Design of Supervisors for Partially Observed Discrete Event Systems Using Quiescent Information

In this paper, we study the nonblocking supervisor synthesis problem in partially observed discrete event systems modeled by finite-state automata. We consider a particular type of supervisors that can observe not only the execution of observable events in a plant but also the quiescence of it. We first define a q-observer to characterize the behavior of a plant with observable quiescence. Comparing with the classical observer structure, the q-observer contains the quiescence information of a plant, which can be used to improve state estimation. Then we propose a method to detect the blocking states in a $q$ -observer. Finally, we develop an iterative method to synthesize a nonblockingness enforcement supervisor from the $q$ -observer. Since quiescence provides additional information on state estimation, the supervisor synthesized by the proposed method is in general more permissive than those synthesized by the existing approaches that do not monitor the quiescence. A manufacturing system example is also given to elucidate the effectiveness of the developed approach.Note to Practitioners-A discrete event system is a discrete-state and event-driven system, covering a deluge of contemporary computer-integrated man-made constructs, such as automated manufacturing systems, smart urban transportation systems and computer communication networks. Such a system is in general partially observed due to the limited sensor deployment, which complicates its controller design. This research studies the typical supervisory control problem for a partially observed discrete event system. A supervisor is designed to restrict the dynamics of the system in order to guarantee a safe operation through a different control scheme by using the quiescence information that can be usually provided by a real-world system. The practitioners in control and automation community are capable of practicing the formulated control scheme for engineering applications.