37.6% compared with the active-joint space method.

reduce conservativeness on the small quadrotors.

localized activity predictions can be incorporated in a human-aware task planner for an assistive robot to reduce the occurrenc
space. More details regarding the real-world experiments are presented in: https://youtu.be/OM-2edHBRfc.
yed to a real-world robot. We compare EditMP with probabilistic-based MPs and empirically demonstrate the state-of-the-art ad

uch as the maximum stiffness, robot base position, friction coefficient of the sawed object, and movement period.
he properties of real objects, explaining its contribution to improving the generalisation of the downstream task.
challenging environments.
model predictive control.
g-rlap.github.io/.

arying levels of proficiency in terms of excision forces.

clude a fixed-base manipulator and a quadruped-mounted manipulator.

ation resources can extend the range of applications of this LfD framework.
near-zero normalized evaluation scores on the medium and large antmaze tasks, while our modification gives an average sco
hion and leverage them through planning, significantly surpassing baseline approaches for skill discovery.

ent representation. RLTSC substantially enhances data efficiency in vision-based RL, surpassing state-of-the-art methods on v

mulated and real robotics tasks demonstrate that our approach outperforms naive domain randomisation and model-free meth

pushing. COMPOSER achieves the highest success rate across all tasks when compared to a centralized baseline and four m
 e underwater environments. Comprehensive experiments demonstrate that our framework not only achieves state-of-the-art (S

and extending the applicability of our framework.

ule (REM) is also designed to determine the temporal distance of subgoals. Moreover, REM computes intrinsic bonuses to enc

monstrate the efficacy of our approach across varied targets and trajectories. All code, data, and additional videos: https://kl-rese
d codes are available at https://sites.google.com/view/lcs-rl.

el and tactile ensemble policies. Results indicate that TEST can achieve a success rate of 90% and is over 4 times more efficie

experiments where the robot watches 20 demonstrations of a tabletop task and transfers zero-shot to a mobile manipulation ta
r impairment and physical perturbations without any retraining.

evel policy into consideration. Comprehensive evaluations demonstrate that using the proposed PNSS module consistently imp

sor setups than the ones used during training. We make the code and trained models publicly available at http://ralf.cs.uni-freib
duce high-quality reconstruction results on the Replica and ScanNet datasets. The source code will be available at https://github
eless digital twin (WDT) built upon simulations of a large class of indoor environments from the AI Habitat dataset augmented w

faster convergence and improves the learned behaviors in various challenging robotic tasks, from climbing boxes with a quadru
ult shows that our approach achieves outperformance over search-based methods in total execution time and computational e
e Waymo Open Motion Dataset (WOMD) demonstrate that GOLD can successfully distill lightweight policies and solve decision

es and their corresponding ghost nodes. The HTP leverages graph neural networks to capture correlations between agents an
n natural environments for elevation mapping and traversability estimation compared to using raw sensor measurements and ex
rnal forces compared to using a locked arm.

ower joint torques, which we successfully transfer to a real Unitree A1 robot without any additional fine-tuning.

at a speed of 2.5 m/s. Remarkably, this framework also allows the robot to maintain stable locomotion while bearing an unexpe

hibits superior task performance in real-world evaluations compared to PPO.

to explore its own locomotion skills within the entire solution space and facilitates smooth transitions between local goals, enab

ntroller with lateral spinal flexion, our proposed optimized controller effectively improves the dynamic balance of the robot and r

ss different robots, the only varying component is the PF layer, which adjusts the scaling parameters for the stride height and le

body weight).
tation time is reduced significantly.

eedback control effort than trajectories generated with traditional methods.

viously impossible for this robot.

h as to indoor environments and aerial imagery.

es 3.6% and 4.1% absolute improvements on mAP on the nuScenes and the Argoverse2 datasets, respectively, demonstrating
formance even when using an average of only 3.7 bits per measurement, equivalent to a data reduction of 8.6 times compared

emory usage on both simulation and real-world datasets. The project page is: https://yuxuan1206.github.io/NFAtlas/.

google.com/view/visual-traversability-learning.
xperiments in diverse conditions, including automobiles and unmanned ground vehicles, herbfields, woodlands, and farmlands

quences from the Replica and ScanNet datasets and demonstrate 3DIML’s effectiveness under mild assumptions for the im
s still guaranteed on NCLT and M2DGR.

eased Colmap-PCD, an open-source tool leveraging the Colmap algorithm, that enables precise fine-scale registration of image
ne approaches fail.

on method significantly outperforms existing satellite-based systems. Moreover, in scenarios where the robot struggles to find lo

ResNet18 as the backbone.

ethod. The implementation of our method has been released at https://github.com/cloudcjf/CCL.

between the two 3D scene representations and 2D re-projection errors between the 3D global scene representation and 2D im

ty of the aggregation network. Extensive evaluations show that VOLoc provides localization accuracy even better than the Lida
nition algorithm, while being several orders of magnitude faster and suitable for real-time deployment – with inference speeds

ilding VPR solutions that may be deployed anywhere, anytime, and across anyview. We encourage the readers to explore our p
hub.com/ilaydayaman/LuViRA_Dataset.

own cases where existing filters fail without careful exception handling or case-specific tuning and benchmark against the ArduP

conventional PF. Moreover, vehicle position estimation experiments conducted in urban environments reveal that the propose
gorithms that use the current, feature-rich cluttered map for localization.

tforms in feature-sparse and low-visibility conditions such as mining, construction, and search & rescue operations.

mputational efficiency.

lation platform and a real-world 7-DOF robot.

rovement for out-of-distribution ropes in simulation, as well as a 26% improvement for ropes and a 50% improvement for cloths

oreover, because DMPO requires fewer samples, it can also achieve these benefits with 4.3X less memory. When we subject th
ess conservative (i.e. faster) as its understanding of the target system dynamics improves with more observations. In particular,
for PCB board assembly, cable routing, and object relocation in less than an hour of training per policy, comparing very favorab

ur framework produces policies that greatly improve over other prior methods. Our video and additional details can be found at
d these insights to further fine-tune the turning gait, enabling the rover to turn nearly 90 degrees at just above 4 seconds with m
that our approach has better task performance with less safety violations than state-of-the-art algorithms.

ation challenge in manipulation by learning a conditional distribution, which is an effective way to move towards generalizable o
ge-scale simulated noisy point clouds mimicing real-world scans, and in the real world scenarios, with superiority over existing

UncertainShapePose.
nts' and ii) search for target objects using natural language queries describing spatial relations of objects. Furthermore, we quan

of higher quality. We also show that our method only incurs a linear factor runtime overhead with respect to a lightweight baselin
ubonan123/OmniColor/.

tics, we make the design open-access and freely available to the community.

g agile robotic fishes.

ollision experiments present the robot’s potential applications in narrow spaces. Overall, this omnidirectional swimming rob

self. Simulation experiments, in different challenging scenarios, showcase robustness and efficient real-time performance of mo
GPU systems. The implementation of this work is made publicly available at https://github.com/ebnerluca/uw_depth.

ade public on https://github.com/rpl-cmu/sonic to facilitate further development in the field.

 present the overall design and implementation of the system, as well as provide analysis of some of its capabilities.
of MPC-CBF to disturbances.

ion. The low computational load of the proposed models makes it ideal for integrating in real-time visual navigation and reconst

hat there is an increased accuracy compared to conventional SLAM and that the software is usable in real-time during a missio

nderwater caves demonstrate the improvements in accuracy and robustness introduced by the proposed VIO extension.

this is the first work to incorporate a vision-based tactile sensor into aerial interaction tasks.

n studies where large overshoot and even collision with a perching site occur. To the best of the authors' knowledge, this work p
ping and transitioning of aerial transportation and manipulation systems to real-world deployment.

is verified to be consistent with the Computational Fluid Dynamics (CFD) and wind tunnel experiment results. Finally, we optim

ated the minimum idle time required by the proposed system after the system perching and motor shut down to save energy on

lose proximity (0.2 mm) to the ground plane and as much as a 20% increase is measured even when operated centimeters aw
paths when μL and μR approach zero.
% or higher. The efficacy of the trajectory planner is also validated.

ly integrating the proposed control method for practical usage. Real world experiments are conducted to validate both the contr

ght is reduced by 60.30 %, on average. To our best knowledge, these are the first flight-test results that thoroughly demonstrat

can adapt to various robotic arms and interactions when performing multi-task aerial manipulation.
as many drones as necessary for the transportation of any single payload.

ith data from numerous experiments and thorough statistical analysis, which indicated promising directions for future developm
ile mechanism for fast prototyping.

ating a torque bandwidth of 66.3 Hz at an amplitude of 6 N·m, ±0.11◦ of backlash and 0.37 N·m of backdrive torque.
gation performance of the proposed scheme.

erage error of 3 degrees in trocar orientation estimation, with an average processing time of 15 fps. This could serve as a found
, such as ascending ramps, avoiding obstacles, standing on tip-toes, and tapping the foot. In this way, PVI-HVC offers the kine
heme to assign an importance factor to each pair of classes, controlling the impact of their confusion on the overall uncertainty.
Impressively, the TSER boasts an 88.11% accuracy rate in haustral fold depth estimation, surpassing the performance of exis
esearch. The hope is that this review will aid users in identifying and addressing common limitations of the systems promptly, th

cts from breathing motion, and yields smoother 3D reconstructions for enhanced screening and medical diagnosis.

registration is more effective and robust in adapting to the inter-patient variations than the ICP (distance error mean±SD: 5.0Â
3 ± 5.3%, which implies that the proposed method is valid.
s. This quality-aware module enables the closed-loop control of robotic screening to further improve the confidence and robustn

o our previous work where the system was firstly introduced. The results of the proposed compliance control demonstrate cons
he entire experiment. The results suggest the potential for advancing the intelligence level of neurosurgery in the future.
mean bidirectional precision of 0.0485 mm. Also, 10.0 mm diameter circles were traceable with the same scaling factor.
ared to systems lacking haptic assistance. With the aid of our system, surgeons were able to safely skeletonize the critical struc

licate blood vessel during lymph node resection. It shows the potential to enhance the safety of robot-assisted lymphadenectom
sual tracking and tactile interaction guidance. Experimental results demonstrate that the proposed method not only allows for FO
mm (4.2% and 57.9%), respectively.
der-follower grasping force control.
ntally validated. The results demonstrated that high-resolution motion could be achieved within 10 um. Furthermore, a demonstr

n process of RMCR is described in detail. Furthermore, a kinematic model of the RMCR is established, simulated, and experime

hanism to quantify the models' confidence through uncertainty estimation. We validate the uncertainty estimation through exten

ed into two categories representing benign and malignant lesions.

ndovascular interventions. The proposed method was validated on a hold-out validation dataset, thus demonstrated robustness
over, for the first time, we compared the performance of the proposed adaptive teleoperation mode with the cooperative mode b

n leveraging technology for innovative medical solutions, particularly in the management of antibiotic-resistant Helicobacter pylo
within 393 mN, which is far less than the control group's 1351 mN.

ns on porcine samples. The average surface and depth resection errors measure 0.73±0.60mm and 1.89±0.54mm, respecti

physical intelligence at the small scale.

s for future medical applications such as a pulling mechanism to assist in guidewire insertion or agents delivering drugs.

tasks and achieve superior results than human operators, showing an average task completion time of 34.8±5.9s and a succe

icromanipulation tasks to provide accurate depth information.

ove that the refined S2worm-G robot is one of the best insect-scale robots for its size, mass, and mobility.
, biopsy, and tissue removal). Hence, potential strategies to raise the ceiling of the ranges are examined to extend the function
models designed for small object segmentation, achieving a significant increase up to 18.62% higher mean intersection over u

feedback control at the control frequency of 100 Hz without latency.

without needing to actuate both stages simultaneously.

performed the state-of-the-art top-down RP-R-CNN by 9.2% AP_vol^p, and the proposed automated tail morphology measurem
or of 1.6 ± 2.1 mm.

ation from grasping poses aids in mitigating the sim-to-real gap, leading to robust performance in challenging real-world experi

rategy to the real world, indicating a promising direction for active-perceptive MoMa.

tudent racing car showing an improvement of 16% in mean lap times across the same track between a common geometric con

optimization speed over the state of the art. Beyond simply a computational speed up, this ten-fold improvement may enable th

n simulation and with a physical automotive wire harness.

first principles physical model and a comparable neural network model trained without the physics regularization proposed.
he robustness of the proposed algorithm.

mimic human grasp actions with varying velocities. A final action adjustment based on tactile detection is performed during the n

m/ChrisChenYh/EfficientHuman.git.
muscle forces and kinematics with visual feedback at 30 Hz with a 110 ms delay. This method is promising to guide rehabilitati

nts of tracking and estimating a heavy-load cable prove that the proposed approach is robust under occlusion and complex ent
ds are validated through physical experiments.
t robots into everyday life and industry, by performing more complex, dynamic tasks.

ased within a 3D-printed mammary duct phantom. This work sets the stage for endoscope-based autonomous navigation of MA
d 4.91 mm average Chamfer distance error during evaluation. Supplemental materials, experiment data, and visualizations are
ansfer to realistic environments provides promise for applications in locomotor design and biomedical engineering.

apping of joint velocity and describe situations where this assumption is realistic (including underwater, aerial, and orbital enviro

hods (MPC, STL-solver), model-free and model-based RL methods in STL satisfaction rate, especially on tasks with complex ST

compared to the existing methods the CTopPRM improves the probability of finding all distinct paths by almost 20%, during sa

tic arm experiments. The results show that RBI-RRT* is more efficient in high-dimensional planning problems.

vailable at https://github.com/tedhuang96/nirrt_star.

analytical findings demonstrate that RETRO’s total regret can be upper-bounded by a function of the chosen time step size

pproach to strike a balance between guaranteed fast plan generation and the pursuit of optimization over time. We support our

cess, significantly improving the success rate compared to classical planners. EDMP performs comparably with SOTA deep-lea

outperforms the existing methods in planning time, path length as well as the number of unnecessary heading changes. Finally,

obot in multi-goal missions fully autonomously on rough terrain.

different directions and distances to arbitrary targets.

discussion, lay the foundation for addressing the technological and research challenges to significantly push the boundaries of

rm for testing advanced control algorithms such as event-triggered model predictive control. The associated code is open-sour

proposed design procedure can be seamlessly extended to other robotics and automation applications

derstand the intent of autonomous control to improve safety. The feasibility and effectiveness of this method have been validate

formance. Closed-loop tracking results show the proposed K-BMPC exhibits elevated tracking precision with the commendable
or initial and target shapes that differ from the training data.
Accompanying media can be found at: https://sites.google.com/view/adaptive-contact-implicit-mpc /home

oordinates, and we validate the approach experimentally on a seven-DoF robot.

