2024 IEEE International Conference on Smart Internet of Things (SmartIoT)

Virtual Reality based Smart Mobile Robot

Navigation
1st Muthu Mariappan H 2nd Gomathi V 3rd Karthikeyan Jothikumar
KPIT Technologies Ltd National Engineering College National Engineering College
Bangalore, Karnataka, India Kovilpatti, Tamilnadu, India Kovilpatti, Tamilnadu, India
2024 IEEE International Conference on Smart Internet of Things (SmartIoT) | 979-8-3503-6644-0/24/$31.00 ©2024 IEEE | DOI: 10.1109/SMARTIOT62235.2024.00020

[email protected] [email protected] [email protected]

4th S.Hari Ram 5th R.Sri Gomathi 6th S.Karann

National Engineering College National Engineering College National Engineering College
Kovilpatti, Tamilnadu, India Kovilpatti, Tamilnadu, India Kovilpatti, Tamilnadu, India
[email protected] [email protected] [email protected]

Abstract—Control and monitoring of IoT devices are changing evolution of hardware and software has paved the way for
as a result of the integration of contemporary ICTs, including new opportunities in Virtual Reality (VR), offering innovative
virtual and mixed reality. Because they increase user engage- control solutions for various applications [6]. As working
ment and productivity, head-mounted displays, or HMDs, are
becoming more and more common. The trend of HMDs becoming conditions become more demanding and control requirements
more practical and accessible is bolstered by the developments stricter, VR-based systems can liberate operators from physical
in computer technology and the subsequent decrease in their constraints, providing immersive experiences by combining
cost. The two-wheeled robot automobile that we propose in this human perception with precise control of robotic systems.
work can be controlled remotely in real-time using an HMD. This project aims to develop a versatile system that operates
The Unity 3D game engine, which is economical and efficient
to develop for, is used to enable the remote control via Virtual in real-time with minimal hardware changes, addressing the
Reality (VR). Users can interact with the system using controllers growing need for Unmanned Aerial Vehicle (UAV) control
and hand movements thanks to distinct items for the screen, stop and other vehicle control methods [8]. By integrating the Intel
button, steering wheel, and transmission. The external cameras of RealSense D455 sensor for depth sensing, the system enhances
the Oculus headset are used for hand identification. To improve spatial awareness, which is crucial for navigating complex
spatial awareness and control precision, the system makes use
of depth sensing technologies. Three main tasks are performed environments [9].
by the Raspberry Pi 4: it processes control signals from the
HMD, streams video from a camera to the HMD, and uses VR presents a unique opportunity to enhance traditional
GPIO pins to operate DC motors. With the Raspberry Pi serving control approaches by integrating real-world feedback with
as a server based on the Flask framework, data is transferred virtual environments, especially in remote vehicle operation,
between the VR headset and the Remotely Operated Vehicle
(ROV) via server-client communication. As a client, the VR such as controlling a two-wheel robot car in hazardous or
headset is implemented in C# on the Unity game engine, and hard-to-reach areas [5] [6].
this server runs asynchronously, processing images using the
OpenCV library. In response to user inputs, the system moves The system’s architecture must prioritize compatibility and
the ROV and streams real-time video to the virtual reality screen. performance, given the challenges of mixed platform devel-
A responsive and immersive control experience is made possible
by the official Oculus Software Development Kit (SDK), which opment, including latency, documentation gaps, and potential
also manages depth data and input setups. compatibility issues [6] [8]. The proposed system will feature
Using data from large-scale experiments, this research analyses a two-way communication setup between a VR headset and a
the advantages and drawbacks of these novel input systems. Remotely Operated Vehicle (ROV), enabling real-time control
As a versatile and affordable replacement for conventional and live video transmission with minimal latency. The VR
physical control panels for ROV operations, the results show
that the suggested approach is workable for realistic real-world environment will simulate essential vehicle controls, such
applications. as steering and display functions, ensuring ease of use and
Index Terms—Virtual Reality (VR), Head-Mounted Display realism.
(HMD), Remotely Operated Vehicle (ROV),Raspberry Pi 4,Unity
3D,Oculus SDK, Depth Sensing, Server-Client Communication, By leveraging depth sensing for more precise navigation
OpenCV, Internet of Things (IoT)
and control, this system could be particularly valuable in
applications where precise maneuvering and remote operation
I. I NTRODUCTION
are required. The final system’s performance will be rigorously
The rapid advancements in technology are continuously analyzed and compared to existing solutions, guiding future
reshaping our daily routines and work environments [5]. The improvements and development [7] [9].

