Modelling of networked systems for remote diagnosis

This paper considers the communication network effects in the remote diagnosis of discrete-event systems, where the diagnostic task is decomposed into an on-board and an off-board diagnostic task. The system to be diagnosed and the communication network are modelled by timed automata with discrete inputs and outputs. A novel serial composition operation for timed automata is proposed. For diagnosis, this operation can be used to obtain timed automata describing the system to be diagnosed including the communication network. As the benefit of this new approach, the off-board diagnostic system is capable of taking network effects into account. The results are illustrated by an example.