CO - DIAGNOSABILITY APPROACH OF DISCRETE EVENT SYSTEMS WITHOUT GLOBAL MODEL BASED ON PETRI NETWORKS AND INTERNAL SUCCESSION EVENTS

Diagnosability ensures that the global model of a centralized system, will always be able to diagnose a predetermined set of faults previously listed unambiguously. However co - diagnosability guarantee that these faults are diagnosed in a decentralized manner using multip le local diagnosticians. The co - diagnosability property is stronger than diagnosability, because if a system is Co - diagnosable it is diagnosable; while a diagnosable system does not necessarily ensure the co - diagnosability of this system. The challenge of decentralized diagnosis approaches is to perform multiple local diagnostics and verify that they are equivalent to the global one without the need for a global model. This paper proposes an approach to obtain a decentralized co - diagnosable diagnosis struc ture of the discrete event system, without the use of a global model, based on internal succession events of local diagnosticians and Petri networks.