Codiagnosability of timed networked discrete-event systems subject to event communication delays and intermittent loss of observation

Abstract

We address, in this paper, the problem of codiagnosability of timed networked discrete event systems (TNDES) subject to delays and losses of observations of events between the measurement sites (MS) and local diagnosers (LD), and, for this purpose, we first introduce a new timed model that represents the dynamic system behavior of the plant based on the, a priori, knowledge of the minimal firing time for each transition of the plant and on the maximal delays in the communication channels that connect MS and LD, and also intermittent packet losses in the communication network, we then convert this timed model in an untimed one, and, based on the untimed model, we present necessary and sufficient conditions for TNDES codiagnosability and an automaton-based algorithm for its verification. Examples illustrate all the results present in the paper.