Dynamic data model and robust controller with high‐frequency component reduction for unknown time‐varying discrete‐time systems

Abstract

This article presents an adaptive controller tailored for discrete‐time systems, emphasizing robustness and addressing time‐varying constraints. An equivalent model is constructed solely from input‐output data collected from an unknown plant. Performance optimization involves determining an appropriate learning rate guided by theoretical analysis. A control law is derived from the equivalent model and a proposed auxiliary controller, with parameter settings guided by theoretical analysis to achieve robust convergence in closed‐loop performance. Comprehensive experimental and comparative assessments confirm the superior performance of both the equivalent model and the proposed controller, especially in brushless DC motor position control. Moreover, compared to similar schemes, the proposed controller generates fewer high‐frequency control effort components. The investigation validates the approach's robustness in real‐world conditions, including uncertainties, disturbances, and common time‐varying constraints in practical plant scenarios.