Scribd
Upload
36 views2 pages

SpookyBotz Control Award

Control Award ftc

Uploaded by

heetsutaria1179
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
36 views2 pages

SpookyBotz Control Award

Control Award ftc

Uploaded by

heetsutaria1179
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Control Award Submission - KookyBotz

Autonomous Objectives
D
High-scoring & reliable auto, that is partner adaptabl
Detect the team prop using a computer vision mode
Delay auto if needed, to work with our partner’s aut P P P
Rapidly scores the purple and yellow preload
Driving by wall or through stage door, intakes 2 stack pixel
Approaches the backdrop and scores the pixel
Repeats steps 3 and 4 until 6 pixels are cycle
Parks in the 1 of 3 possible locations S

Autonomous Map
Autonomous Enhancements

Uniquely during autonomous, our robot performs these functions


to ensure accuracy, consistency, and performance
Detecting the partner pixel on the backdrop, we score in the S
other column, ensuring 50 points
Relocalizes using ultrasonic sensors and apriltags
Detects robot collisions and ceases robot operations while
attempting to re-align or re-score I

Intake

I S Star t

P
Park

D Deposit
Sensors Used (3x) (1 of 2) (1 of 3) (4x)

Cameras (x2)
Absolute Encoder (x1)
IMU (x1)

Read team prop for Allows us to track the Tracks the angular
autonomous and relocalize position of our arm with orientation of our robot.
using april tags. 12 bit precision
Distance Sensors (x2)
Odometry (x3)

Assist with redundancy of 3 deadwheel odometry


localization and alignment. pods localize our robot. After discovering an lack of
sensors and actuators purpose
V oltage Sensor (x1)
Motor Encoder (x1)
built for FTC, our captain founded
Enables consistent We use a motor encoder the company Axon-Robotics. We
power output (scaling to track the position of primarily use sensors from Axon,
power values according our box tube extension and our captain handles the
to battery charge). for consistent intaking. manufacturing, and distribution.
Driver Enhancements
Finite State Machine Control
Drone FSM: Fired Armed
By modeling every possible transition and state the robot can be in, we
queue certain behaviors and outputs in order to actuate and control our Claw FSM: Drop Closed Open
system efficiently and without error. We do this with a command based Arm FSM: Scoring Stored Intake
system, making additions of new states trivial.
Pivot FSM: Scoring Stored Intake
Slew Rate Limiting

Scaling robot acceleration non-linearly through a piecewise cubic model

reduces wheel slip and traction loss when autonomously driving.


Simple Input, Advanced Performance: Arm Inverse Kinematics

Presets on our controller streamline scoring on the nth row, by providing data
to our arm inverse kinematics that move it to convenient scoring positions,
separating the driver from manually controlling fine details such as deposit
height or angle. Read more about how our kinematics work on the next page.
Control Award Submission - KookyBotz
Key Algorithms
Arm Physics and Inverse Kinematics

Our robot’s extremely large and fast arm presented a unique opportunity to
introduce physics modeling and driver enhancements- control that could show
a considerable performance benefit over robots of the same archetype. This
also presented a learning opportunity for us, applying math and physics to FTC.

Inverse Kinematics

Using a mathematical model we take in target row of the backdrop and

distance from the backdrop. The model then translates them into angles the
robot can understand and travel to, which are arm and wrist positions. We
derived this model using trigonometry and dimensions from CAD, and it has
trivialized our deposit code. This mathematical arm control allows higher
accuracy and precision, compared to arbitrary scoring height presets.

Arm Feedforward Gravity Compensation

Calculating the center of mass of our arm and the effects of gravity on it enable

us to calculate the exact gravitational effects exerted on our arm motor. Based
on that force, we can then apply a modeled feedforward to our PID control
which will keep the arm balanced at whatever angle it was sent to.

Partner Pixel Detection


Stack Detection

Detecting partner ’s preload Using a CV model, we analyze

and scoring in the adjacent the centroids of visible white


column, ensures 50 preload contours, and based on each
points, lessening alliance contour ’s width and height,
coordination requirements. we use a linear regression to
determine the distance and
April Tag Relocalization
angle of the stacks from our
Every time our robot scores,
camera, allowing for error-
we use the april tags on the less realignment and intaking.
backdrop to relocalize.

CSV Data Logging

Loop T ime Optimizations


We timestamp and log robot data,

A mix of performant code, publisher/subscriber format lynx


such as robot positions or
reads and write, and UUID hub identification fixes, our teleop extension ticks, allowing for a
improved from running at 50hz to running at 450hz (+700%!). seamless debugging process.

Custom Pure Pursuit Pathing


Arm data over time.
Our custom pure pursuit pathing gives us smooth trajectories compared to rigid point to
point paths. Our following system performs movements tangential to the path, which is 2300ms Path 1640ms Path
the most efficient and fastest for mecanum characteristics, which results in a higher
velocity on average and quicker path completion (one such test shown on the right). Old PID to Point Our Custom
Pathfinding Pure Pursuit

Engineering Portfolio References


Arm Inverse Kinematics (pg. 14) Vision Localization (pg. 13)

Dynamic Feedforward System (pg. 14) Open-Source Github Repository (pg. 3)

Robot Control Architecture (pg. 12) Autonomous Strategy (pg. 4)

You might also like