2770-2677/24/$31.00 ©2024 IEEE 71

DOI 10.1109/SmartIoT62235.2024.00020
Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on February 15,2025 at 17:12:10 UTC from IEEE Xplore. Restrictions apply.
 TABLE I system, these parts interact with one another via a software
S PECIFICATIONS OF THE ROVER stack.
Length 380mm
Width 200mm A. VR headset
Height 10mm
Ground clearance 40mm Oculus Quest 2 An advanced virtual reality headset that
Mass 162.89 grams
Wheel Diameter 80mm runs entirely on its own is the Oculus Quest 2. [9] As the
Camber Angle Adjustable VR environment that is required to run the remotely operated
Transmission Shaft Driven vehicle (ROV), it plays a vital part in the system. Control
Chassis ½ Scale 4WD on-road 510mm
and engagement with the ROV are improved by the headset’s
capabilities, which enable the receiving of real-time video
feeds and human commands.

Fig. 1. The rocker-bogie design

Fig. 2. Oculus Quest 2

II. S PECFICITION
III. A RCHITECTURE B. Raspberry Pi 4
The system employs a server-client model with the server on The system’s primary processing unit is the Raspberry Pi 4.
the mobile robot and the client (HMD) connecting via Wi-Fi. It The combination of this single-board computer’s adaptability
uses a Raspberry Pi 4 for computation, an L298N Motor Driver and processing capacity made it the choice [13]. With its
Module for controlling DC motors, and the Oculus Quest 2 quad-core Cortex-A72 processor, 4GB RAM, and multiple
for VR control [5] [6]. The Intel RealSense D455 sensor adds connectivity choices (such as Wi-Fi and Bluetooth), the Rasp-
depth perception, ensuring effective real-time operation. List berry Pi 4 is well-suited to manage hardware components and
of Components: run sophisticated software programs. Its 40-pin GPIO header
• Raspberry Pi 4 allows it to accept a large variety of peripherals, which makes
• - L298N Motor Driver Module it a great option for integrating with other sensors and motor
• - 6 DC Motors drivers [5].
• - Oculus Quest 2 The system’s primary processing unit is the Raspberry Pi 4.
• - Intel RealSense D455 Sensor The combination of this single board computer’s adaptability
• Vehicle Body with Wheels and processing capacity made it the choice.
The Raspberry Pi 4 is well-suited to manage hardware com-
IV. H ARDWARE
ponents and run sophisticated software programs. Its 40-pin
The Raspberry Pi 4, the Oculus Quest 2 VR headset, and GPIO header allows it to accept a large variety of peripherals,
mobile robot elements make up the three main aspects of the which makes it a great option for integrating with other sensors
project’s hardware architecture. To build a useful and effective and motor drivers.

72

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on February 15,2025 at 17:12:10 UTC from IEEE Xplore. Restrictions apply.
 1) Specifiction: control over the vehicle’s forward and backward motion. The
• The Raspberry Pi 4 with 4GB Ram features dual display chosen DC motors can operate between 5 and 35 volts DC,
micro-HDMI ports which supports H 265 decode for 4K which is compatible with the L298N and guarantees accurate
video at 60p and dependable control over the vehicle’s movements [5] [6].
• It is integrated with quad-Core 64-bit Broadcom 2711 and
cortex A72 processor and multiple connectivity choices
(such as Wi-Fi and Bluetooth).
• The Raspberry Pi 4 - 4 GB RAM has an operating power
of 5V at 3A via USB type-C port.

