Diagnosability analysis of partially observed Petri nets
Ning Ran, Xinyue Gu, Shaokang Zhang, Zhou HeAbstract
This paper proposes an approach for analyzing the diagnosability of a partially observed Petri net (POPN). In particular, POPNs are equipped with sensors in both places and transitions, allowing the observation of token counts in some places and the firing of transitions. We first extend the definition of diagnosability of labelled Petri nets (PNs) to POPNs. We then transform a POPN into a labelled PN and prove that the diagnosability of a POPN is equivalent to that of its transformed labelled PN. A special automaton, called Verifier, is constructed for diagnosability analysis, and necessary and sufficient conditions are given for diagnosability of POPNs based on the Verifier. Finally, two practical examples are given to illustrate the proposed approach.