Model matching for asynchronous sequential machines with adversarial inputs using state bursts

Asynchronous sequential machines are referred to as finite state machines in which state changes are not governed by a global clock. This paper presents model matching for asynchronous sequential machines where an unobservable adversarial input can infiltrate and provoke unauthorized state transitions. The objective is to build an automatic state feedback controller so that the closed-loop system can match a prescribed model and all the unauthorized transitions by adversarial inputs can be invalidated. We address reachability and detectability properties of the asynchronous machine with adversarial intervention, and present necessary and sufficient conditions for the existence of appropriate controllers that realize model matching. Whenever controllers exist, algorithms for their design are outlined and an experimental verification is provided.