Fig. 3. Raspberry Pi 4 Model B

Fig. 4. L298N Motor Drive
C. Motor Driver
This L298N Motor Driver Module is a high-power motor D. Mobile Robot Body
driver module for driving DC and Stepper Motors. This The physical platform that supports all other components.
module consists of an L298 motor driver IC and a 78M05 Houses the Raspberry Pi, motor driver, motors, and camera,
5V regulator. L298N Module can control up to 6 DC motors, and provides the physical means for movement and manip-
or 2 DC motors with directional and speed control. The system ulation. Ensures that all electronic components are mounted
makes use of the L298N motor driver, a widely used module securely and function together as a cohesive unit.
for operating two DC motors at once. The Raspberry Pi can
be directly linked to this module, giving it control over the E. Raspberry Pi Camera
vehicle’s forward and backward motion. The Pi4B has 1x Raspberry Pi 2-lane MIPI CSI Camera that
The system makes use of the L298N motor driver, a widely are connect to backwards compatible with legacy Raspberry Pi
used module for operating two DC motors at once. The boards, and support all of the available Raspberry Pi camera.
Raspberry Pi can be directly linked to this module, giving it IDUINO RPI Pi Camera Module 3 Wide-angle, Sony IMX708

73

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on February 15,2025 at 17:12:10 UTC from IEEE Xplore. Restrictions apply.
Sensor, 11.9mp Resolution, 7.4mm Sensor Diagonal, f/2.2 Integration and Communication. The hardware parts work
Aperture, 15 x 1mm FPC Cable Connecter, 150mm Cable together to create a unified system. The Raspberry Pi manages
Ribbon Length. Built-in 2D Dynamic Defect Pixel Correction the image processing and control instructions on a server that
(DPC) enhances image clarity. Phase Detection Autofocus runs in Python and makes use of the Flask and OpenCV
(PDAF) ensures rapid autofocus. libraries [6] [7].
Through a Wi-Fi network, the server and Oculus Quest 2
communicate, allowing for the minimal latency transfer of
video feeds and user inputs [5] [6]. The motor driver, which
translates user commands from the VR headset into the robot’s
actual movements, is controlled by the Raspberry Pi’s GPIO
pins [5].

Fig. 5. Raspberry Pi Camera Module V2

F. Components of Mobile Robots

The DC motors discussed earlier are used to power a
wheeled vehicle body that is used to construct the mobile robot
[12]. The robot’s ability to travel and manoeuvre in a variety
of situations is ensured by this configuration. The robot also
has a camera module that can record full HD video and take
high-resolution pictures[14]. By using the camera’s output to Fig. 6. Flow chart of Hardware Mechanism
supply a live video feed to the VR headset, the operator can
view the world in realtime [9].

74

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on February 15,2025 at 17:12:10 UTC from IEEE Xplore. Restrictions apply.
 V. S OFTWARE A RCHITECTURE
This paper presents a software architecture designed for
real-time control and monitoring of a remotely operated
vehicle (ROV) using a Raspberry Pi and a Virtual Reality
(VR) headset. The architecture employs a client-server model,
utilizing Unity for the VR interface and Python with Flask
for server-side operations to ensure efficient communication
and low-latency data transmission [5] [6]. This setup allows
for seamless control of the ROV and an immersive user
experience.

