On most permissive observers in dynamic sensor optimization problems for discrete event systems

This paper considers the problem of dynamic sensor selection under the constraint of maintaining the property of K-diagnosability. We use an information-state-based approach to define the most permissive observer, a structure introduced in prior work that provides all feasible solutions for the controller (i.e., for sensor selections). The principal result of this paper is that the information-state-based controller results in no loss of generality relative to any other admissible controller.