CMP methods.

algorithms pose a serious threat to robotics, which should be taken into account in future research and development.
c trajectory to ensure that the robot has enough energy to return to the charging station when needed. The proposed eware filte

ofile for robot execution. Compared to directly applying the commonly used untethered trajectory deformation and tracking algor
red by 30%, compared with vanilla RRT on hard test cases. Given LiDAR input where vanilla RRT is not directly applicable, we

effectively avoid high-risk areas in unknown off-road environments.

dover success rate above 86.7% for all experiments by successfully combining the strengths of both fields of view of human po
eractive experience.

We conduct extensive experiments on the large-scale multi-view dataset InterHand2.6M and CAMInterHand achieves a subst

to five to achieve a more generalized low-dimensional classifier. This simplification both reduces the amount of training data re

perspicuity and efficiency.

tem can efficiently and stably construct large-scale dense maps in real-time with high accuracy and robustness.

eduling with dynamic subteam formation and voluntary waiting, and outperforms the previous and baseline methods.
nstraints. It is shown that the robots can switch between active and inactive roles dynamically online as required by the unknow
stacle density in physical deployment. Website: https://sites.google.com/view/obst-avoid-swarm-rl.
ur methods have better performance compared with some SOTA methods, showing strong robustness and remarkable adapta
red to 31% for state-of-the-art one-agent-at-a-time algorithms. In experiments involving MCTS, MLAT-R reached a terminal sta

e performance in various multi-agent reinforcement learning algorithms. We also implement OWC devices on real mobile robot

rithm. Further, we implemented our approach to achieve cohesive flocking, leader following and obstacle avoidance by a group

ctiveness of a team while allowing for intuitive mission specification and flexibility in the team composition.

Earth science missions.

s information-greedy baselines in terms of the Optimal Sub-Pattern Assignment (OSPA) metric for different numbers of targets

incentive is found to be at par with the original bigraph matching approach that used expert-specified heuristics, with the former

t methodology for multi-AGVs traffic-aware management. Moreover, we propose a coordination algorithm, which does not rely

ellaneous user-defined joint steady state distributions and incorporating other objectives, e.g., entropy maximization of individua
 an optimization approach that minimizes the number of timesteps to complete the structure, we observe a 100x reduction in co

set through mathematical transformations. Our submission outperformed that of other participants. Site: https://github.com/wq1

able at https://cloudgripper.org.

pecially, the occupancy map also makes difference in improving prediction precision of object orientation dramatically to averag
with +2.2%/+2.7% NDS/mAP improvements on NuScenes benchmark, and gain a comparable performance with 55.9%/45.7%

Based on these results, we propose a novel anomaly score by integrating these two scores, which we term as unknown objectn

best results on both benchmarks, with more significant improvements on smaller objects.

s computational power but also achieves significant performance gains compared to the baseline.
 model on an edge-computing device. We will make our code repository publicly available.

rease of 28.2% when compared with other state-of-the-art UOIS methods.

.18 m and 2.7 degrees of mean frame alignment error in our most challenging simulation scenario. In addition, we conducted h

i-task loss, and the implementation of a post-processing stage for uncertainty refinement. Our experimental results demonstrate

rcycles. Through testing actual low-resolution data, the accuracy of our method is verified to improve a lot than the state-of-the

cant amount of detection errors caused by rotations of BEV and objects inside a BEV, thereby further enhancing the performan

bits superior performance and expanded robustness in the 3D object detection task. The code will be publicly available at https

ved while maintaining an economical annotation budget of 1.44% of the training data. We foresee further research exploring the

blished information measures, our approach achieves comparable performance to training on the entire SemanticKITTI datase
multi-disturbed conditions. Specifically, when the target moves at a speed of 1.0 m/s, the UGV can maintain a tracking error wit

ming algorithms.
 of dynamic uncertainties and unknown external disturbances. The simulation results reveal the efficiency and practicality of the

ly improve the computation time and refer to the modified labeling algorithm as GPLA*. Computational results are presented to
or the three methods respectively.

h coordination method. We also present preliminary results that show how informing each coordination method about human pr

uadrotor swarm of up to 12 quadrotors in obstacle-rich environments. The code, supplementary materials, and videos are relea

endencies at different spatial scales. Our simulation experiments demonstrate that ALPHA outperforms both globally-guided M
ts and one outdoor experiment with three real Quadcopters.

hile minimizing the cable angular motion.

ach with Sim2Real adaptation and the MSC-RCNN architecture achieves a remarkable 7.6% performance improvement in insta
ld experiments.

alculating and tracking psi_d to perceive and avoid two static obstacles with an average computation time of 371 mus.
ted and real-world data. The software is released open-source to benefit the robotics community.

https: //hkust-aerial-robotics.github.io/FC-Planner.

of forest areas in multi-UAV setups up to 30% faster than the state of the art.

l world, without prior knowledge of the environment.

hout any external infrastructure or prior knowledge of the environment. We verify the efficacy of the perceptive safety filter in re

at our method can achieve the lowest position error (0.11m) and a comparable velocity error (0.23m/s) across the benchmarkin

ce the computational demand.

ocess and empowers state-of-the-art 6-DoF robotic grasping algorithms to exhibit markedly improved accuracy.

HTY and GEM-C schema's potential for a wide range of long-term and complex robotic manipulation applications.

results on synthetic data suggest Residual-NeRF outperforms the baselines with a 46.1% lower RMSE and a 29.5% lower MA
superior performance over state-of-the-art algorithms for the class of wiry structures and the potential to be easily used for sub

system achieved an 80% success rate for autonomous opening of unknown articulated objects.
a distance to slip, the safety margin is a powerful metric for regulating grasp forces. As a testament of this effectiveness, we sh
h positive and negative matches between pegs and holes. Robot experiments further validate the practical application of this a
 sing solution for enhancing robotic manipulation in liquid-related tasks.
n intervention.

further validated in a physical experiment. For videos, please visit https://sites.google.com/view/multi-view-rearr.

ring them with an existing heuristic policy, as well as several baseline attention architectures. Then, we validate the performanc

nsive experimentation on a representative set of the prediction models using the large scale Argoverse dataset. We show that

Sawyer, and they responded positively to robot-assisted skill coaching.

s were conducted and the results showed favorable performance, indicating the effectiveness of this approach.

w a higher significant p
s for the same intervention. The results of this study can provide insights into the optimal design of leg exoskeleton-based reha
uracy and reliability of PC are also found to be consistent regardless of the variations in our experimental procedure design.

ated through human studies. The experiments show that the robot correctly recognizes and responds to human input, commun

n operator training and the appropriate selection of cobot’s control parameters.

ion, without affecting the real-world mechatronic system or hindering the subject’s safety. The proposed tools can optimize
ility in the optimized parameters highlights the importance of personalized control settings when walking with an assistive devic
planning for agents with merged sub-maps to further improve exploration efficiency. Our experiments show that our approach i

evaluate our approach in simulated environments and demonstrate its efficacy in handling multi-modal belief hypotheses with a
nline planning.

and-action space) from the executed ones. Simulation and experimental results show that the proposed approach increases th

ul and safe navigation in a number of case studies with real-world pedestrian data.

ps, we improve the state estimation by correctly selecting the viewpoints and considering the measurement noise correctly whe

ess of the proposed framework in a real-world plug insertion task in the presence of port position uncertainty and an assembly t

trajectory execution time while enabling robust collision avoidance in complex environments. Our software is available on GitHu
e present extensive experiments and compare our RL approaches against their classical counterparts.

onments to validate the performance of the proposed algorithm. The C++ implementation of the proposed algorithm is released

nt state of the system to sequence movement decisions that optimize short-term coalition formation and long-term task schedul

great potential to apply A-mode ultrasound in orthopedic surgery from safe, convenient, and efficient perspectives.

riments with a powered knee-ankle prosthesis show that this simpler, unified model produces competitive results when compar

nd synchronization. This study highlights the importance of monitoring coordination changes to fully understand the impact of ex
esults show that the neuroprosthesis hand control scheme based on MU spike trains has the capacity of generating accurate a
man intention. To the best of our knowledge, this is the first cooperative push recovery framework for the lower-limb human exo
while the torque magnitude varied; participants may be much more sensitive to torque timing than torque magnitude, and there

he mass of the object being lifted, namely reducing by 45% and 54% respectively the Biceps activity while lifting the MW and H

captured in an indoor corridor environment with precise localization and ground truth mapping information.

gorithm and has a property to put samples into around one optimal sample. As a result, we can achieve robust reactive navigat

n several real-world applications, showcasing its ability to deal with cluttered scenes, viewpoint mismatch, new instances of cat

error of the torque to less than 60% in relation to pneumatic pressure alone. This demonstrated the ability to generate consisten
been successfully applied to a 4-DOF parallel robot for lower-limb rehabilitation.

ovide assistance as needed and prevent the early termination of physical therapy due to muscle fatigue.
ceived the SWS to be useful and easy to use. They also reported mild to moderate perceived exertion (RPE).

, in comparison to six different state-of-the-art Safe RL methods. The results demonstrate the significant superiority of the prop
the latency of DMC tasks by as much as 27%.

periments, demonstrating a significant reduction in vertical mismatches and an improvement in the overall visibility of automotiv
d AI. The experimental results demonstrate the effectiveness of our method in successfully reconstructing all information from t

ain event detectors.

e the efficacy of MOSAIC's unified representations, showing competitive performance in category recognition through a simple

e giving lower weights to those with weaker representation abilities. Such blending operation automatically modulates the multip

n of the input point cloud (e.g. 75%) will still yield a nontrivial self-supervisory performance, which enables us to achieve 3-4 tim

o/SKTHang/.

light the accuracy of SlotGNN in downstream robotic tasks, including challenging multi-object rearrangement and long-horizon

thod reduces the execution time of visual odometry by 4.3% and also reduces reprojection errors. We demonstrate the need to

al SDF grid. Videos and code are available at: https://samsunglabs.github.io/HIO-SDF-project-page/

odical strategy, selecting industrial sites characterized by ambient machinery noise to mirror real-world scenarios. This approac
requirement. The dataset split and code is available under www.github.com/evinpinar/supergb-d.

nto various architectures and demonstrates state-of-the-art (SoTA) performance on both synthetic and real weather datasets. S
s of communication bandwidth for connected autonomous driving. In addition, our approach achieves over 30% improvements
scenarios and demonstrate its effectiveness when combined with state-of-the-art scene flow methods to establish dense corre

rface from a sensor mounted on the end effector of a robot arm.

e-of-the-art performance for action recognition on both NTU RGB+D 60 dataset and NTU RGB+D 120 dataset.
 at rely on re-sampling, re-weighting, and pre-training.

s today, and find that OriTrack demonstrates 23% increased energy production. This result suggests OriTrack could be used as

in the grasping mode.

neral and can be applied to other robot arms.

world tasks for table-top manipulation and diverse in-the-wild manipulation. https://homangab.github.io/hopman/

e problem for an autonomous mobile robot and spacecraft rendezvous problem.

e framework with various safety thresholds and demonstrate the feasibility of result trajectories on a surgical robot. The effects

' knowledge, are first-of-its-kind revealed in real-world testing of legged robots.

diction error-based detection and ensembling, thereby enhancing the overall safety and robustness of autonomous systems.

ated our approach on a simulated robot manipulator, empirically demonstrating that it leads to less constraint violations than sta

robot safety while adhering to driver intention.

set and the real-world ScanObjectNN dataset. The results show the superior performance of HEGN in 9DoF point cloud registra
e a principled objective that more accurately balances training the classification and density estimation networks. We demonstr
has superior numerical stability, and remarkably, achieves efficient and robust performance on 32-bit single-precision float at a

porally extended exploration for reinforcement learning. We extensively evaluate UVD on both simulation and real-world tasks,
ds, we achieved a maximum improvement of 12.38% in prediction accuracy, demonstrating that our method provides the best e
peedup. Kappa statistic tests are performed to confirm the consistency between our system and human experts. It is expected
en degree-of-freedom alignment. Experimental results show the advantage of our approach over other matching algorithms and

with lower segmentation accuracy.

