Preferential Discrete Model-based Diagnosis for Intermittent and Permanent Faults

Diagnostic à base de modèle préférentiel pour fautes intermittentes et permanentes

Valentin Bouziat (1) , Xavier Pucel (1) , Stéphanie Roussel (1) , Louise Travé-Massuyès (2)

1 DTIS, ONERA, Université de Toulouse [Toulouse]
2 LAAS-DISCO - Équipe DIagnostic, Supervision et COnduite

In this paper we consider the diagnosis of intermittent and permanent faults in discrete event systems. We present a logic based modeling approach associated with conditional preferences in order to produce a single diagnosis at each time step. Like all incomplete diagnosis approaches, ours is subject to deadlocks between the system and its diagnoser. In this paper, we address the detection of such deadlocks at design time with the rich semantic model-checker ELECTRUM.

Domaines: Intelligence artificielle [cs.AI]; Robotique [cs.RO]

Valentin Bouziat, Xavier Pucel, Stéphanie Roussel, Louise Travé-Massuyès. Preferential Discrete Model-based Diagnosis for Intermittent and Permanent Faults. 29th International Workshop on Principles of Diagnosis (DX 2018), Aug 2018, Varsovie, Poland. 8p. ⟨hal-01796310⟩