Robust Control of Metric Discrete Event Systems Against Bounded Disturbances
Yiding Ji, Xiang Yin · IFAC-PapersOnLine · 2024
This work investigates robust supervisory control problems of discrete event systems modeled as finite state automata equipped with metric functions to measure the distance between states. The system may deviate from its nominal behaviors and fail to achieve the specification under disturbances whose effects are considered to be bounded. Accordingly, the supervisor should be designed to ensure that the controlled system degrades gracefully against such adversary. We formally formulate two problems: robustness bound verification and optimal robust supervisor synthesis. For the special case of verification under constant disturbances, a control Lyapunov function approach is introduced. Then we develop a two player game framework for the general verification and synthesis problems. Specifically, a dynamic programming method is proposed on the game structure, which uniformly tackles both problems.