Exploiting distinguishing sensors in supervisory control of DES

In discrete event systems, sensors can be used to observe occurrences of an event or to distinguish the occurrences of an observable event. In a previous paper, we formulate the supervisory control problem with distinguishing sensors to refine the set of events in a way that specifications can be represented by simpler models while the behavior of sensor is incorporated into the plant model. To avoid computational complexity, an approximation of the distinguished plant is used in the synthesis of suboptimal supervisors. In this paper, we introduce more general conditions on the approximating automata that can partially incorporate the sensor complexity in the synthesis to improve the quality of solution. An example of a manufacturing transfer line with material feedback shows the advantages of the results.