Optimal dynamic output-feedback controller design for linear output regulation problems with its applications

This paper designs an optimal dynamic output-feedback controller (ODOFC) for linear output regulation problems based on adaptive dynamic programming (ADP). First, the overall model of the controller and compensator is established using the internal model principle and the minimal polynomial of the exosystem. Second, a state reconstruction technique is employed to reconstruct all system states using measurable input and output information. Subsequently, the delayed form of the minimal polynomial of the exosystem is introduced, eliminating the need to collect the exosystem states. Without relying on system dynamics, a value iteration based off-policy ADP algorithm is proposed to learn ODOFC using online input and output information. Finally, the effectiveness of this method is verified through simulations of a DC motor speed control system and a double integral system, thereby demonstrating its practicality and robustness.