Diagnosis using labeled Petri nets: Faults may either be silent or undistinguishable events

In this paper we generalize our previous results on the diagnosis of discrete event systems using Petri nets based on the notions of minimal explanations and basis markings. In particular, in our previous submissions we assumed that fault events only correspond to silent transitions, and there also exist silent transitions modeling regular behavior. Moreover, labeled transitions model regular behavior but they introduce a further form of nondeterminism because in general the same label can be shared by two or more transitions. Now, we assume that fault events can also be modeled by labeled transitions that share the same label with other fault transitions (e.g., belonging to different fault classes) and with other transitions modeling regular behavior. This requires redefining the main concepts on which our approach is based on and adapting the algorithms for diagnosis.