A. Unity
The Unity component in the VR headset serves as the Fig. 8. Sample Code
main interface for the user. It captures user inputs, such as
steering and movement commands, through a virtual steering
F. OpenCV Camera Control
wheel, transmission, and terminate button. These inputs are
sent as HTTP GET requests to the Flask server running on the OpenCV is utilized for camera control within the Python
Raspberry Pi [10].Additionally, the Unity interface displays the component. It captures the live video feed from the ROV’s
live video feed from the ROV’s camera, providing real-time camera, processes the video stream, and prepares it for trans-
visual feedback to the user. mission back to the VR headset via the Flask server [7].
G. Flask
B. Web Client
Flask serves as the web framework for the server-side
The web client within the Unity interface handles the
operations on the Raspberry Pi. It manages the communication
communication between the VR headset and the Raspberry Pi
between the VR headset and the Python scripts, ensuring
server. It sends user commands to the Flask server and receives
efficient processing of commands and video streaming [7].
the video stream to be displayed in the VR environment [5].

C. Python
The Python component on the Raspberry Pi processes
the commands received from the VR headset. It includes
control scripts for the GPIO pins to operate the motor driver
and control the ROV’s movement. Python is also used for
managing the OpenCV-based camera control to capture and
process the video feed from the ROV’s camera [6] [7].

Fig. 7. Installation of the necessary Python libraries

D. Web Server
The web server, implemented using Flask, is the core of
the server-side operations on the Raspberry Pi. It handles
incoming HTTP requests from the VR headset, processes these
requests by invoking corresponding Python scripts, and con-
Fig. 9. Illustration of the software architecture
trols the ROV’s motors and camera. The server also streams
the processed video feed back to the VR headset [5] [6].
VI. F INITE E LEMENT A NALYSIS (FEA)
E. GPIO Control
Results of static load analysis (SolidWorks) on Rocker-
The GPIO control module in the Python component is Bogie mechanism commonly used in rovers Image by Author
responsible for interfacing with the Raspberry Pi’s GPIO pins. In the analysis This help us to proceed towards evaluating
It controls the motor operations of the ROV, translating user stress, strain & displacement along with mesh quality under
commands into physical movements of the vehicle [5] [6]. given loading condition.

75

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on February 15,2025 at 17:12:10 UTC from IEEE Xplore. Restrictions apply.
A. Von Mises Stress Analysis (top left) VII. NAVIGATION
The contour shows color gradient for stress field in the A. UserCommunication
structure. The highest von Mises stress does not seem to 1) Configuration for Virtual Reality: The user puts on the
exceed (1.531 times 101̂3) N/m², which is orders of magnitude Oculus headset, which creates a captivating virtual reality
above typical material yield strengths and should raise a flag setting. Several virtual controls, including a steering wheel,
for concerns in regards to the choice of materials or loads transmission lever, and other control panels, are intended to
applied. be interacted with in this Unity 3D-built environment.
B. Strain Distribution (top right)
B. Hand tracking and input capture
This plot indicates the strain realized by the mechanism,
To precisely track the user’s hand movements and gestures,
having a maximum value of 9.976E+1 The strain uniformity,
the Oculus headset makes use of external cameras [9].Unity
variance of stain under loading maintains within a pretty
3D records these inputs in real time and converts physical
regular range.
movements into virtual commands.
C. Displacement (bottom left) 1) Interaction with VR Controls: In Unity 3D, users can
use virtual controls by using their hands or the VR controllers
With a maximum displacement of (2.649 times 108̂) mm,
themselves. For instance, pressing a virtual button or rotating
the displacement figure shows how much every part of the
a virtual steering wheel will produce particular control signals
mechanism deforms under load.
(such as forward, backward, left, and right turns).
D. Mesh Quality (bottom right)
C. Transfer of Commands
The simulation’s finite element mesh is evaluated in this
mesh quality graphic. 1) Data Encoding: Unity 3D converts user inputs into
control commands by processing and encoding them. These
E. Real-Time Mapping Capabilities commands set the ROV’s speed, direction of travel, and other
Developing real-time mapping and localization capabilities operating characteristics.
using Simultaneous Localization and Mapping (SLAM) tech- 2) Server Communication: A Flask server operating on a
niques is essential for improving navigation. This allows the Raspberry Pi receives the encoded directives that Unity 3D
ROV to build a map of its surroundings dynamically while sends over the network. Typically, WebSocket connections or
it navigates, enhancing its situational awareness and decision- HTTP requests are used for this kind of communication, which
making. guarantees effective command transmission [11].

