Skip to main content

Understanding PID Control

Learning Objectives​

  • Explain the fundamental concepts of a Proportional-Integral-Derivative (PID) controller.
  • Understand the role of each PID term (P, I, D) in controlling a system.
  • Identify scenarios where PID control is applicable and its limitations.

Core Concepts​

In robotics, simply commanding an actuator to move doesn't guarantee it will reach the desired position or maintain a specific velocity accurately, especially in the presence of disturbances or friction. This is where control systems come into play. A PID controller (Proportional-Integral-Derivative) is one of the most widely used feedback control mechanisms in industrial control systems and robotics due to its simplicity and effectiveness.

A PID controller continuously calculates an "error" value as the difference between a desired setpoint (SP) and a measured process variable (PV). It then applies a correction based on three terms:

  1. Proportional (P) Term: This term produces an output value that is proportional to the current error. A larger error results in a larger corrective action.

    • Effect: Reduces the current error quickly.
    • Problem: Can lead to oscillations and a steady-state error (offset).

  2. Integral (I) Term: This term sums up the past errors over time. It is designed to eliminate the steady-state error that occurs with a pure proportional controller.

    • Effect: Eliminates steady-state offset.
    • Problem: Can cause overshoot and make the system less stable if tuned poorly.

  3. Derivative (D) Term: This term predicts future errors by observing the rate of change of the current error. It helps damp oscillations and improve stability.

    • Effect: Reduces overshoot, improves settling time, and increases stability.
    • Problem: Sensitive to noise in the process variable, can amplify small measurement errors.

The outputs from these three terms are summed to calculate the final output that is applied to the actuator (e.g., motor speed, servo angle). Tuning a PID controller involves finding the right balance of the Proportional (Kp), Integral (Ki), and Derivative (Kd) gains.

Hands-On Exercise​

Exercise: Specifying a PID Controller for a Robot Wheel

  1. Specification (SDD Phase 1): You need to control a robot's wheel to maintain a constant target velocity of 1 meter/second, even when the robot encounters varying terrains or inclines.

    • Task: Define the Setpoint (SP) and Process Variable (PV) for this control system.
    • Task: Describe the "error" calculation.
    • Task: Explain, conceptually, how the P, I, and D terms would contribute to achieving the target velocity and handling disturbances. For example, what would a large P term do? How would the I term help with a consistent incline? What role would the D term play if the robot suddenly hit a flat surface after an incline?

  2. Visualizing Output (SDD Phase 2): Sketch a graph showing (conceptually) how the robot's velocity might behave over time if:

    • Only P control is used.
    • P and I control are used.
    • Full PID control is used. Label axes and key characteristics (e.g., overshoot, steady-state error).

Summary​

PID controllers are fundamental to achieving precise and stable control in robotic systems. By continuously monitoring the error between a desired state and the current state, and applying corrective actions based on proportional, integral, and derivative components, PID controllers enable robots to perform tasks accurately and robustly in dynamic environments. Mastering PID tuning is a crucial skill for any Physical AI engineer.