Diagnosis of I/O Automata Networks

This chapter is devoted to complex discrete-event systems that are represented by input–output automata networks. A method for decentralised diagnosis is developed where the diagnostic units have only access to the model and the measurement sequences of a subsystem. It guarantees that the diagnostic result is complete in spite of the lack of information about the subsystem interactions. Completeness means that the set of faults found by the diagnostic units include all fault candidates. Sufficient conditions for the autonomy of subsystems and for the kind of asynchronous state transitions are derived for which the result of decentralised diagnosis coincides with the result of centralised diagnosis.