r all the investigated scenarios. We also evaluate the embeddings to assess the quality of the fully unsupervised approach and

ed skeletonization technique in tracking different organs of the plant by finding good quality correspondences. Experiments are

mode transition and shows the best transient performance in overshoot minimization among three controllers.
er to operate with smaller batch sizes and achieve real-time performance even on embedded hardware devices like NVIDIA Jet

icultural environments while presenting a metrically accurate and human-understandable representation.

structive picking of sweet pepper, with an average picking time of about 15 seconds.

ltural standards.

a sampling-based NBV planner, while taking ten times less computation and generating 28% more efficient trajectories.

egmentation, and instance segmentation on the Sugar Beets 2016 and our potato dataset demonstrate that it achieves state-of

ntation errors and deficiencies in the final height adjustment algorithm for grasping logs. We also present detailed timing results

gnated goal.

e actively sampled during the training phase. We empirically show that our approach outperforms existing meta-learning and do
orrelation of findings from both IE and GPR sensors give a comprehensive evaluation of concrete structures that will benefit the
eamless integration of our proposed approach compared to conventional time-triggered MPC strategies.

on actions to ensure the robot's compliance with safety regulations, paving the way for safer and more widespread adoption of r

hecker in a dynamic and collaborative environment.

esults even when hands and arms were heavily occluded, adjusting to different variations.

s our attack, and adopting our method as the defense. We find that RAST can defend against 60% of untargeted attacks and 2
atively and qualitatively with experimental setpoint baselines. The methodology introduced shows promise for enhancing UHRI

pproximately 16 %.
explored area and response time trade-off depending on the type of robot: wheeled vs quadruped.
al graph in cases of obstructions. As such, the exploration behavior is directing the robot to follow another route in the global re

aversability estimation methods and achieves comparable performances with supervised learning methods trained on manually
duplication to increase the number of distant foreground object instances and mixes the radial duplication with another point clo
and indoor environments. To foster future research, we make the code and trained models publicly available at http://spino.cs.u

SemanticKITTI and NuScenes benchmarks.

erence with different depth sensors. Code and data will be available upon paper acceptance.
onal frontier-based methods, our findings indicate that the proposed approach leads to faster exploration and requires less com
s on our project page: url{https://uark-aicv.github.io/OpenFusion}
emanticKITTI test with a score of 68.4 LSTQ.
ation. We plan to provide the code and pre-trained models in case of paper acceptance.
p from Habitat to ROS. The migration evaluations show that HSPNav can generalize to realistic settings well and achieve prom
f the classical 3D scene graph concept and aims to enable high-level reasoning for task planning and optimization of a variety o

hensive experimental results on the KITTI road and semantics datasets validate the effectiveness of our proposed architecture

sting HD map construction methods. The satellite maps and our proposed module notably enhance their performance in both H

cenarios using a photorealistic simulator. We release our code at http://curb.cs.uni-freiburg.de.

demonstrate mmEMP's effectiveness.

, which contribute up to 30% discrepancy in terms of the average precision.
-day detection.In addition, to compensate for the lack of paired RGB-Event datasets for all-day training and evaluation, we prop
usion-based methods on SemanticKITTI and nuScenes datasets. The code will be released at https://github.com/Jun-CEN/CMD
utperforms state-of-the-art multi-modal fusion and distillation models in scenarios with missing modalities when using only IR da

ed by our method can lead to better path planning for autonomous vehicles. Our code and self-recorded dataset have been rel

and public datasets. Results confirm the potential of the GICI system to provide continuous precise navigation solutions in a wid
und. Finally, we fuse the results from the two reasoning stages to extract the final moving object region. This joint spatio-tempo

Meanwhile, it utilizes an output head to generate a fused mask, which is iteratively updated for learning consensual spatial prior

periments recording the motion of a mechanical lidar subject to angular velocities four times higher than other similar datasets a
 get microphone array, which can span diverse array types. As a result, our method demonstrates robust performance across tw

s, medical students, and patients. The functionalities of ARIS 1.0 were validated in an emulated indoor environment to evaluate

nment of interest.

ing, it is written in C++ and comes with a Python interface. We demonstrate our system’s ability to reproduce complex real

omes with demonstrations.

ation of SLAM, odometry and loop closure algorithms, and we provide several state-of-the-art baselines. We also include 40k se

LAM. All our software and datasets are freely available for further research.

backend, along with the VIO+Backend example, with the aim of contributing to the betterment of the community.
iverse public datasets demonstrate the advantages of our system compared to other state-ofthe-art methods. Our code is avail

ource implementation of the proposed method.

al complexity of correspondence matching. Extensive experiments conducted on three public datasets demonstrate the superio
zing the objects and points-based covisibility graphs in our visual object mapping process. Experiments show that VOOM outpe
ure sharing module alleviates the impact of object boundary in depth estimation, and the designed multi-sharing module focuse

utperforms the accuracy of the second-best method by a margin of 10.67%, while extending the robot situational awareness by

-grade IMU sensor using fixed-lag smoothing optimization. Finally, we evaluate our approach using real-world 3D motion data.

is applicable for both non-repetitive scanning LiDARs and traditional scanning LiDAR.

rmance of the proposed trade-off optimization algorithm, we apply our approach to CSAIL, Intel, and Manhattan datasets.

n indoor environment while improving robot localization.

. Lastly, we demonstrate the system’s potential for planning.

s. Empirical benchmarks, conducted on a new dataset of simulated indoor environments, substantiate the efficacy and efficienc

es, respectively.
 he experimental results show that the approach can improve the robustness of the navigation policy. The source code for the tr
an average 15% improvement on the robot experiments over baselines that do not use active perception. The dataset, code, a

cycle. The soft gripper based on the actuator can grasp repeatedly or laterally. The gripper can grasp soft objects such as a pie

soft grippers for delicate fruit harvesting, and indicates how to increase the impact of robotics in agriculture.

lling locomotion attained between 15.6 and 19.4 mm/s (0.12-0.15 BL/s), and crawling 7.8 mm/s (0.06 BL/s). The rolling gait wa

obot-object interaction.
shes the unlocking task. The ECR has great application potential for on-orbit service.

ccessfully picked, with 0% of them sustaining damage. The ripeness probe classified 56 ripe and 11 near-ripe berries, with 89%
mm/s. And this design principle can be extended to the research and design of various flexible devices and soft robots.

electromagnetic surfaces but for any application benefitting from light-controlled actuation.

he trade-off between the actuation distance and design constraint. Finally, three optimized designs are fabricated and tested, de

es a large workspace and stable motion capabilities.

echanism. It also presents the design guidelines and fabrication details for the mechanism, as well as a demonstration of its sw
each other, allowing the robot to exhibit higher forces and tighter curvatures than previous vine robot architectures. This work a
n exploitable in a wide range of applications, from medical to rescue missions, where actions behind obstacles in highly confine

nce within 3D space, laying a robust foundation for future implementation of tactile localization on UAVs. The design files, datas

he softness and compliance while being proprioceptive relying only on embedded sensing and not on external systems such as

ata collection and training taking only 70 minutes.

ble of delivering a relatively high force to the target objects.

he reactor to trigger a biochemical reaction for powering a fluidic soft actuator. This work opens up the possibility to power and

github.com/ika-rwth-aachen/dorotos.

d action at runtime.

navigation tasks that would be otherwise not practical with the standard components.

nsumption by the MUs. Reward shaping and surrogate models are used to respectively guide and speed up the RL process. S

an effective data filtering method to improve the quality of training data, thus enhancing model performance. A multifaceted bas
real-world testing. This technique will be used for the local mapping and collision avoidance on NASA's CADRE lunar rovers.
and analyzes the kinematics and dynamics of the VWDER. The experimental results verified the performance of the VWDER, in

lation time, respectively.

images with pose annotations of a surgical needle that was used to evaluate a state-of-the-art pose estimation network. The tr
orms. The code is available at: https://github.com/SamuelSchmidgall/SurgicalGym.
rformance and task versatility compared to existing methods. The learned trajectories can be deployed on the da Vinci Researc

https://github.com/wrld/Ada-Tracker.
ment of simulation-based deformable tissue manipulation and holds potential for improving surgical autonomy.

on, MP recognition models perform best if trained on similar tasks and/or tasks from the same dataset.

re, we demonstrate path following capability in hardware with a line path with resulting errors of 4.37 +/- 2.39 mm or 5.61 +/- 3.

stimation algorithm is implemented for real-time visual feedback to the operator with a low latency of 400 ms. Experiments sho

ulator interface to verify performance and robustness of the algorithm before deployment to a Crazyflie hardware experiment in
n baseline. An appendix comparing the framework to bandit methods and proposing a corresponding tree policy version is foun
which takes into account all the information gathered along the robot's movement. The proposed method is evaluated in both s

the significance of leveraging historical multi-modal data, particularly in scenarios where the agent encounters intermediate fai

from research institutes worldwide is needed to create a comprehensive human injury avoidance database for any pHRI scena

d accuracy of detection; and introduce a Bayesian-guided evolutionary strategy(BGE) for efficient task allocation, which can utili
g-wheel quadruped with a humanoid dual-arm upper body.

task board for benchmarking, and in two illustrative bimanual aerial manipulation tasks.
e proposed framework can achieve up to 50% communication load reduction with improved task completion probability.

using an offline estimated OBM improves the accuracy of the force feedback and is reliable against packet loss and short temp

ck-KF and VF showed superior interaction and tracking ability, respectively. Moreover, these two properties are in trade-off rela
ttps://github.com/JaronTHU/DynamicAggregation.
-UDA benchmark. Code will be available at https://github.com/AronCao49/MoPA.
nce of the transferred policies but also reduces alignment errors significantly compared to the baselines.

o a shared geometric embedding space and decreases data distribution distance between two domains. To verify the effectiven

and underwater (3D) robots, both requiring time-consuming physical simulations. Our results demonstrate a 50% decrease in le

e for the development of more advanced cooking robot systems.

cy-aware HVAC energy saving, living service, search& rescue, and law enforcement.
deployment of our prototypes, including the Synchronous Team-Robot Van (SVan) with air, land, and water guardians and Co-R

r accurate depth estimation. The predicted 3D human meshes are then tracked and refined in physical 3D space with real-time

posed method is able to achieve sub-meter level localization accuracy across both indoor and outdoor environments.

ntrol. The project aims to eliminate X-ray exposure risks for operators, and increase the level of autonomy in percutaneous proc

e best heuristic tuning.

his unique dynamic configuration of BALLU allows for energy-efficient turning, where the cost of spinning becomes favorable co
obots, ultimately streamlining the mapping process in ISRU scenarios—a pivotal advancement for the emerging space econom

h smaller and lighter than other limbed climbing robots, and only costs $1250 per unit. It can climb at 7.3 cm/s, making it one of
m to lead more fulfilling lives.
ble way for human-machine interaction in the deep sea in the future.
ories with precision. This work represents a significant leap in autonomous tensegrity robot capabilities, offering exceptional ma
oad operation, overhead support, assisted delivery and cable installation. The capability of the proposed SRL to perform succes
medium or high stiffness levels can effectively reduce both muscles' activity.

sed method outperforms existing NRSfM methods.

non-stationary (moving) cameras, validating its performance in both structured and unstructured scenarios. The results achieve

set are available for future comparisons at: https://github.com/Lio320/AgriSORT

system, as well as the easy setup of UWB anchors. Experimental results show that the proposed method outperforms previou
ethod, in terms of absolute trajectory error and relative rotation error, while maintaining competitive computational speed. Notab

experiments on real hardware.

predicting future events and seamlessly integrating the vehicle's kinematic model for enhanced decision-making.

LED light quantity, recognition distance, and angle, external lighting conditions, and camera or marker movement on accuracy

date the effectiveness of our proposed navigation strategy. More details are accessible via our project website https://sites.goo

the planner is trained purely in simulation, enabling a highly scalable training data generation. Experimental results demonstra
ation about the environment or localization. Moreover, a robot using UIVNav compared to complete coverage method surveys
pability to extrapolate preferences and generalize to novel terrains and challenging lighting conditions. Compared to baseline ap
pared to state-of-the-art methods.

coring the significance of this approach in real-world applications.

hallenges of real-world surgical scenarios. Our code is publicly accessible at github.com/longbai1006/OSSAR.

 olute improvement in mAP for 6D pose estimation while running in near real-time at 5 Hz. Project page: https://fsd6d.github.io/
e available at url{https://github.com/yuyangpoi/FE-DeTr}.

visible images.
produce human-level cuboid annotation on the image without needing manual adjustment, and can accelerate the monocular-3

mentations as a condition to guide a fine-tuned latent diffusion model. This finetuning process ensures both view and style cons

a systems within SLAM. Consequently, this fusion method enjoys the high-resolution and accuracy benefits of advanced LiDAR

g the number of required views for high-quality reconstructions in dataset refinement and NBV planning.
ag maintains an average processing time of 10ms per frame on a standard laptop, effectively meeting the real-time demands of
clock devices. We will release code related to this work.

uations. Our code is publicly available at https://github.com/MJ-Kor/OP-GCN.

sure images that are combined to create a haze-free, visually pleasing, and higher-quality final image. The method effectively r
ance using unlabeled data. The code is available at https://github.com/Junyu-Z/Semi-BEVseg.

ectively augments the performance of semantic segmentation tasks. In addition, we propose a novel paradigm for representing

model, which avoids error accumulation in target spike data. To verify the effectiveness of BiCross, we conduct extensive expe
racy is achieved.
era position optimisation framework.
ng a strong competitiveness compared to other outstanding methods. Consequently, we can offer clinicians automated, efficien

ate-of-the-art approaches.

hat RIDE sets a new state-of-the-art performance on matching and relative pose estimation tasks and scores competitively on s
hen put forth a novel data processing method that transforms complex LLM embeddings into logits compatible with our CL fram
ow a user to initiate a grasp on an object by specifying a grasping point in a 2D image.
twelve different food items.

y to noise compared to velocity controllers.

e is available at https://github.com/bearyi26/DCPT.

to forecast the object positions. Finally, both the refined feature embeddings and the predicted positions are integrated to enha

s score of the propagated boxes as well as forward-backward optical flow consistency. We re-identify objects across occlusions
tate of the target with a small densely connected convolution network. Deep supervision is added to each stage to further boos
video. In the second stage, we fine-tune the model to reproduce these outputs, while also applying augmentations on the input

materials will be available at https://github.com/TsingWei/LiteTrack.

