Data-Driven Model-Reference Control With Closed-Loop Stability: The Output-Feedback Case
T. O. de Jong, Valentina Breschi, Maarten Schoukens, Simone Formentin
Abstract
We generalize a recently introduced data-driven approach for model-reference control design with closed-loop stability guarantees to the case of single-input single-output systems with inaccessible state. By considering a dynamic controller with fixed structure and leveraging a data-based description of the closed-loop dynamics, we propose a two-stage strategy for the optimization of the controller’s parameters to match the desired closed-loop behavior. By means of a benchmark simulation example, we show the potential of the proposed approach and the impact of a simple strategy to handle noisy measurements.
Topics & Concepts
Control theory (sociology)Benchmark (surveying)Stability (learning theory)Computer scienceController (irrigation)Loop (graph theory)Simple (philosophy)Closed loopReference modelFeedback loopControl (management)Control engineeringMathematicsEngineeringArtificial intelligenceEpistemologyAgronomySoftware engineeringGeographyComputer securityMachine learningPhilosophyGeodesyBiologyCombinatoricsControl Systems and IdentificationAdvanced Control Systems OptimizationFault Detection and Control Systems