PID Controller
What is PID Controller in Humanoid Robotics?
A control algorithm that adjusts outputs based on Proportional, Integral, and Derivative calculations.
PID controllers are fundamental in robotics for maintaining desired positions, velocities, or forces by continuously correcting for errors.
How PID Controller Works
A PID controller continuously calculates an error value - the difference between desired setpoint and actual measured value. The Proportional term produces output proportional to current error (big error = big correction). The Integral term accumulates error over time, eliminating steady-state offset (if error persists, increase correction). The Derivative term predicts future error based on rate of change, dampening oscillations (if approaching setpoint quickly, reduce correction to prevent overshoot). The controller sums all three terms (each multiplied by tuning gains Kp, Ki, Kd) to produce the control output sent to the actuator. This cycle repeats continuously, often 100-1000 times per second, automatically correcting any deviations.
Types of PID Controller
- P Controller: Proportional only, simple but can have steady-state error
- PI Controller: Adds integral term to eliminate steady-state error, common choice
- PD Controller: Proportional plus derivative for reduced oscillation
- PID Controller: All three terms for robust control
- Cascaded PID: Multiple PID loops in series for complex systems
- Adaptive PID: Automatically adjusts gains based on performance
- Fuzzy PID: Uses fuzzy logic to tune parameters
Applications in Humanoid Robots
Joint position control uses PID to maintain desired angles for each robot joint during movement. Balance control employs PID to keep the robot upright by correcting lean angles detected by IMU. Force control in manipulation uses PID to maintain desired grip pressure. Velocity control regulates motor speeds for smooth walking gaits. Temperature regulation in motors and electronics uses PID. Vision-based servoing uses PID to adjust position based on visual feedback from cameras.
