Robust Control of Networked Discrete-Event Systems Using Predictive Supervisors

Control of networked discrete-event systems is important because many engineering systems are now connected via shared communication networks where communication delays and losses are unavoidable. In this paper, we investigated robust control of such systems, which is useful when the exact model of the plant is unknown or when one supervisor is designed to control many different but similar plants. We consider nondeterministic control, where the language generated by the closed-loop system is not unique. We use predictive control that takes control delays and losses into consideration. Necessary and sufficient conditions for the existence of predictive supervisors are derived. Examples are given to illustrate the results.