Design of prediction governors for linear systems with input saturation

This paper focuses on the tracking control of linear systems with inaccurate reference signals generated through prediction methods. In our previous study, we proposed a shaping mechanism, called prediction governor, which shapes predicted signals to improve the tracking control performance of systems. Then, we analytically derived an optimal prediction governor whose performance is theoretically guaranteed. However, the applicability of the prediction governor was limited to minimum-phase systems without any input saturation as the previous result was obtained under some assumptions. In this paper, we propose a design method of prediction governors without considering the assumptions made in our previous study. First, we formulated a design problem of prediction governors by considering their gain constraints. Then, we solved it based on linear programming and minimal statespace realization method. Finally, we verified that the proposed method provides satisfactory stable prediction governors for linear systems with input saturation through a simulation.