SO on public datasets and real-world experiments, and our system outperforms state-of-the-art algorithms. The code is publish
e and also show the effectiveness of online GNSS-IMU extrinsic calibration. The experimental validation on the public datasets

ng the level of normal people in jogging. Our MAT dataset is publicly available from https://github.com/snoopy-l/MAT

ubject and reducing mean muscle activity by ∼31%. Further results highlight air regeneration's potential for energy saving in

proposed control scheme.

he exploration process and acquire more satisfying optimized results for naive wearers.
ates knee flexion/extension while accommodating the knee joint’s center of rotation in both the Sagittal and Coronal planes
ptimization is a novelty for wheeled-legged robots.

ments on a car-like robot demonstrate the effectiveness of the proposed control scheme.
n a stable standing configuration from falls with various horizontal velocities and different angular perturbations.

ual feedback for long-distance planning.

lization architectures while having a negligible effect on their runtime.

nd TLESS. In comparison to MegaPose, we improve the Average Recall on all three datasets and compared to OSOP we impr

dangerous oscillatory swing-up motions.

elay UAV in order to maintain LOS links between the mission UAV and the ground station

ral advancements for the entire robotics community.

ore precise and rapid adaptation to multi-source disturbances than other adaptive controllers.
to verify the system's effectiveness and robustness.
us robot navigation. The related codes, data, and models will be made publicly available at https://github.com/KangchengLiu to

ethod. Our evaluation spanned 48 instances across 16 distinct categories, revealing that our framework not only outperforms tra
bility and generalizability of working with natural language instructions.
d LLM as a planner. Extensive experiments demonstrate that our method exhibits efficient and adaptive ability in complicated m

MDP with parametric beliefs, as well as nonparametric beliefs represented by particles.

feedback. Our observations lead to a prompt recipe we refer to as PromptBook that combines: example-based, instruction-bas

ing and expanding task and environment development, and by facilitating the discovery of reinforcement-learned behaviors thro

ks relying on LTM information. Notably, similarity-based forgetting methods outperform frequency- and time-based approaches

es data efficiency in RoA estimation.

hub.io/Autonomous-Vehicle-Pursuit/.

nts, highlighting its potential for practical utility.

such as wheel speed sensor odometry while outperforming it in terms of drift-impairment.

e perception and trajectory prediction empowered entirely by passive thermal sensing technology, heralding a new era where a

babilistic evaluations are used. This demonstrates effective prediction for supporting a pla

sical robots show that the hybrid planner can achieve better social compliance compared to using either the geometric or learnin

human intervention. Code is available at https://github.com/KubeEdge4Robotics/ScaleNav.

nt confirm the superiority of our method in terms of learning robustness and driving performance. Videos and source code are p

of 0.61 fps compared to the lidar data publishing rate, and the generated MTraMap can clearly delineate the multi-directional tra

ilities compared to conventional approaches. Finally, we conduct real-world experiments to demonstrate our feasibility.

show that the LiDAR data brings improvement in the motion forecasting task. We hope that WOMD-LiDAR will provide new opp

also demonstrate in simulation experiments that our method outperforms prior works that use different RL algorithms or differe

rovided by the model proves to be an appropriate approach for aggressive lateral maneuvers of bird-size ornithopters.

ther hydrodynamic interactions among schooling fish.

elays of 90 ms, and tracking performance for successful sim-to-real transfer for navigation tasks.

bilities to incorporate microelectronics, enabling autonomous flight functionality.

al comparisons among wheel-wheel, tail-wheel, and 2-DoF tail robots s
ed in the modules.

at mias.group/StereoCalibrator.

d cameras.
rameters, LCCRAFT achieves competitive performance as compared to the state-of-the-art method, which entails 69M parame

r, respectively. The tool alignment experiments after using GBEC achieve a mean accuracy around 0.2mm in Euclidean distanc

tness, accuracy, and generalization capabilities.

demonstrate our method is highly accurate and robust.

dge of the environment. The accomplishments of VLFM underscore the promising potential of vision-language models in advan
collision rates, despite utilizing smaller models than state-of-the-art approaches.
ed state space representation methods to address the associated challenges. Curriculum learning is employed to expedite the t

e proposed algorithm publicly available.

5% increase in AU-ROC compared to the state-of-the-art navigation methods.

ase our dataset and provide further details and visualizations of our results at https://iliad.stanford.edu/pg-vlm/.

st of only a 9% performance drop, making it the only model among current literature capable of real-time operation while mainta

bots and image splicing technology.

ce levels, and to sliding manipulation planning by clarifying

-driving, and corrupted Cityscapes datasets.

ss and lower buffer size requirement compared to other methods.

orld. Project material will become available at the webpage https://gtziafas.github.io/LRLL_project/

e exploration efficiency through collaborative work.

ture in cleaning operations.

OI and HOI-based features in the context of Action Recognition. Notably, our experimental results show th
onmental settings, making them a valuable resource for SLAM benchmark development.

cognitive intervention rate 46% lower than the zero-shot state of the art visual language model (VLM) baseline and is able to gu

visual navigation dataset on our project page.

cal structures and extract rich local information with linear computational complexity. Experiments demonstrate that our method

n shifts, which are fine-tuned to avoid the catastrophic forgetting problem. In addition, since CTTA is a temporal task, we introd

in any stage of this process.

us SoA work and an endurance performance up to the entire battery lifetime of 4min.

forms the state-of-the-art render&compare method RoboPose and achieves higher inference speed. Furthermore, the tests acc

anipulators.
 coding methods through empirical validation in downstream tasks without sacrificing point cloud compression performance.

safety rate compared to the baselines, under unseen disturbances in real-world experiments.

lementary material, code, and videos can be found at https://xinyiz093

f autonomous grasp.

-shot overlap success rate on previously unseen target shapes.

site at: http://sites.google.com/view/gamma-articulation
h exhibits its usability in real robot experiments without any refinement steps, even when only trained on a synthetic dataset, du
publicly available at: https://erdiphd.github.io/HER_force/

on efficiency. Our experimental results in both simulated and real-world environments showcase the state-of-the-art effectivenes
ons from the cohort. We validate the TCC framework on Cityscapes and Pascal VOC 2012 datasets, which outperforms existin
hub.com/zy1296/CrackNex.
n of data-sensitive components. Furthermore, we make use of a novel hierarchical training scheme to obtain the optimal model

s available online at: https://youtu.be/q1l4BJVDbRw

ate increase of over 60%. Additionally, when compared to current visual pre-training methods, our success rate exceeds them b

and FastMaxSum.
 ding to three attitude angles, which verify the performance of employing MPC to control an autonomous underwater robot with

outperforms other approaches in cluttered and unstructured environments and demonstrates an enhanced capacity for autonom
uiring any fine-tuning.

ate the greatly improved utility, accuracy, and robustness of multi-session mapping over the comparative methods.

eover, for small problem instances where the exact solution can be found, we demonstrate that our proposed heuristic has a co

mental results.

LURE dataset [1] with 99 new multimodal manipulation recordings and annotate them with

for the upkeep of electricity networks.

ing the capabilities of ULRs and lays the foundation for future underwater intervention missions in various scientific and industri

e autonomous inspection has been realized in unknown and dynamic tunnel environments.

design instead of efficient robot mass distribution or motion control.

ore the efficiency of our model, achieving safe landings with minimal errors (≤ 7 cm) in winds up to 5 Beaufort (8.1 m/s). In e
s, including significant savings on the effort and operational cost of deploying the ASVs.

st to ViTac's 0.12mm and 0.15N. Results indicate that ViTacTip outperforms single-modality sensors.

CH could further empower existing Vision-based Tactile Sensors (VBTSs) in replication and deployment.

th different inclinations. Finally, we used the relation between specimen compliance and the
ipulation. Code will be openly released at https://github.com/ac-93/tact

al automation, enhancing operational efficiency and strengthening situation awareness and decision-making capabilities in verti

edded force information within the tactile images.

ve lookup table method in the GelSight sensors. The results show that the proposed method enhances both reconstruction accu
m the proposed model show a 28.3 % decrease in marker positional error and a 93.5 % decrease in the image similarity m

upper-body dressing.
d using the proposed method, and the cutting process was made easier compared to that of convention

e solution space, generating better solutions and improving the performance of multi-agent route planning with uncertain travel

y of relationship detection in complex scenes and exhibits robust generalization capabilities in real-world applications.

with varying numbers of objects. Videos and Code at icgraspnet.github.io

eployments. Consequently, our method reduces the human workload needed over time while ensuring reliable task execution.
n applied to a different robotic hand and task. It also exhibits excellent performance on real robots with a success rate of 70%.
n of these ingredients in four complex visuo-motor manipulation problems in both simulation and the real world. To the best of o

h. The results demonstrate that our proposed method exhibits excellent performance in reconstructing finger's joint configuratio

o the ideal single-point-focused target vibration pattern. Four point-wise patterns focused on the top, bottom, left, and right part

actile scene.

nsparency compared to the conventional zero-torque controller, using force readings for position control and torque measurem

s. Statistical support is provided along with experimental outcomes.

ed for 53,026 iterations of HIL dataset enhancement, leading to the removal of 2,884 images and the incorporation of ground tr

os of the experiments can be found at https://sites.google.com/view/rl-sequel

nment using a simulated human expert. We further show success transferring this approach to a new domain with a real human

eduction among the subjects are also reported as 66.4%, and 67.11% for 0.5 kg, and 1 kg, respectively.

w to apply these modifications to other RL policies and demonstrate that key high-level behaviors of our approach transfer to a p
ous modalities like speech, touchscreen and gestures. We provide crucial XAI design insights.

ncept we introduce in this paper—IJP efficiently searches over diverse motion plans. Closed-
enarios, where the robot goes through gaps safely, finds trajectories quickly, passes pedestrians politely.

/EEGTrust.

pout and is recovered using the proposed architecture and the training paradigm.

sk to simulate the visually impaired. They were instructed to run by following the perc

redictions with errors over the time window growing as 2.3 ± 3.4 mm at 50 ms, 2.4 ± 4.4 mm at 100 ms, and 5.5 ± 6.7 mm.
ging existing point-based value iteration algorithms with essential modifications. Human subject experiments with 21 participant

pants' effort also reveals a significantly lower perception of the workload.

he provision of a comprehensive comparison with competing approaches, thereby significantly enhancing the robustness and a

example is presented to show the versatility of the proposed framework.

ious advantages of our method over single-round self-thinking.

om/view/rl-polite
g results as an effective method for reducing sim-to-real gaps.

tence remained stable across conditions, the recalled, post-experiment ratings of competence and trust were significantly lowe
n motion prediction. We demonstrate the superior prediction performance of LaCE-LHMP through benchmark comparisons with
a provided by the users (experimenters), and use the learned knowledge to perform object fetching tasks.

pants in both VR groups compared to participants in the simple task group.

dback to human users and allow local processing with low energy consumption, such as cognitive prosth

our method.

ucaszyp/MonoOcc.
place amongst LiDAR-only trackers that use CenterPoint detections. The open-source code for reproducing

erscores the pressing need to address security implications within multi-agent perception systems, thereby underscoring a criti
ems not only from differences in precision but also from the nature of imperfect inputs provided by perception modules, and tha

dataset demonstrate the superior performance-bandwidth trade-off of CenterCoop, which achieves the state-of
ation costs compared with the state-of-the-art method in V2V4Real.
ks in robotic applications.

ed reward function. We demonstrate the effectiveness of Rank2Reward at learning behaviors from raw video on a number of ta

models, weights, assessment tool, and dataset, for easy replication and extension.
 perturbation.

en selected tactile features are reduced.

reate a specific instantiation for the problem of dexterous, in-hand reorientation of a cube by a highly actuated robot hand.

e scalability and generalization under varying city scales, vehicle numbers, and task numbers.

ce by only taking observations from a force/torque sensor attached to the arm's wrist, exerting a prescribed force, and handling

we present the design and software to replicate the electronic task board to promote task-centric benchmarking.
d on the original PDE model.
 o identify different motion patterns of courteous behaviors according to the scenarios.
behaviors compared to humans. Furthermore, the digital twin learning framework can identi

oU) by 7.15% compared with state-of-the-art schemes.

