Strong Tracking Iterative Learning Control for Linear Parameter-Varying Systems*

Abstract

This paper presents an advanced predictive Iterative Learning Control (ILC) methodology tailored for discretetime Linear Parameter-Varying (LPV) systems subject to strict input and state constraints, as well as bounded process and measurement disturbances. The proposed framework operates through a dual-phase architecture. First, a receding-horizon ILC scheme generates an optimal feedforward control sequence for a virtual, deterministic reference model. Second, the actual perturbed system tracks this learned trajectory using a robust feedback controller designed via the Lyapunov-Metzler (L-M) approach and Parallel Distributed Compensation (PDC). To strictly enforce physical hardware limits despite system uncertainties, the state and input constraints are systematically tightened leveraging Quadratic Boundedness (QB) theory. Consequently, the closed-loop system achieves guaranteed global stability and strong tracking performance, effectively decoupling the learning process from real-time noise. The efficacy and practical applicability of the proposed methodology are comprehensively demonstrated through the simulation of a multi-mode two-tank liquid level control system.