Distributed synchronous diagnosis of discrete-event systems

Abstract

Recently, the centralized and decentralized synchronous diagnosis of discreteevent systems have been proposed in the literature. In this work, we propose a di erent synchronous diagnosis strategy called distributed synchronous diagnosis. In this scheme, local diagnosers are computed based on the observation of the fault-free behavior models of the system components. It is considered that these local diagnosers are separated into networks, and are capable of communicating the occurrence of events and their current state estimate to other local diagnosers that belong to the same network. The diagnosers are implemented considering an speci c communication protocol that re nes the state estimate of the faultfree behavior of the system modules, reducing, therefore, the augmented fault-free language considered for synchronous diagnosis. In order to do so, boolean conditions are added to the transitions of the fault-free component models, which check if the occurrence of an observable event is possible according to the current state estimate of other local diagnosers. This leads to the notion of distributed synchronous diagnosability. An algorithm to verify the distributed synchronous diagnosability with polynomial complexity in the state-space of the system component models is proposed.