on is available. Lastly, we deliberately degraded the state to observe how the framework performs under non-ideal data collecti

sing precise information on environments, including ring, figure eight, intersection, merge, and bottleneck. In certain scenarios,

u.be/Wjla071BPBA

erize the number of taxis m that is sufficient for IA to be stable; 2) derive a necessary condition on m to maintain stability for IA
utilization of the vast pre-collected scenario library for closed-loop driving policy optimization but also facilitates AV improvemen

g the versatility of our framework.

del can obtain a superior performance to other methods with clear explainability. The code is available in https://github.com/JW

endent estimation method and its availability in the upper-limb rehabilitation system.
uitive nature and potential efficiency gains to be leveraged by introducing the proposed HMIP into HRC contexts. The opensour

ur method outperforms SOFA when the deformation is highly nonlinear.

ctive-joint space method.

t to reduce the occurrences of undesirable human-robot interactions by 29.2% on average.

rate the state-of-the-art adaptability of EditMP.

ment period.

tion gives an average score over 40.

te-of-the-art methods on various continuous and discrete control benchmarks. Remarkably, RLTSC even outperforms RL meth

tion and model-free methods in terms of sample efficiency and task performance of sim-to-real transfer.

alized baseline and four modular policy baselines. Additionally, we show enhanced robustness against module corruption and s
chieves state-of-the-art (SOTA) performance in both UIE and UOD tasks, but also shows superior adaptability when applied to d

s intrinsic bonuses to encourage the collection of novel states for training. We evaluate our REPlan in three vision-

onal videos: https://kl-research.github.io/dyncatch.

s over 4 times more efficient than the heuristic policy in both in-distribution and generalization settings, suggesting a scalable sk

o a mobile manipulation task in a much larger setup. Project website: https://equivact.github.io

S module consistently improves performances over the use of Lidar only or Lidar and encoded RGB features.

le at http://ralf.cs.uni-freiburg.de.
e available at https://github.com/NeuCV-IRMI/VPE-SLAM.
bitat dataset augmented with electromagnetic radiation simulation for wireless signals. It is shown that the PIRL significantly ou

mbing boxes with a quadruped to dexterous manipulation.

time and computational efficiency, and also verifies the generalization of our method to different numbers of objects. In addition
policies and solve decision-making problems in challenging safety-critical scenarios.

ations between agents and graph nodes in a coarse-to-fine manner for effective global goal selection. Extensi
sor measurements and existing smoothing methods. Quantitative analysis is performed in various natural e

n while bearing an unexpected load of 10 kg, or 83% of its body mass. We further demonstrate the generalizability and robustn

between local goals, enabling long-term navigation towards distant goals. In simulation, our hierarchical approach succeeds at

balance of the robot and retains the desired locomotion during trotting.

or the stride height and length. Subsequently, we evaluate the sim-to-real transfer by testing the single policy on both the Unitr
espectively, demonstrating the superiority of our method.
ion of 8.6 times compared to transmitting single-precision measurements.

ub.io/NFAtlas/.

woodlands, and farmlands, demonstrate that the proposed method is compatible for various robot platforms and adaptable to a

mild assumptions for the image sequences. We achieve a 14-24× speedup over existing implicit scene representation metho
scale registration of images to the point cloud map.

e robot struggles to find loop closures, employing our localization method improves the SLAM accuracy.

representation and 2D image pixels, which improves localization accuracy. At inference time, our mo

even better than the Lidar-to-Lidar place recognition, setting up a new record for utilizing the compressed Lidar map by low-en
– with inference speeds over 50 Hz on CPU. VPRTempo could be integrated as a loop closure component for online SLAM o

he readers to explore our project page and interactive demos: https://anyloc.github.io/

 chmark against the ArduPilot’s EKF3, the most sophisticated EKF implementation currently available.

ts reveal that the proposed method outperforms conventional GNSS positioning techniques, yielding more accurate position es

ue operations.

0% improvement for cloths in the real world), demonstrating the effectiveness of our approach in one-shot deformable object ma

mory. When we subject the quadrotor to turbulent wind fields with an attached drag plate, DMPO can adapt zero-shot while stil
observations. In particular, our approach results in an approximately 41% improvement in lap-time over the non-adaptive baseli
y, comparing very favorably to state-of-the-art results reported for similar tasks in the literature. We hope that these promising r

al details can be found at https://dibyaghosh.com/vptr/.

st above 4 seconds with minimal downhill slip. Combining gait optimization and human-tuning approaches, we found that fast tu

e towards generalizable object manipulation. More results can be found on our website: utkarshmishra04.github.io/ReorientDiff
h superiority over existing methods, demonstrating the effectiveness in tackling the noisy real-world point cloud problem.

cts. Furthermore, we quantitatively analyze data association at the segment level, which underpins inter-image connectivity dur

ect to a lightweight baseline normal estimation procedure and is therefore suited for operation in computationally demanding en

nidirectional swimming robot shows a great balance among stability, adaptability and maneuve

al-time performance of more than 16 Hz.

uca/uw_depth.
 its capabilities.

ual navigation and reconstruction frameworks. We provide full math derivations of the proposed models as well as example C++

n real-time during a mission with the low-cost hardware.

sed VIO extension.

ors' knowledge, this work presents the first autonomous aerial unperching framework using a fully actuated tiltrotor.

results. Finally, we optimized the tolerance of the power and dynamic system according to the theoretic model. With this pipeli

ut down to save energy on a mission, which is 48.9% of the operating time. Overall, the integrated system expands the operatio

n operated centimeters away.

 to validate both the control design and the guidelines.

hat thoroughly demonstrate, evaluate, and compare the performance of a real-time controller capable of selecting the stable CL

ctions for future developments.

·m of backdrive torque.
his could serve as a foundation to improve robotic systems' automation, integration, and efficiency of robotic systems for retina
, PVI-HVC offers the kinetic and kinematic performance of the PVIC during level ground walking, along with the freedom to voli
on the overall uncertainty. We demonstrate our visualization concept on iOCT volumes acquired at different stages during subre
g the performance of existing geometric shape fitting methods. These results demonstrate that the TSER holds great potential
f the systems promptly, thus helping to accelerate progress in the field.

cal diagnosis.

nce error mean±SD: 5.0±1.9mm vs 8.6±6.7mm on seven CTs).

he confidence and robustness of image segmentation. The experimental results demonstrate that the prop

control demonstrate consistent performance in the force range of 40 mN during mem

rgery in the future.
me scaling factor.
keletonize the critical structures without breaching any critical structure even under obstructed view of the surgical site.

-assisted lymphadenectomy.
thod not only allows for FOV adjustments with surgical instrument guiding but also adapts well to large-scale FOV adjustment ta

Furthermore, a demonstration using an eyeball model confirmed the successful implementation of 1R1T RCM.

d, simulated, and experimentally validated.

estimation through extensive experiments, demonstrating its effectiveness and reliability on out-of-distribution (OOD) data, add

demonstrated robustness to ultrasound noise and a wide range of scanning angles. It was also tested on data collected from si
th the cooperative mode by conducting a vessel-following experiment inside an eye phantom under a microscope.

esistant Helicobacter pylori infections.

1.89±0.54mm, respectively. The resection accuracy is demonstrated to be on par with manual glossectomy performed by an

s delivering drugs.

of 34.8±5.9s and a success rate of over 90%.

ned to extend the functionality catalog and expand the operating scope of these robots.
mean intersection over union (mIoU). Additionally, the CSG network excelled in sperm morphometric analysis, achieving errors

tail morphology measurement method achieved high measurement accuracies of 95.34%,96.39%,91.20% for length, width and
 llenging real-world experiments.

a common geometric controller. The LMPC is analyzed in a high-fidelity simulator, achieving an improvement of 8.9% in mean

provement may enable this method to be successfully deployed for in-situ behavior refinement, injury recovery, and terrain ada

gularization proposed.

n is performed during the near-grasp time to enhance grasp quality further.

mising to guide rehabilitation and enrich assistive device control laws with personalized force estimations.

cclusion and complex entanglement scenarios.

onomous navigation of MAMMOBOT and similar soft growing steerable robots.
ta, and visualizations are available at sites.google.com/view/acoustic-poe.
engineering.

r, aerial, and orbital environments). The core insight of this work is that when perturbation of the floating-base has affine structu

y on tasks with complex STL specifications while being 10X-100X faster than the classical method

by almost 20%, during same run-time. The source code of our method is released as an open-source package.

the chosen time step size. Moreover, our approach delivers smoothly optimized robot trajectories within the joint-space, offerin

over time. We support our approach by elucidating its analytical properties, showing the convergence of the anytime componen

rably with SOTA deep-learning-based methods while retaining the generalization capabilities primarily associated with classica

heading changes. Finally, field navigation experiments are carried out toverify the icability of APP.
ly push the boundaries of the state-of-the-art 3D printing mechanism.

ociated code is open-sourced and available at: https://github.com/jchenee2015/jetracer-4WS.

method have been validated through simulations and experiments.

on with the commendable computational efficiency. The experimental results demonstrate the feasibility of K-BMPC.
nd development.
. The proposed eware filter is applicable to nonlinear robot models and is computationally lightweight. We demonstrate the work

rmation and tracking algorithm into tethered cases, the proposed algorithm demonstrates improved performance in terms of mi
not directly applicable, we demonstrate that our CBF-INC-RRT can improve the success rate by 10%, compared with planning w

fields of view of human pose tracking under partial occlusions without sacrificing handover duration.

terHand achieves a substantial performance improvement over existing methods for multi-view and single-view interactive hand

amount of training data required and improves real-world classification accuracy.

eline methods.
as required by the unknown monitoring task. The proposed methods are validated via large-scale simulations of up to $100$ ro
ss and remarkable adaptability to the dynamic substitution in multi-robotic systems.
T-R reached a terminal state in 99% of experiments compared to 92% for MCTS. MLAT-R achieved these results while conside

vices on real mobile robots and demonstrated the proposed multi-agent operation.

acle avoidance by a group of Crazyflie drones. The performance of STGNN underscores its potential as an effective and reliable

ferent numbers of targets and varying teamsizes.

heuristics, with the former offering notable robustness benefits. During training, the learned incentive policy is found to get initia

ithm, which does not rely on ad hoc assumptions or rules, to prevent collisions and deadlocks and to deal

maximization of individual steady state distribution, into the objective. We compare E

rve a 100x reduction in computation time for comparable solutions.

te: https://github.com/wq13552463699/Real-Robot-Challenge-III-Winning-Solution

on dramatically to averaged 64.0%.

mance with 55.9%/45.7% NDS/mAP. Furthermore, we conduct extensive experiments to demonstrate its generality based on va

e term as unknown objectness score. Quantitative evaluations show that the proposed method outperforms state-of-the-art met
 addition, we conducted hardware experiments and successfully achieved 0.29 m and 2.59 degrees of frame alignment error. T

mental results demonstrate that probabilistic detectors trained using the outputs of MEDL-U surpass deterministic detectors trai

a lot than the state-of-the-art approaches.

enhancing the performance of object detection.The experiment results show that the network architecture significantly improve

publicly available at https://github.com/ hexunjie/Ro- temd.

her research exploring the potential of OVA-based active selection to address challenges in cold start scenarios and resource-

ire SemanticKITTI dataset, despite using only 5% of the annotations, outperforming existing active learning methods. We also m
aintain a tracking error within 0.346 m.
ency and practicality of the proposed data-driven control method.

al results are presented to illustrate the effectiveness of GPLA* in solving the ASPP.

n method about human presence can help the coordination method to find faster paths for the robots.

rials, and videos are released for reference.

ms both globally-guided MAPF solvers and communication-learning based ones, showcasing its potential towards scalability in r
ance improvement in instance segmentation accuracy with only a 6% increase in model size.

ime of 371 mus.

erceptive safety filter in real-time MAV experiments using exclusively onboard sensing and computation and show that the tele

) across the benchmarking algorithms running on the robot's onboard computer.

applications.

