Advanced Control
Overview
Beyond PID, advanced control techniques like State-Space, LQR, and Feedforward offer superior performance for complex systems. These are useful for multi-joint arms, swerve drives, or systems requiring high precision.
Techniques
Key concepts:
- State-Space: Models systems using state variables (position, velocity) and matrix equations.
- LQR (Linear Quadratic Regulator): Optimal control that balances error reduction with control effort.
- Feedforward: Predicts required output based on physics models (gravity, friction, inertia).
- MIMO: Multi-Input Multi-Output control for coupled systems (e.g., Swerve).
State-Space Control
Uses the equation
αΊ‹ = Ax + Bu to model how system state changes. This allows controlling multiple coupled variables (like arm angle and velocity) simultaneously. WPILib provides LinearSystem and LinearSystemLoop for implementation.LQR (Optimal Control)
LQR finds the 'best' control gains by minimizing a cost function. You tune Q (penalty for error) and R (penalty for effort).
High Q = aggressive correction.
High R = smooth/efficient motion.
High Q = aggressive correction.
High R = smooth/efficient motion.
When to Use
PID: Sufficient for 90% of FRC mechanisms (shooters, simple arms, intakes).
Advanced Control: Needed for inverted pendulums, balancing robots, or highly coupled multi-joint arms.
Advanced Control: Needed for inverted pendulums, balancing robots, or highly coupled multi-joint arms.