Design of observations graphs for partially observed Petri nets: Application to the diagnosability analysis of DES1

This paper addresses the problem of diagnosability for bounded or unbounded Petri nets and monitored with sensor configurations that are specified with partially observed Petri nets. The proposed method gives a necessary and sufficient condition for diagnosability. It is based on the computation of the finite coverability graph and its transformation into the graph of observation sequences that includes all sequences of measured markings and events with respect to the sensor configuration. Diagnosability is obtained from the analysis of the observation graph.