SE and a 29.5% lower MAE. Real-world qualitative experiments suggest Residual-NeRF leads to more robust depth maps with
al to be easily used for subsequent robotic tasks.

this effectiveness, we show that a simple proportional controller can robustly grasp a wide variety of objects. The results show
actical application of this approach.
view-rearr.

e validate the performance of our approach in a real-world scenario.

se dataset. We show that the proposed benchmark can produce a more accurate ranking of the models and serve as a means

g exoskeleton-based rehabilitation paradigms for effective gait rehabilitation.

ental procedure design.

s to human input, communicates its intent clearly, and resolves conflict. We report success rates and draw comparisons with hu

oposed tools can optimize the design of wearable robots for seamless integration with the human body.
ng with an assistive device, aiming to enhance users' comfort and mobility while ensuring reliable physical support.
s show that our approach is up to 50% more efficient than the baselines on average while merging

al belief hypotheses with ambiguous data associations.

ed approach increases the success rate in execution and reduces the number of replannings needed to reach the goal.

ement noise correctly when affected by environmental factors. We furthermore deploy our approach on a ground robot to showc

rtainty and an assembly task in simulation in the presence of pose uncertainty.

ware is available on GitHub, along with the developed autonomy framework, as open-source ROS packages.

sed algorithm is released for the benefit of the robotics community

nd long-term task scheduling. We further propose a novel leader-follower technique to boost cooperation learning and compare

perspectives.

itive results when compared to a more complex hybrid impedance-kinematic model over varying walking speeds and inclines.

nderstand the impact of exoskeletons on human motor control and thus enable safe and widespread adoption of those devices.
of generating accurate and intuitive hand movements for amputees in a physical environment.
the lower-limb human exoskeleton that relies on the simultaneous adaptation of intra-stride parameters in both frontal and sagit
que magnitude, and there may be a set of torque timing parameters that are generally effective.

while lifting the MW and HW compared to the 32% of the "Dynamic Arm". The int

ve robust reactive navigation that quickly finds a path to avoid obstacles while keeping smoothness of control input sequence. W

atch, new instances of categorical objects, and large object pose and shape variations.

bility to generate consistent torque with high robustness to motion disturbances.

ant superiority of the proposed SORL algorithm in safety-critical applications.

erall visibility of automotive AR displays. Our findings provide valuable insightsfor enhancing driving convenience and safety in
cting all information from the data in 120 room layouts. Check our website for videos.

ognition through a simple linear probe setup and excelling in the fetch object task under zero-shot transfer conditions. This wor

cally modulates the multiple MPIs to appropriately represent the diverse local density and color information. Experiments on the

ables us to achieve 3-4 times faster during training but also improve accuracy. Experimental results show that our approach ou

gement and long-horizon prediction. Finally, our unsupervised approach proves effective in the real world. With only minimal ad

e demonstrate the need to select only the most important features and show the competitiveness using various feature detectio

d scenarios. This approach enhances the dataset's applicability, capturing nuanced challenges faced during proximate vehicula

d real weather datasets. Source code and data are available at url{https://github.com/wangjiyuan9/WeatherDepth}.
over 30% improvements in driving safety.
s to establish dense correspondences.
OriTrack could be used as a future solution for solar tracking on rooftops.

surgical robot. The effects of the regularization terms are analyzed in an additional ablation study.

f autonomous systems.

nstraint violations than state-of-the-art approaches, while retaining reasonable performance in terms of tracking cost, number o

9DoF point cloud registration and its competitive performance in conventional 6DoF point cloud registration.
n networks. We demonstrate the accuracy of our uncertainty estimation approach under long-tail OOD obstacle classes for sem
single-precision float at a speed nearly twice as fast as the SOTA methods. We also conduct comprehensive real-world experim

tion and real-world tasks, and in all cases, UVD substantially outperforms baselines across imitation and reinforcement learning
method provides the best estimation accuracy.
an experts. It is expected that AV4GAInsp will alleviate inspectors' workloads and inspire further r
r matching algorithms and demonstrate the importance of retraining using crop scenes, and a training method customised for g

supervised approach and see a higher performance of our domain-specific postprocessing.

dences. Experiments are performed on three datasets on real and synthetic sequence of spatio-temporal plant point cloud data

e devices like NVIDIA Jetson Xavier.

fficient trajectories.

e that it achieves state-of-the-art performance.

ent detailed timing results of the main parts of the autonomy pipeline, which support the feasibility of deployment in operational

sting meta-learning and domain adaptation methods for excavation tasks.

ctures that will benefit the inspection and maintenance industry of civil infrastructure.
e widespread adoption of robotic systems in various service domains.

untargeted attacks and 20% of targeted attacks. Therefore, our work highlights the benefits of
mise for enhancing UHRI by improving robotic perception of human orientation underwater.

other route in the global re-positioning phase through path-way updates in the global graph. Finally, we showcase the performa

hods trained on manually labeled data.

tion with another point cloud for enriching the training samples. With the joint improvements of the radial imbalances, our metho
vailable at http://spino.cs.uni-freiburg.de.

tion and requires less computation time.

ngs well and achieve promising performance.

optimization of a variety of robotics missions. The efficacy of the overall framework has been evaluated in an object search sce

our proposed architecture and encoder representation sharing strategy. SG-RoadSeg also demonstrates superior performance

heir performance in both HD map semantic segmentation and instance detection tasks. Our code will be available at https://gith

g and evaluation, we propose an event synthesis approach based on the randomized optical flow that allows for directly genera
github.com/Jun-CEN/CMDFusion.
ies when using only IR data during inference in two public benchmarking datasets. This performance increase is achieved witho

ded dataset have been released at https://github.com/haomo-ai/SuperFusion.

vigation solutions in a wide spectrum of urban environments.

on. This joint spatio-temporal reasoning framework can effectively detect the moving object from motion confidence and geomet

g consensual spatial priors. Secondly, we propose the Reliability-oriented Depth Refinement (RDR) module to estimate dense

an other similar datasets available. The dataset is available online at https://github.com/norlab-ulaval/Norlab_wiki/wiki/TIGS-Dat

bust performance across two DoA estimation tasks using three different types of arrays.

or environment to evaluate their feasibility. The results indicate that ARIS 1.0 is feasible for providing remote medical services. F

o reproduce complex real-world experiments where a mobile vacuum robot interacts with a towel on different floor types and to

es. We also include 40k semantic BEV maps rendered from a digital twin. This novel approach to ground truth generation allow

community.
methods. Our code is available at https://github.com/tiev-tongji/LOG-LIO.

s demonstrate the superior performance of our proposed method. Specifically, it achieves an overall performance in terms of m
ts show that VOOM outperforms both object-oriented SLAM and feature points SLAM systems such as ORB-SLAM2 in terms o
ulti-sharing module focuses on fine-grained features, thereby boosting the performance of Lite-SVO. Experimental results demo

situational awareness by a richer scene model. Moreover, we make the software available as a d

eal-world 3D motion data.

Manhattan datasets.

the efficacy and efficiency of our approach.

 The source code for the training and the simulation platform is released online at h
ption. The dataset, code, and videos of our approach can be found at https://minaek.github.io/grounded_commonsense_reason

soft objects such as a piece of tofu and a balloon of water.

BL/s). The rolling gait was the most accurate gait, achieving 2.3% linear deviation.

near-ripe berries, with 89% (50) of the ripe and 64% (7) of the near-ripe berries being correctly classified. In a second in-lab exp
s and soft robots.

e fabricated and tested, demonstrating a performance improvement of over 140% compared to an

s a demonstration of its swift and buckling-free retraction.
architectures. This work advances the abilities of vine robots–and robots more generally–to not only access tightly constra
obstacles in highly confined areas are needed.

Vs. The design files, dataset, and source code are available on: https: //github.com/BioMorphic-Intelligence-Lab/Whisker-3D-Lo

external systems such as motion capture, which is essential for deployment in real-world scena

e possibility to power and control tetherless and fully soft robots.

eed up the RL process. Simulation-based experiments show that this approach allows the successful capture of debris at fuel c

mance. A multifaceted baseline evaluation was conducted using different event representations, event filtering strategies, and a
A's CADRE lunar rovers.
rmance of the VWDER, including its autonomous opening and closing, obstacle crossing, automatic reorientation and slope clim

estimation network. The trained model obtained a mean translational error of 2.59mm on a challenging dataset that presented v

ed on the da Vinci Research Kit (dVRK) for validating its practicality in real surgical robot settings. Our project website is at: http

+/- 2.39 mm or 5.61 +/- 3.11 % robot length.

400 ms. Experiments show that an average localization error of 0.95 mm can be achieved a

e hardware experiment in the Georgia Tech Robotarium.

 tree policy version is found on the supplementary webpage.
thod is evaluated in both simulation and reality under different scenarios emulating indoor environmental conditions.

ncounters intermediate failures, showcasing its robust re-planning capabilities.

abase for any pHRI scenario and thus for safe pHRI applications including edged and pointed geometries.

allocation, which can utilize the current and historical information to find the optimal allocation scheme in a global perspective.

pletion probability.

acket loss and short temporal breakdown of the communication link.

erties are in trade-off relationship that cannot be compatible. Finally, we proposed a seamless and automatic conversion mech
ns. To verify the effectiveness of our method, we conduct BEV 3D object detection experiments on three cross-domain scenario

trate a 50% decrease in learning time compared to a direct RL approach, while achieving similar control performance.

water guardians and Co-RIFT, a collaborative telepresence interface for multi-domain robotic teams.

al 3D space with real-time capabilities, accelerated using multi-threading and Tensorrt. The experiments demonstrate that the p

environments.

omy in percutaneous procedures, potentially enhancing safety in the process.

ing becomes favorable compared to that of a direct fall.

e emerging space economy.

7.3 cm/s, making it one of the fastest for its size, and can carry a 3.4 kg payload, making it one of the strongest.

s, offering exceptional maneuverability and robustness in challenging environments.

ed SRL to perform successfully various tasks suggests a great potential in future applications.

narios. The results achieved show an average error of 0.05 m in localizing the robot's base position and 6 degrees in joint state

thod outperforms previous methods in terms of estimating anchor position and improving localization accuracy.
omputational speed. Notably, even with 5 keyframes for initialization, our method consistently surpasses the state-of-the-art app

ion-making.

er movement on accuracy. Notably, using the 5 cm marker, we confirmed its compatibility to recognize IDs at distances exceed

t website https://sites.google.com/view/pixnav/.

mental results demonstrate resistance to noise, zero-shot sim-to-real transfer, and a decrease of 38.02% in terms of traversab
overage method surveys on average 36% more oysters when traveling the same distances. We also demonstrate the feasibilit
. Compared to baseline approaches, our findings indicate that PATERN robustly generalizes to diverse terrains and varied light
ge: https://fsd6d.github.io/

ccelerate the monocular-3D training tasks.

both view and style consistency. Our model leverages the generative capacity of large pretrained diffusion models within traffic

enefits of advanced LiDAR systems with the cost-effectiveness of single-beam LiDAR technology. Through this innovative comb

the real-time demands of robotics applications.

. The method effectively reduces haziness, enhances sharpness, preserves natural colors, and minimizes artifacts in challengi

paradigm for representing event data in a patch format compatible with transformer-based models, employing multi-spatiotemp

e conduct extensive experiments on four scenarios, including Synthetic to Real, Extreme Weather, Scene Changing, and Real

nicians automated, efficient, and reliable Cobb angle measurement.

scores competitively on surgical tissue tracking.

mpatible with our CL framework. We also design an adaptive weight assignment approach that balances the generalization ab
ons are integrated to enhance the object association. Comprehensive evaluations on VisDrone2019-MOT and UAVDT datasets

objects across occlusions using deep feature matching. We show that our model achieves strong performance on multiple est
each stage to further boost the performance as well.
ugmentations on the input. We complement this boot-strapping method with simple techniques that densify and re-balance the

thms. The code is published at https://github.com/luodongting/VI-HSO.

on on the public datasets further reveals that the proposed SRI-GVINS outperforms the state-of-the-art methods in terms of bo

/snoopy-l/MAT

ential for energy saving in portable pneumatic exoskeletons, showing that the proposed device extends exoskeleton operation b

agittal and Coronal planes. The experimental outcomes demonstrated a substantial reduction in misalignment with the knee wh
mpared to OSOP we improve on two datasets. The code is available at https://github.com/PhilippAuss/ZS6D.

hub.com/KangchengLiu to benefit the community.

rk not only outperforms traditional methods on 8 seen categories but also shows a powerful zero-shot capability for 8 unseen a
ve ability in complicated motion tasks.

ple-based, instruction-based and chain-of-thought prompting to write robot code; as well as a method to build the prompt levera

ent-learned behaviors through temporal task decomposition in RL. Our work contributes hundreds of simulated assets, tasks, a

d time-based approaches. The combination of online frequency-based, online similarity-based, offline similarity-based, and tim
ralding a new era where autonomous navigation is both feasible and reliable irrespective of the time of day. We make our code

er the geometric or learning-based approach alone.

eos and source code are provided.

ate the multi-directional traversability of the surrounding environments.

ate our feasibility.

DAR will provide new opportunities for boosting end-to-end motion forecasting models.

nt RL algorithms or different approaches for predicting rewards. Project website: https://robofume.github.io

size ornithopters.
which entails 69M parameters.

.2mm in Euclidean distance. When compared to some existing methods, the proposed method relies solely on the geometry of

anguage models in advancing the field of semantic navigation. Videos of real world deployment can be viewed at naoki.io/vlfm.
employed to expedite the training process, while the reward function also takes into account the smooth transition between adja

me operation while maintaining decent performance.

baseline and is able to guide real robots through long-horizon tasks. The performance gap with zero-shot state-of-the-art mode
monstrate that our method achieves state-of-the-art performance on 3DMatch and KITTI benchmarks. The code for our method

a temporal task, we introduce the Parameter Accumulation Update (PAU) strategy to collect the updated DSP and TRP in targe

Furthermore, the tests accentuate our method's robust cross-camera generalisation capabilities. We intend to release both the
pression performance.

on a synthetic dataset, due to the probabilistic grasp distribution modeling.

tate-of-the-art effectiveness of our approach. We believe that our method will inspire further research on real-world-oriented rob
which outperforms existing SSL methods by a large margin. Project page: url{https://vlislab22.github.io/TCC/}.

obtain the optimal model and binarization parameters. These above improvements allow the proposed binarization method to

cess rate exceeds them by more than 50%.

 us underwater robot with fully vectored propulsion.

anced capacity for autonomous navigation in real-world settings.

ative methods.

oposed heuristic has a competitive performance ratio.

ious scientific and industrial applications, aligning with the goals of the Decade of Ocean Science for Sustainable Development

o 5 Beaufort (8.1 m/s). In essence, our mobile hub concept significantly boosts UAV mission availability, allowing for autonomou
making capabilities in vertical farms. Beyond agriculture, the versatility of this sensor extends to application in areas such as wa

s both reconstruction accuracy and robustness, thereby demonstrating the potential of the proposed sensor in large-scale surfa
he image similarity m
ning with uncertain travel durations.

rld applications.
g reliable task execution. We demonstrate that MoMo outperforms the baselines across system-level and unit-test metrics, with
th a success rate of 70%.
real world. To the best of our knowledge, our work presents the first successful system for demonstration-augmented visual MB

g finger's joint configurations.

bottom, left, and right parts of the tactile board were applied, to guide continuous directional movements without visual assistan

rol and torque measurements for velocity optimisation and control.

incorporation of ground truth information for 30,292 images. The enhanced dataset, named the Jacquard V2 Grasping Dataset

domain with a real human expert, improving success from less than 0.1 with an autonomous agent to 0.92 with a 0.23 human i

ur approach transfer to a physical robot.