F. Multi-Agent Coordination D. Server Processing Command Reception

Another promising area for future exploration is the coor- The Raspberry Pi’s Flask server processes and accepts
dination of multiple ROVs for cooperative tasks. Investigating incoming control commands. These commands are handled
frameworks for multi-agent systems could improve exploration asynchronously by the lightweight Python web framework
efficiency and coverage in complex environments. This col- Flask, so other processes are not impeded in the process.
laboration may enable tasks such as shared mapping, resource 1) Motor Control: The commands are decoded by the
collection, or coordinated search and rescue operations. server and converted into actions for the DC motors on
the ROV. The Raspberry Pi’s GPIO (General Purpose In-
put/Output) pins are used to transmit signals to the motors,
which then adjust their speed and direction in response to
commands.
2) Video Streaming: The Flask server manages the ROV’s
camera’s video streaming simultaneously. After processing, the
camera footage is transmitted back to the Oculus headgear,
giving users a real-time view of the ROV’s surroundings.

E. Manage Performance
1) Motor Activation: To move the ROV, the Raspberry Pi
activates the DC motors in response to control inputs. To
perform the desired actions, such turning or moving forward,
the motors change their direction and speed.
2) Real-Time Adjustment: The VR environment is updated
in real time with the video stream and sensor data. This enables
Fig. 10. Design and analysis of a stairs-climbing users to see the results of their orders right away and make
any necessary revisions.

76

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on February 15,2025 at 17:12:10 UTC from IEEE Xplore. Restrictions apply.
F. Modification
[1] Real-time video feedback and depth information are
offered by the Oculus headset’s sensors. This information aids
the user in more accurately determining the ROV’s location
and surroundings.
1) User Adjustment: [1], [2] Based on the feedback they
receive, users can modify the control inputs they make. To
prevent a collision, the user could change the direction or
speed, for example, if the video stream shows the ROV
approaching an obstruction.
G. Constant Functioning
1) Continuous Communication: The Flask server, Rasp-
berry Pi, and Unity 3D client are always connected to each
other via the system. This guarantees prompt processing and
execution of control commands as well as seamless delivery
of real-time feedback.
2) System Monitoring: The Flask server keeps an eye on
the functionality of the system, responds to faults, and makes
sure that there is consistent communication between the VR
interface and the ROV. This continuous observation aids in
preserving a responsive and trustworthy control environment.
Fig. 11. Working of A* Algorithm
H. A* Algorithm
The A* algorithm is a popular pathfinding and graph traver-
sal algorithm used in various applications, including robotics, can be used for multi-robot coordination, allowing multiple
video games, and artificial intelligence. It is particularly effec- ROVs to plan paths that avoid collisions and efficiently cover
tive for navigating through complex environments where the an area.
shortest path needs to be found while avoiding obstacles
1) Graph Representation:The environment is modeled as VIII. F UTURE W ORKS
a graph where nodes represent positions and edges rep- The suggested VR-based Remotely Operated Vehicle (ROV)
resent the possible movements between those positions control system presents a viable basis for future improvements
2) Cost Function: A* uses a cost function, typically repre- and applications as technology advances. There are multiple
sented as f(n)=g(n)+h(n)f(n)=g(n) + h(n)f(n)=g(n)+h(n): primary areas into which the project’s future work will be
a) g(n)g(n)g(n): The cost from the start node to node divided:
nnn.
b) h(n)h(n)h(n): A heuristic estimate of the cost from
A. Improved Sensing of Depth
node nnn to the goal.
1) Sensor Integration: Use cutting-edge depth sensing tech-
I. Heuristic Function
nology to increase object detection and spatial awareness.
The heuristic function h(n)h(n)h(n) plays a vital role in Examples of this are upgraded LiDAR or Intel RealSense sen-
the efficiency of the A* algorithm. Common heuristics in- sors. This will increase the system’s precision when navigating
clude Euclidean Distance, suitable for environments with intricate or dynamic environments.
free movement, Manhattan Distance, used in grid-based 2) Sensor Fusion: Investigate how to integrate several sen-
environments where movement is restricted to horizontal and sors to give a more thorough picture of the environment around
vertical directions, and Diagonal Distance, a combination the vehicle. This can be done by merging depth data with
of Euclidean and Manhattan distances for environments with additional sensory inputs like thermal imaging or ultrasonic
diagonal movement. A* offers significant advantages, such as sensors. .
guaranteeing the optimal path if the heuristic is admissible and
improving efficiency by reducing the search space compared
B. Performance Enhancements & Optimization
to uninformed algorithms like Dijkstra’s. However, A* also
has limitations, including high memory consumption in large 1) Reducing Latency: Prioritize streamlining server-client
environments and dependence on the quality of the heuristic, communication to minimize additional latency. Real-time re-
which can impact performance. In the context of ROV nav- sponsiveness could be guaranteed by investigating methods
igation, A* can be adapted to dynamically update paths as like enhanced data compression, quicker data transmission
obstacles move, enabling real-time reactions. Additionally, it protocols, or edge computing.

