Robust supervisory control of discrete event systems against intermittent loss of observations

We address in this paper the design of robust supervisors that are able to cope with intermittent loss of observations and also make the controlled system achieve the specification language under nominal operation. In order to do so, we introduce a definition of robust observability that leverages possible observations of the events that are subject to intermittent loss of observations and address language permissiveness by extending the recently introduced definition of relative observability to robust relative observability. We present necessary and sufficient conditions for the existence of robust supervisors that make the controlled system achieve robustly controllable and observable or relatively observable languages and present a characterisation of all achievable languages. A running example illustrates all the results presented in the paper, and an example taken from the open literature is used to illustrate the efficiency of the robust design strategy proposed in the paper.