0 ms, and 5.5 ± 6.7 mm.

iments with 21 participants on a driving simulator demonstrate the effectiveness of the proposed approach.

cing the robustness and adaptability of robotic systems.

ust were significantly lower in the condition with decreasing competence than in either of the other two conditions, suggesting a
nchmark comparisons with state-of-the-art LHMP methods, offering an unconventional perspective and a more intuitive underst
ereby underscoring a critical area of research. The code is available at https://github.com/jinlong17/AdvGPS.
rception modules, and that (2) is not trivially reduced by simply finetuning on perception outputs. Based on extensive experime

he state-of

w video on a number of tabletop manipulation tasks in both simulations and on a real-world robotic arm. We also demonstrate h
actuated robot hand.

cribed force, and handling the motion control in the base and the arm so that both can follow their respective task requests. Ov

chmarking.
der non-ideal data collection conditions.

neck. In certain scenarios, our approach even outperforms using precision observations, e.g., up to 8% increase in average veh

to maintain stability for IA as time goes to infinity. Our numerical results show that our approach achieves stability for an $m$ th
facilitates AV improvement by individualizing its training with more challenging cases out of those poorly organized scenarios. E

e in https://github.com/JWFangit/eMem-NDT.
C contexts. The opensource implementation of the framework is made available at https://github.com/maleenj/hmip_ros.git.
ven outperforms RL methods based on physical state features in terms of data efficiency for continuous control benchmarks. T

st module corruption and significantly superior zero-shot generalizability in our proposed method. The videos of this work are av
aptability when applied to different underwater scenarios. Our efficiency analysis further highlights the substantial potential of ou

n three vision-

s, suggesting a scalable skill transfer approach for contact-rich manipulation.

t the PIRL significantly outperforms both e2e RL and heuristic-based solutions in terms of generalization and performance. Sou

bers of objects. In addition, we show the effectiveness of our method deployed on the real robot in the supplementary video.
eneralizability and robustness of the same policy where it realizes zero-shot transfer to different robot platforms like Go1 and Al

al approach succeeds at navigating through demanding confined 3D environments, outperforming both pure end-to-end learnin

le policy on both the Unitree Go1 and A1 robots. Remarkably, we observe robust performance, even when adding a 15 kg load
forms and adaptable to a range of terrains.

ene representation methods with comparable quality, showcasing its potential to facilitate faster and more effective 3D scene u
ssed Lidar map by low-end mobile cameras. The code are publicly available at href{https://github.com/Master-cai/VOLoc}{https
ponent for online SLAM on resource-constrained systems such as space and underwater robots.
more accurate position estimates.

shot deformable object manipulation.

adapt zero-shot while still outperforming all baselines. Additional results can be found at https://tinyurl.com/mr2ywmnw.
er the non-adaptive baseline while remaining safe across different environments.
ope that these promising results and our high-quality open-source implementation will provide a tool for the robotics community

ches, we found that fast turning is empowered by creating anisotropic torques with the sweeping wheel.

a04.github.io/ReorientDiff
oint cloud problem.

er-image connectivity during mapping and segment-level localization when revisiting the same place. Finally, we show prelimin

putationally demanding environments.

ls as well as example C++ header files for easy incorporation in independent projects.

uated tiltrotor.

etic model. With this pipeline, our real-time flight achieved 8m/s controlled vertical-descent-speed (CVDS), which is a leading re

stem expands the operational capabilities and enhances the energy efficiency of aerial robots for long-term monitoring tasks.
of selecting the stable CL equilibrium AEQ during operation.
 robotic systems for retinal surgery. The dataset created for this work is made publicly available.
ng with the freedom to volitionally pursue alternative activities.
ferent stages during subretinal injection on ex-vivo porcine eyes. We show that our processing pipeline achieves sufficient upd
SER holds great potential for effective and efficient movement and inspection within the colon, offering a promising solution for i

the surgical site.

e-scale FOV adjustment tasks.

stribution (OOD) data, adding an additional layer of safety and precision to minimally invasive surgical robotics.

d on data collected from silicon-based aorta phantoms, thus demonstrated its potential for translation from sim-to-real. This wor
microscope.

sectomy performed by an otolaryngologist.

analysis, achieving errors below 20%. Moreover, sperm motility parameters were further derived from the segmentation result

.20% for length, width and curvature, respectively.

ovement of 8.9% in mean lap times when compared to the MPCC.

recovery, and terrain adaptation, particularly in domains where simulations provide poor guides for the real world.
ng-base has affine structure with respect to joint velocity, it provides the system a class of kinematic reduction that permits the

hin the joint-space, offering flexibility and adaptability for various tasks. It seamlessly integrates task-specific requirements such

of the anytime component towards optimal solutions. Additionally, we provide empirical validation through simulation and real-w

y associated with classical planners.

ity of K-BMPC.
 We demonstrate the working of the framework through a simulation case study.

erformance in terms of minimising the risk of self-entanglement and maximising robot safety. These are validated in both simula
compared with planning with other steering controllers. Our project page with supplementary material is at https://mit-realm.git

ngle-view interactive hand reconstruction.

ulations of up to $100$ robots and targets.

hese results while considering an exponentially smaller action space than MCTS.

as an effective and reliable approach for achieving cohesive flocking, leader following and obstacle avoidance tasks.

policy is found to get initially closer to the expert-specified incentive and then slightly deviate from its trend.
 its generality based on various detectors with about +2% NDS improvements.

forms state-of-the-art methods when applied to the publicly available datasets.

f frame alignment error. Together with the alignment framework, our planner ensures safe navigation in unknown environments

eterministic detectors trained using outputs from previous 3D annotators on the KITTI val set for all difficulty levels. Moreover, M

cture significantly improves object detection on mAP and NDS, respectively.In addition, in order to demonstrate the effectivenes

t scenarios and resource-constrained training environments.

arning methods. We also match the existing state-of-the-art in active learning on nuScenes.
ntial towards scalability in realistic deployments.
on and show that the teleoperated MAV is able to safely navigate through unknown environments despite arbitrary inputs sent

e robust depth maps with less noise and fewer holes. Website: https://residual-nerf.github.io

objects. The results show that this control method outperforms slip detection methods, by reducing regrasp reaction times while
els and serve as a means of characterizing their behavior. We further present ablation studies to highlight contributions of differe

draw comparisons with human-human experiments to demonstrate the effectiveness of the approach.

sical support.
 to reach the goal.

n a ground robot to showcase its applicability for real-world active perception missions.

ion learning and compare our performance to conventional baselines in a wide variety of scenarios. There, our method closely

ing speeds and inclines.

adoption of those devices.

rs in both frontal and sagittal directions. The proposed control scheme has been validated with human subject experiments.

f control input sequence. We validate our proposals in a simulator of a quadrotor. Results show that our proposal involving both

onvenience and safety in real-world conditions.

 nsfer conditions. This work pioneers the application of sensory grounding in foundation models for robotics, promising a signific

mation. Experiments on the KITTI dataset and ScanNet dataset demonstrate that our proposed MMPI synthesizes high-quality im

how that our approach outperforms the state-of-the-art on a diverse range of downstream discriminative and generative tasks. C

orld. With only minimal additional data, our framework robustly predicts slots and their corresponding dynamics in real-world co

g various feature detection baselines.

during proximate vehicular operations. MMAUD emerges as an indispensable resource for scholarly investigation and practical

eatherDepth}.
of tracking cost, number of completed tasks, and computation time.

D obstacle classes for semantic segmentation in both off-road and urban driving environments.
hensive real-world experiments to validate the efficiency, accuracy, and robustness of the proposed SR-VINS.

and reinforcement learning settings on in-domain and out-of-domain task sequences alike, validating the clear advantage of aut
method customised for growing crops, with an average registration error of 27 cm across a season.

oral plant point cloud data to demonstrate the effectiveness of the proposed method.

deployment in operational environment.

e showcase the performance of the method both in simulation runs as well as deployed in real-world scene involving a legged r

dial imbalances, our method archives state-of-the-art performance on nuScenes and SemanticKITTI datasets, and consistently
 ed in an object search scenario, and has also been tested in a real-life experiment to emulate human common sense of unsee

ates superior performance than other SoTA collision-free space detection approaches. Our source code, demo video, and supp

be available at https://github.com/xjtu-cs-gao/SatforHDMap.

t allows for directly generating the event frame from a single exposure image. We further build two new datasets, E-MSCOCO a

increase is achieved without additional computational costs compared to the baseline segmentation models.

on confidence and geometric structure. Moreover, we conduct extensive experiments on various datasets to verify that the prop

module to estimate dense depth with the fine-grained structure based on the fused features and masks. We evaluate the effecti

Norlab_wiki/wiki/TIGS-Dataset.
emote medical services. Furthermore, the paper discusses several promising research directions related to the proposed conc

different floor types and towel geometries. Our system is able to reproduce 100% of the qualitative outcomes observed in the la

und truth generation allows us to produce more accurate and crisp semantic maps than are currently available. We make the si
performance in terms of mean matching accuracy of 0.68, 0.92, and 0.95 with respect to the tolerances of 1, 3, and 5 pixels, res
as ORB-SLAM2 in terms of localization. The implementation of our method is available at https://github.com/yutongwangBIT/VO
Experimental results demonstrate that our method is at least $84.46%$ faster than the state-of-the-art textit{Single-Stream} app
d_commonsense_reasoning/.

ed. In a second in-lab experiment, 16 of 20 berries were successfully picked, with 2 sustaining damage.
 nly access tightly constrained environments, but perform useful work once accessed.

gence-Lab/Whisker-3D-Localiz ation.

capture of debris at fuel costs that are notably lower than nominal baselines, including in scenarios where the debris is significa

filtering strategies, and algorithmic frameworks, and the results are summarized. The dataset will be made available at http://c
eorientation and slope climbing etc.

g dataset that presented varying levels of occlusion. These results highlight our pipeline's success in training and evaluating no

project website is at: https://med-air.github.io/SurRoL.

ntal conditions.

e in a global perspective. We conduct a comprehensive test of the proposed strategy in the ROS system as well as in the real

utomatic conversion mechanism from POstick-VF to KF, and vice versa, to cover up inherent limits of haptic devices.
ee cross-domain scenarios and achieve state-of-the-art performance. The code will be released at https://github.com/liujiaming

rol performance.

nts demonstrate that the propo

nd 6 degrees in joint state estimation. We believe this novel technique opens up possibilities for enhanced robot mobility and na

ses the state-of-the-art approach using 10 keyframes in rotation accuracy.

e IDs at distances exceeding 20 m, and pose recognition at 2.5 m was confirmed.

02% in terms of traversability cost compared to purely geometric-based approaches. Code and models are made publicly availa
demonstrate the feasibility of real-time deployment of UIVNav in pool experiments with BlueROV underwater robot for surveyin
e terrains and varied lighting conditions, while navigating in a preference-aligned manner.
 usion models within traffic contexts, effectively yielding diverse and condition-coherent street view images.

ough this innovative combination, we demonstrate a SLAM system that not only maintains high fidelity in mapping and localizat

mizes artifacts in challenging real-world scenarios. The proposed approach surpasses five SOTA methods in both qualitative an

mploying multi-spatiotemporal scale encoding to efficiently extract motion and semantic correlations from event representations

cene Changing, and Real Spike. Our method achieves state-of-the-art (SOTA) performances, compared with RGB-oriented uns

ces the generalization ability of the LLM and the domain expertise of the old CL model. Finally, we construct a new dataset for
MOT and UAVDT datasets show that DroneMOT provides substantial performance improvements over the state-of-the-art in th

rformance on multiple established benchmarks, and can produce reasonable tracks in manipulation data. In particular, our mod

ensify and re-balance the pseudo-labels, ensuring that we do not merely train on "easy" tracks. We show that our method yields

rt methods in terms of both accuracy and efficiency.

s exoskeleton operation by ∼27%.

lignment with the knee when compared to a commercial knee-support brace with a one-degree-of-freedom revolute joint.
capability for 8 unseen articulated object categories with only 17 demonstrations. Moreover, the real-world experiments on 7 d
 to build the prompt leveraging LLMs and human feedback. Secondly, we show PromptBook can write code for new low-level m

simulated assets, tasks, and demonstrations, taking a step towards fully autonomous robotic manipulation skill acquisition in sim

e similarity-based, and time-based decay methods shows superior performance compared to using individual forgetting strategie
of day. We make our code and thermal trajectory dataset public.
solely on the geometry of the flange and the pose of the rigid-body marker, making it independent of workspace constraints or

be viewed at naoki.io/vlfm.
th transition between adjacent paths. In our experiments, PathRL achieves better success rates and reduces angular rotation v
shot state-of-the-art models indicates that a lot of grounded data remains to be collected for real-world deployment, emphasizin
The code for our method is publicly available at https://github.com/CGuangyan-BIT/TrT.

