PID SIMULATOR

One of our mentors has created a PID simulator to assist new programmers in understanding this important concept for the positioning of robots and their components. You can play with the simulator here: https://glowmouse.github.io/wasm_pid/

PUBLIC CODE

Our code for all of our robot’s mechanisms is public on Github for other teams to use as a reference or example, from 2017 to the present. The link to our current code which is written in Kotlin (a language based on Java) is here: https://github.com/Team865/FRC-2019

Some information of our code:

Subsystems are organized as finite state machines.
States are actions that subsystems can perform.
Actions can have sub-actions, which allows for complex routines
Meta-subsystems (RobotControl and Superstructure) control other subsystems
Action Management and Gradle’s Multi-Project Build for modular and easy-to-read code

SPECIFICS

Teleop Cycles

Limelight Vision Alignment: Dual-mode angle PID (quick turn and drive forward) to the rocket and loading station for optimized game piece acquisition and placement
Limelight Stream: Automatically switching between camera stream and vision mode
Lift Velocity Control: Squared inputs with a gravity-countering feedforward prevents the carriage from falling and allows for precise adjustments
Lift Setpoint Control: Position PID control loop with seven setpoints. Automatically accounts for drift using a Hall Effect Sensor. Precisely raises/lowers to any setpoint in under 1 second. Setpoint does not change when going to the bottom allowing faster cycles for the same level
Cargo Bi-Directional Passthrough: Control shared between the driver and the operator. Motor speeds are tuned to stop cargo after going through the lift to prevent falling. Cargo passes from intake to outtake in under 1 second
Curvature Drive: Driver controls throttle and radius of curvature with the addition of precise quick turning at a reduced speed
The Robot runs autonomous programs during sandstorm control until the driver touches the controller

Autonomous Programs

Wheel encoders combining with the navX-MXP sensor provides accurate odometry
Complex actions in series and parallel enable multi-state autonomous programs
Drive trajectory planning for both straight lines and curves using a motion planner. Curves use parametric spline functions to generate smooth position and curvature
Drivetrain kinematics is modelled through calculations and empirical measurements
To turn the robot in place, the robot uses a PID controller with an integral zone. This gets the robot to within 3 degrees to the target angle
To follow a trajectory, the robot uses a PDVA controller with a velocity intercept constant to account for static friction and a proportional controller to correct for drift in angle

Google Sites

Report abuse