Testability analysis using a discrete event systems framework
S. Bavishi, Edwin K. P. Chong · 2005
For a complex natural or a man-made system, automated fault detection and diagnosis often presents a challenging task. We previously (1994) presented results on the testability of a system whose fault behavior is modeled by a nondeterministic automaton. We also discussed issues pertaining to testability including minimal testable observation and the finest testable partition of the fault space. In this paper, we show that the definition of testability has a natural equivalent interpretation in terms of "one-step" faults. Specifically, we show the equivalence of the given automaton model and a directed bipartite graph under the testability definition. The equivalent bipartite graph representation allows us to derive an efficient fault diagnosis algorithm, when the system is assumed to be testable.