ted DSP and TRP in target domain sequences. We conducted extensive experiments on two widely-used semantic segmentati

ntend to release both the dataset and code.

on real-world-oriented robotic manipulation.
ed binarization method to outperform binarization methods applied to convolution neural networks when used in point cloud Tra
Sustainable Development.

ty, allowing for autonomous operations even under challenging conditions.

ation in areas such as warehousing, logistics, and the food and beverage industry.

sensor in large-scale surface reconstruction. Codes and mechanical structures are available at: https://github.com/ZhangZhiyua
and unit-test metrics, with on average 23% and 40% higher success rates in simulation and on physical hardware, respectively

tion-augmented visual MBRL trained directly in the real world. Visit sites.google.com/view/modemv2 for videos and more detail

nts without visual assistance, which shows significant implications for machine-assisted cognition based on vibration tactile feed
uard V2 Grasping Dataset, served as the training data for a range of neural networks. We have empirically demonstrated that th

o 0.92 with a 0.23 human interaction rate. This approach provides a practical means for robots to interact and learn from human
o conditions, suggesting a notable recency bias. We conclude with implications for human-subjects experiment design (i.e., how
d a more intuitive understanding of human movement patterns.
ed on extensive experiments, we provide recommendations for critical areas that require improvement and guidance towards m

m. We also demonstrate how Rank2Reward can be easily extended to be applicable to web-scale video datasets.
pective task requests. Overall, we have developed a user-friendly control scheme, where an operator with little training and usi
% increase in average vehicle velocity in the merge environment, despite only using local traffic information as opposed to globa

eves stability for an $m$ that satisfies the theoretical conditions. We also empirically demonstrate that our proposed two phase
orly organized scenarios. Experimental results clearly indicate that CLIC surpasses other curriculum-based training strategies, s
maleenj/hmip_ros.git.
us control benchmarks. This may enlighten us: effective representation learning is more effective in reinforcement learning than

videos of this work are available on our project page: https://sites.google.com/view/composer-snake/.

substantial potential of our framework for onboard deployment.

ion and performance. Source code is available at url{https://github.com/Panshark/PIRL-WIN}.

e supplementary video.
platforms like Go1 and AlienGo robots for load carrying. Code is made available for the use of the research community at https

th pure end-to-end learning approaches and parameterized locomotion skills. We further demonstrate the successful real-world

when adding a 15 kg load, equivalent to 125% of the A1 robot’s nominal mass.

more effective 3D scene understanding.
m/Master-cai/VOLoc}{https://github.com/Master-cai/VOLoc}.
rl.com/mr2ywmnw.

or the robotics community to study new developments in robotic RL. Our code and videos can be found at https://serl-robot.gith
Finally, we show preliminary trials on segment-level `hopping' based zero-shot real-world navigation. Project page with supple
DS), which is a leading result in both quadrotors and coaxial UAVs.

-term monitoring tasks.

ne achieves sufficient update rates for surgical display and discuss the impact of our visualization concept on the acceptance of
g a promising solution for improved colon cancer screening.

from sim-to-real. This work represents a significant step towards safer and more efficient endovascular surgery using interventi
m the segmentation results for comprehensive sperm fertility analysis.
he real world.
eduction that permits the use of sample-based motion planners (specifically a nonholonomic RRT). We show that this allows ra

pecific requirements such as collision avoidance while maintaining real-time control rates. We validate the effectiveness of our

ough simulation and real-world demonstrations on a 6 degree-of-freedom robot manipulator, applied to an assembly domain.
 re validated in both simulated and real scenarios. An open-sourcesourcing implementation has also been provided for the bene
al is at https://mit-realm.github.io/CBF-INC-RRT-website/.
oidance tasks.
in unknown environments and collision avoidance in decentralized settings.

ifficulty levels. Moreover, MEDL-U achieves state-of-the-art results on the KITTI official test set compared to existing 3D autom

monstrate the effectiveness of the proposed local-multi-global equivariant components, we conduct sufficient ablation experime
spite arbitrary inputs sent by the teleoperator.

egrasp reaction times while decreasing grasping forces to 1-3 N.

ght contributions of different elements that are used to compute the proposed metrics
There, our method closely matches or outperforms the baselines; in particular, it yields higher-quality solutions and is at least 2 o
n subject experiments.

ur proposal involving both the spline interpolation and SVGD outperforms other base line methods.
botics, promising a significant leap in multi-sensory perception capabilities for autonomous systems. We have released the cod

synthesizes high-quality images from diverse camera pose distributions and is fast to train, outperforming the previous fast-trai

ve and generative tasks. Code is available at https://github.com/junshengzhou/3D-OAE.

dynamics in real-world control tasks.

nvestigation and practical research, facilitated by meticulous methodologies. It plays a pivotal role in advancing UAV threat det
he clear advantage of automated visual task decomposition within the simple, compact UVD framework.
scene involving a legged robot carrying camera and lidar sensor.

datasets, and consistently shows significant improvements along the radial distances. Our code is publicly available at https://git
 common sense of unseen-objects. For a video of the article, showcasing the experimental demonstration, please refer to the f

de, demo video, and supplement are publicly available at mias.group/SG-RoadSeg.

w datasets, E-MSCOCO and E-VOC based on the popular benchmarks MSCOCO and PASCAL VOC. Extensive experiments

sets to verify that the proposed method can improve the moving object detection accuracy by 13%.

s. We evaluate the effectiveness of our method on the synthetic and real-world datasets, which shows that, even without pretra
ated to the proposed concept.

tcomes observed in the laboratory environment. The simulation pipeline, named Stark (the German word for strong, as in stron

available. We make the simulation environment available to allow researchers to adapt the dataset to their specific needs.
es of 1, 3, and 5 pixels, respectively, on the HPatches dataset, outperforming all other SoTA algorithms.
b.com/yutongwangBIT/VOOM.git.
textit{Single-Stream} approaches, and excitingly, our pose accuracy is about $79.83%$ higher than theirs.
where the debris is significantly off-centered compared to the approaching chaser spacecraft.

made available at http://cvi2.uni.lu/spades.

training and evaluating novel vision algorithms for surgical robotics applications.
tem as well as in the real world, which proves the efficiency of our strategy.

haptic devices.
tps://github.com/liujiaming1996/BEV_UDA.
nced robot mobility and navigation in future exploratory and search-and-rescue missions.

ls are made publicly available: https://github.com/leggedrobotics/viplanner.

erwater robot for surveying a bed of oyster shells.
y in mapping and localization but also ensures affordability and broad applicability.

hods in both qualitative and quantitative evaluations across three key metrics, utilizing two real-world and three synthetic hazy im

om event representations. Exhaustive empirical evaluations conducted on the DSEC-Semantic and DDD17 datasets provide va

ed with RGB-oriented unsupervised depth estimation methods. The code and dataset are available at https://github.com/anony

nstruct a new dataset for surgical VQA tasks. Extensive experimental results demonstrate the superiority of our method to othe
er the state-of-the-art in the domain of MOT on drones.

ata. In particular, our model outperforms previous state-of-the-art in UVO and BURST, benchmarks for open-world object track

how that our method yields reliable gains over fully-supervised methods in real videos, for both short-term (flow-based) and long

edom revolute joint.

-world experiments on 7 different object categories prove our framework's adaptability in practical scenarios. Code is released a
e code for new low-level manipulation skills on the fly zero-shot. The prompt extracts motion trajectories from LLMs that the rob

ation skill acquisition in simulation.

dividual forgetting strategies.

workspace constraints or robot accuracy, without sacrificing the orthogonality of the rotation matrix. Our results validate the acc
educes angular rotation variability compared to other DRL navigation methods, facilitating stable and smooth robot movement.
d deployment, emphasizing the critical need for scalable data collection approaches. Finally, we show that video VLMs significa
used semantic segmentation CTTA benchmarks, achieving competitive performance and efficiency compared to previous state-
en used in point cloud Transformer structures. To demonstrate the superiority of our algorithm, we conducted experiments on tw
//github.com/ZhangZhiyuanZhang/GelRoller
cal hardware, respectively. More information at https://ut-austin-rpl.github.io/sirius-runtime-monitor/

or videos and more details.

ed on vibration tactile feedback.

ically demonstrated that these dataset improvements significantly enhance the training and prediction performance of the same

act and learn from humans in real-world settings.

periment design (i.e., how subjective measures are influenced by ordering effects) and provide design recommendations to min
 t and guidance towards more robust motion forecasting in the real world. We will release an evaluation library to benchmark m

deo datasets.
with little training and using a joystick, can guide the robot system to perform a physically interactive task on the surface of an
mation as opposed to global traffic information.

our proposed two phase algorithm has equivalent performance to the one-at-a-time rollout over the entire map, but with signific
ased training strategies, showing substantial improvement in managing risky scenarios, while still maintaining proficiency in han
einforcement learning than you think!
search community at https://github.com/DRCL-USC/RL_augmented_MPC.git

e the successful real-world deployment of our simulation-trained controller on a real robot.

nd at https://serl-robot.github.io.
Project page with supplementary details: oravus.github.io/RoboHop/.
cept on the acceptance of robotic task autonomy for subretinal injection procedures.

ar surgery using interventional ultrasound.

We show that this allows rapid, exploration-geared motion planning for high degree of freedom systems in obstacle rich environm

e the effectiveness of our framework through extensive simulations and real-world robot experiments in closed-loop manipulatio

o an assembly domain.
een provided for the benefit of the robotics community
ared to existing 3D automatic annotators. Code is publicly available at https://github.com/paathelb/MEDL-U.

ufficient ablation experiments. The results show that the individual components are indispensable for the object detection perfor
solutions and is at least 2 orders of magnitude faster than exact solver in cases where frequent coalition updates are required.
We have released the code, datasets, and additional results: https://github.com/gtatiya/MOSAIC.

ming the previous fast-training NeRF methods for novel view synthesis. Moreover, we show that MMPI can encode extremely lo

advancing UAV threat detection, classification, trajectory estimation capabilities, and beyond. Explore the dataset, codes, and d
licly available at https://github.com/jialeli1/lidarseg3d.
ation, please refer to the following link: https://youtu.be/hsGlSCa12iY

C. Extensive experiments demonstrate that our EOLO outperforms the state-of-the-art detectors, e.g., RENet, by a substantial m

s that, even without pretraining, our method outperforms the prior methods, e.g., RAMNet, especially in night scenes.
ord for strong, as in strong coupling) is made open-source.

their specific needs.

and three synthetic hazy image datasets. Notably, it achieves a substantial improvement in real-world datasets over the second

DD17 datasets provide validation of the effectiveness and rationality of our proposed approach.

t https://github.com/anonymous-4869/BiCross.

ority of our method to other advanced CL models.

or open-world object tracking and segmentation, despite never being explicitly trained for tracking. We hope that our approach c

erm (flow-based) and long-range (multi-frame) pixel tracking. Our code can be found here: https://github.com/AlexSunNik/refini
narios. Code is released at href{https://github.com/GeWu-Lab/LLM_articulated_object_manipulation}{https://github.com/GeWu
es from LLMs that the robot can execute directly with an IK controller. Finally, we evaluate the new skills on a mobile manipula
ur results validate the accuracy and applicability of the approach, providing a new and generalizable methodology for obtaining
smooth robot movement. We demonstrate the competitive edge of PathRL in both real-world scenarios and multiple challengin
that video VLMs significantly outperform single-image VLMs with an average error rate reduction of 19% across all VQA tasks
ompared to previous state-of-the-art methods.
nducted experiments on two different tasks: point cloud classification and place recognition.
performance of the same network, resulting in an increase of 7.1% across most popular detection architectures for ten iteratio
n recommendations to minimize them. We further discuss practical applications of these results in creating risk-aware robotic p
on library to benchmark models under standardized and practical conditions.
task on the surface of an asset.
 ntire map, but with significantly lower runtimes.
intaining proficiency in handling simpler cases.
s in obstacle rich environments, even on floating-base systems with nontrivial dynamics.

in closed-loop manipulation scenarios. For further details and supplementary materials, please visit: https://sites.google.com/vie
he object detection performance improvement of the overall network architecture.
on updates are required.
PI can encode extremely long trajectories and produce novel view renderings, demonstrating its potential in applications like aut

the dataset, codes, and designs at https://github.com/ntu-aris/MMAUD.

RENet, by a substantial margin (+3.74 mAP50) in all lighting conditions.Our code and datasets will be available at https://vlisla

in night scenes.
datasets over the second-best method, with gains of 0.5702 and 0.129 in FADE metrics for the RTTS and Fattal datasets, resp
e hope that our approach can serve as a simple and extensible framework for future research and enable imitation learning from

ub.com/AlexSunNik/refining-motion-code.
{https://github.com/GeWu-Lab/LLM_articulated_object_manipulation}.
kills on a mobile manipulator with 83% success rate at picking, 50-71% at opening drawers.
methodology for obtaining the transformation from the end-effector to a sensor.
os and multiple challenging simulation environments.
19% across all VQA tasks. Dataand videos are available at https://robovqa.github.io
chitectures for ten iterations. This refined dataset will be accessible on Google Drive and Baidu Netdisk,
ating risk-aware robotic planners capable of trust calibration.
https://sites.google.com/view/retro-optimal-control/home
tial in applications like autonomous driving.
e available at https://vlislab22.github.io/EOLO/.
S and Fattal datasets, respectively.
able imitation learning from videos with unconventional objects.