LPV Control Design for Precision Motion Systems

Motion control for optical and manufacturing applications is becoming evermore demanding due to the constant drive to increase speed and precision. Linear system theory alone is often inadequate to meet the demanding requirements. Gain scheduling control design addresses system nonlinearity by using a family of linear time invariant (LTI) systems parameterized by the operating point. Instead of switching between local controllers, this paper proposes a method to construct a multi-input/multi-output (MIMO) controller with a Linear Parameter Varying (LPV) formulation around multiple operating points. The goal is to preserve good closed-loop performance in local linear neighborhood while tackling the nonlinear dynamics by suitably interpolating the local LTI controllers. The proposed approach is applied to the simulation and experiment of a two-degree-of-freedom fast steering mirrors. The system response matches well with the nonlinear system response in simulation for both small and large motion ranges. The closed-loop performance is further improved by input trajectory optimization for large-motion FSM steering, demonstrating the advantage of establishing such a global LPV controller.