Reference Tracking via Output Feedback for Constrained Uncertain Linear Systems
Output feedback, Robust Control, Constraints, Invariant Sets, Set-membership Observers, Disturbance Rejection.
This thesis presents a robust output feedback control approach designed to address constant reference tracking in linear systems affected by uncertainties, disturbances, and measurement noise. Our research contributions revolve around three primary aspects. Firstly, we establish the conditions for achieving the Output Feedback Controlled Invariance (OFCI) property within the context of linear systems that incorporate polytopic uncertainties. This OFCI property guarantees robust constraint satisfaction through output feedback even in the presence of uncertainties. Secondly, we introduce a dynamic output feedback controller tailored to uncertain models, which employs set-membership observers. These observers effectively reduce the set of feasible states, enhancing tracking performance by minimizing errors. Thirdly, to further reduce tracking errors, we propose a model update procedure that adjusts the nominal model used for tracking based on output measurements. We conduct extensive numerical experiments to assess the controller's effectiveness, demonstrating its capability to achieve reference tracking with significantly reduced errors for the uncertain systems under consideration. Our research offers valuable insights into addressing constant reference tracking in linear systems with output feedback, constrained control and state, while accounting for uncertainties, disturbances, and noise. The potential applications of our approach hold promise for advancing control strategies in practical systems operating under uncertain conditions.