PID Control

Overview

PID (Proportional-Integral-Derivative) is a feedback loop that corrects errors. It drives a mechanism to a target setpoint (like an elevator height or shooter speed) by adjusting motor output.

The Terms

P (Proportional): Reacts to current error. "I'm far away, go fast." Main driver of motion.
I (Integral): Reacts to accumulated error. "I've been close for a while, push harder." Fixes steady-state error (e.g., gravity).
D (Derivative): Reacts to rate of change. "I'm closing in fast, slow down." Dampens oscillation.

Implementation Methods

Where to run PID:

  • On-RIO (WPILib PIDController): Runs in user code (20ms loop). Flexible, easy to debug.
  • On-Controller (Spark/Talon): Runs on motor controller (1ms loop). Faster, frees up RIO CPU.

WPILib PIDController

Best for learning and complex logic.

RIO PID Example

package frc.robot.subsystems;

import edu.wpi.first.wpilibj2.command.SubsystemBase;
import edu.wpi.first.math.controller.PIDController;

public class Elevator extends SubsystemBase {
    // Create PID controller: kP, kI, kD
    private PIDController pid = new PIDController(0.5, 0.0, 0.0);
    
    // Example: encoder and motor would be defined elsewhere
    // private Encoder encoder;
    // private Motor motor;
    private double targetHeight = 0.0;
    
    // ... (other methods and fields)
    
    @Override
    public void periodic() {
        // Calculate PID output: current position vs target
        double currentPosition = encoder.getPosition();
        double output = pid.calculate(currentPosition, targetHeight);
        
        // Apply output to motor
        motor.set(output);
    }
    
    // ... (rest of class)

On-Controller PID

Best for performance.

Hardware PID

Configuration:
package frc.robot.subsystems;

import com.ctre.phoenix6.configs.Slot0Configs;
import com.ctre.phoenix6.controls.PositionVoltage;
import com.ctre.phoenix6.hardware.TalonFX;

public class Example {
    private TalonFX motor = new TalonFX(1);
    
    public void configurePID() {
        // Configure PID gains for Slot 0
        Slot0Configs config = new Slot0Configs();
        config.kP = 0.1;  // Proportional gain
        config.kI = 0.0;  // Integral gain
        config.kD = 0.01; // Derivative gain
        motor.getConfigurator().apply(config);
    }
    
    public void setPosition(double targetRotations) {
        // Use position control with configured PID
        motor.setControl(new PositionVoltage(targetRotations));
    }
}

Tuning Process

Tune feedforward first, then PID. Use Phoenix Tuner or similar tools for quick constant changes. Log desired vs actual position using a vendor-specific tool (Phoenix Tuner, REV Hardware Client) or an external tool (AdvantageScope, Elastic) to see system behavior. (Credits to SuperNURDS FRC 3255 for the original method)

Feedforward Tuning

Steps:

  • Set all PID constants to 0.
  • Command the mechanism to a non-zero position.
  • Manually raise to that position and release.
  • Observe if it holds, raises, or falls.
  • Adjust kG (gravity feedforward) and repeat until it holds position.

PID Tuning

Steps:

  • Log desired position and actual position.
  • Slowly increase kP until you see overshoot, then set kP slightly below that value.
  • (For a riskier strategy, double kP until you see overshoot, then set kP to the midpoint between the old and current kP values until oscillation mostly disappears)
  • If steady-state error appears (doesn't reach target), increase kS (static feedforward) slightly.
  • This makes kP temporarily too large (overshoot), but your final kP will be close.
  • Repeat adjusting kS and kP until properly tuned.

Resources

Trapezoidal Motion ProfilingWPILib PIDMechanism PID Tuning Practice

Open full interactive app