77

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on February 15,2025 at 17:12:10 UTC from IEEE Xplore. Restrictions apply.
 2) Hardware Upgrades: Look into using more capable avoidance, and real-time video processing. Upgrading the
hardware or specialist parts to improve the Raspberry Pi DC motors and leveraging the computational power of the
4’s performance, especially when handling intricate control Raspberry Pi 4 is essential for improved precision and con-
signals and processing high-definition video feeds [3], [4] and trol. Additionally, integrating LED lights and infrared sensors
[13],. will enhance remote operation in low-light conditions and
improve proximity detection, significantly boosting overall
C. Enhanced Functionality
performance and user experience.
1) Multi-Device Compatibility: Increase the system’s
adaptability to several VR headsets and controllers made R EFERENCES
by different companies, hence improving its usefulness and [1] J. J. Leonard and H. F. Durrant-Whyte, “Mobile robot localization
accessibility on a variety of platforms. by tracking geometric beacons,” IEEE Transactions on Robotics and
Automation, vol. 7, no. 3, pp. 376–382, Jun. 1991. doi:10.1109/70.88147
2) Advanced Control Features: Add more control features [2] B.-Y. Xie et al., “Visualization of higher-order topological insulating
to give users a more natural and immersive control experience, phases in two-dimensional dielectric photonic crystals,” Physical Review
like haptic feedback. This could use virtual reality feedback Letters, vol. 122, no. 23, Jun. 2019. doi:10.1103/physrevlett.122.233903
[3] T.-M. Wang, Y. Tao, and H. Liu, “Current researches and future
from the car’s surroundings or simulated in-person encounters. development trend of Intelligent Robot: A Review,” International Journal
of Automation and Computing, vol. 15, no. 5, pp. 525–546, Apr. 2018.
D. User Interface and Experience doi:10.1007/s11633-018-1115-1
1) Enhanced User Interface: To enhance usability and [4] D. Bie et al., “Parametric L-systems-based modeling self-reconfiguration
of Modular Robots in obstacle environments,” International Jour-
control precision in the VR world, create more complex user nal of Advanced Robotic Systems, vol. 15, no. 1, Jan. 2018.
interfaces. This might entail more user-friendly interaction doi:10.1177/1729881418754477
techniques, adaptable control layouts, or more intricate visu- [5] A. Flores-Abad, O. Ma, K. Pham, and S. Ulrich, “A re-
view of Space Robotics Technologies for on-orbit servicing,”
alizations. Progress in Aerospace Sciences, vol. 68, pp. 1–26, Jul. 2014.
2) User Training and Simulation: Provide venues for sim- doi:10.1016/j.paerosci.2014.03.002
ulation and training to assist users in mastering the system. To [6] J. Schmiedeler and R. Gregg, “The next generation of robotics: The
2023 IEEE/RSJ International Conference on Intelligent Robots and
get people ready for real-world operations, this could involve Systems in Detroit, MI, USA [conference highlights],” IEEE Robotics
virtual training situations and lectures. & Automation Magazine, vol. 31, no. 2, pp. 135–140, Jun. 2024.
doi:10.1109/mra.2024.3388269
E. PracticalUses and Evaluations [7] L. Yang, T. L. Henthorne, and B. George, “Artificial Intelligence and
