Robust control of partially observed discrete event systems in the input/output semantics

Abstract

This paper considers a situation where it is only known that the exact model of the plant belongs to a set of possible models. The objective is to synthesize a robust supervisor such that it achieves legal behavior for all possible models. Previous works have solved the robust supervisory control problem in the original Ramadge-Wonham framework. This paper solves the problem in the input/output semantics introduced by Balemi (1994). Necessary and sufficient conditions for the existence of its solution are derived. When there does not exist a solution, a maximally permissive robust supervisor is synthesized.