State-based vs Simulation-based Diagnosis of Dynamic Systems

In Model-Based Diagnosis there is an increasing interest in the diagnosis of dynamic systems. Some recent contributions in the literature show that in some cases such systems can be diagnosed with pure state-based diagnosis, i.e. reasoning on single states of the system rather than on transitions of the system from one state to another. In this work we discuss how in a different context the same results do not hold, and show how reasoning on the causality in the system and using simulation can provide more precise diagnostic results with respect to state-based diagnosis. Essential to this result are component fault models that characterize the discontinuity in the behavior associated with abrupt faults, i.e. the sudden transition of a component from the correct mode of behavior to a faulty behavior.