A Communication Protocol for Decentralized Fault Diagnosis of Discrete Event Systems

In this paper, a decentralized MPO-based diagnosis technique is developed. The proposed decentralized structure consists of a set of local diagnosers that each diagnoser has a most permissive observer (MPO) whose role is to turn on/off the sensors dynamically and observe the occurrence of events. The local diagnosers send their diagnosis information to a coordinator by a communication link. The responsibility of the coordinator is to infer the received data and make decision about the fault occurrence in the system. We developed a communication protocol to coordinate the local diagnosers. Also, an analyzing algorithm is developed whose role is to analyze the received data from the local MPOs and make accurate decision about the fault occurrence. We proved that the proposed decentralized MPO-based structure has less computation cost than a centralized MPO-based diagnoser.