Robotics Technology in the hospitality industry: Current applications
Field testing is a crucial step in assessing the system’s and future trends,” Digital Transformation in Business and Society, pp.
performance in real-world scenarios. It involves conducting 211–228, Oct. 2019. doi:10.1007/978-3-030-08277-2 13
comprehensive tests in many situations, such as dangerous or [8] Y. Yang et al., “Combined intrinsic and extrinsic proton conduction
in robust covalent organic frameworks for hydrogen fuel cell applica-
isolated places. Gather and evaluate information to evaluate tions,” Angewandte Chemie International Edition, vol. 59, no. 9, pp.
efficacy, safety, and dependability. 3678–3684, Jan. 2020. doi:10.1002/anie.201913802
1) Application Extension: Examine possible uses for un- [9] C. S. Chin, N. B. Kamsani, X. Zhong, R. Cui, and C. Yang,
“Unity3D serious game engine for high fidelity virtual reality training
manned aerial vehicles (UAVs), industrial robots, and au- of remotely-operated vehicle pilot,” 2018 10th International Confer-
tonomous cars in addition to potential uses for ROVs. Expand ence on Modelling, Identification and Control (ICMIC), Jul. 2018.
the system’s impact and applicability by customizing it for doi:10.1109/icmic.2018.8529900
[10] K. Premkumar and K. Gerard Joe Nigel, “Smart phone based robotic
various use scenarios.Including Emerging Technologies in arm control using Raspberry Pi, Android and Wi-Fi,” 2015 International
Integration. Conference on Innovations in Information, Embedded and Communica-
2) Artificial Intelligence: Look into combining AI with tion Systems (ICIIECS), Mar. 2015. doi:10.1109/iciiecs.2015.7192973
[11] C. Zhao et al., “Multiscale construction of Bifunctional Electrocatalysts
machine learning algorithms to improve the capacity for for long-lifespan rechargeable zinc–Air Batteries,” Advanced Functional
autonomous operation and decision-making. Activities like Materials, vol. 30, no. 36, Jul. 2020. doi:10.1002/adfm.202003619
course planning, environmental analysis, and obstacle avoid- [12] S. S. Pujari, M. S. Patil, and S. S. Ingleshwar, “Remotely controlled
autonomous robot using Android application,” 2017 International Con-
ance could benefit from AI support. ference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-
3) 5G Connectivity: Consider the system’s possible gains SMAC), Feb. 2017. doi:10.1109/i-smac.2017.8058248
from 5G connectivity, especially in terms of faster data trans- [13] E. Bertolazzi et al., “Efficient re-planning for Robotic
Cars,” 2018 European Control Conference (ECC), Jun. 2018.
fers and decreased. doi:10.23919/ecc.2018.8550215
IX. C ONCLUSION
This project successfully developed a real-time, remotely
operated vehicle using a Raspberry Pi and a virtual reality
headset. The system demonstrated robust functionality and
efficient operation, with joystick inputs performing flawlessly.
Challenges included hand input recognition due to unsuit-
able camera angles, indicating areas for improvement. Future
work will focus on enhancing features and integrating ad-
vanced functionalities like path planning algorithms, obstacle

78

Authorized licensed use limited to: ANNA UNIVERSITY. Downloaded on February 15,2025 at 17:12:10 UTC from IEEE Xplore. Restrictions apply.