Output Feedback Tracking Control of a MEMS Electromagnetic Micromirror Using Adaptive Q-Learning Approach

This study investigates a model-free output feedback (OPFB) tracking controller design in optimal quadratic performance sense for a MEMS electromagnetic micromirror. Specifically, the proposed control scheme works under the adaptive Q-learning framework and uses a state reconstruction method for assuring the OPFB tracking of the given reference trajectories. Employing the augmented system method, we first transform the optimal tracking control problem for the MEMS micromirror into a regulation problem with a discounted sense in performance. Then, we apply an adaptive Q-learning algorithm to learn the model-free tracking control law according to the collected data sequences including the past input and output information of the augmented system. Finally, we illustrate the validity of the adaptive Q-learning based OPFB tracking control solution, where the parameter variations of the micromirror and the influence of the excitation noises are both conducted.