Event-based fault diagnosis for an unknown plant
Mohammad Mahdi Karimi, Ali Karimoddini, Alejandro White, Ira Wendell Bates · 2016
This paper presents an active-learning technique for constructing a fault diagnoser for an unknown finite-state Discrete Event System (DES). The proposed algorithm actively asks some basic queries from an oracle through which the algorithm completes a series of observation tables leading to the construction of the diganoser. The resulting diagnoser is a deterministic-finite-state automaton, which detects and identifies occurred faults by monitoring the observable behaviors of the plant. An illustrative example is provided detailing the steps of the proposed algorithm.