Active diagnosis of discrete event systems

This paper addresses the problem of fault detection and isolation for a particular class of discrete event dynamical systems called hierarchical finite state machines (HFSMs). A new version of the property of diagnosability for discrete event systems tailored to HFSMs is introduced. This notion, called L 1 -diagnosability, captures the possibility of detecting an unobservable fault event using only high level observations of the behavior of an HFSM. Algorithms for testing L 1 -diagnosability are presented. In addition, new methodologies are presented for studying the diagnosability properties of HFSMs that are not L 1 -diagnosable. These methodologies avoid the complete expansion of an HFSM into its corresponding flat automaton by focusing the expansion on problematic indeterminate cycles only in the associated extended diagnoser.

Section: Propositionmentioning

confidence: 88%

“…Therefore it is itself diagnosable with respect to fault f L1 using the results in Sampath et al (1998) (see Proposition 1 therein): a live sublanguage of a diagnosable language is itself diagnosable. Hence, given the definition of L 1 -diagnosability in Eq.…”

Section: Proposition 4 a Level One Fault Fmentioning

confidence: 96%

Diagnosability Analysis of a Class of Hierarchical State Machines

Paoli

Lafortune

2008

“…With slight technical modifications, all the main results of this paper hold true when the liveness assumption is relaxed, cf. the approach employed in Sampath et al (1998). Some of the events in @ S are observable, i.e., their occurrence can be observed by sensors, while the rest are unobservable.…”

Section: System Modelmentioning

confidence: 99%

“…Most discrete-event fault diagnosis methodologies necessitate the construction of a Bmonolithic'' model of the system under consideration for diagnosability analysis and implementation; see Aghasaryan et al (1998), Bouloutas et al (1994), Console et al (2002), Contant et al (2004b), Debouk et al (2000), Hashtrudi Zad et al (2003), Jiang and Kumar (2002), Jiang et al (2003), Lafortune et al (2001), Zanella (2003, 2004); Lin (1994), Lunze (2000), Pencolé et al (2002), Sampath (2001), Sampath et al (1995Sampath et al ( , 1996Sampath et al ( , 1998, Sengupta (2001), Sinnamohideen (2001), Yoo and Garcia (2004).…”

Section: Introductionmentioning

confidence: 99%

Diagnosability of Discrete Event Systems with Modular Structure

Contant

Lafortune

Teneketzis

2006

The diagnosis of unobservable faults in large and complex discrete event systems modeled by parallel composition of automata is considered. A modular approach is developed for diagnosing such systems. The notion of modular diagnosability is introduced and the corresponding necessary and sufficient conditions to ensure it are presented. The verification of modular diagnosability is performed by a new algorithm that incrementally exploits the modular structure of the system to save on computational effort. The correctness of the algorithm is proved. Online diagnosis of modularly diagnosable systems is achieved using only local diagnosers.

“…By designing for diagnosability as an intrinsic system property, detection of special behaviors is improved in detection capability, information management, and time response. Behavioral analysis of DES is an active area of research (Contant et al 2004;Jiang et al 2001Jiang et al , 2003aZad 1999;Sampath et al 1995Sampath et al , 1998Yoo and Lafortune 2002a, b;Fabre et al 2005; Baroni et al 1999;Pencole et al 2005;Lunze 2000). One relevant field is failure analysis, in which special events are identified as faults.…”

Section: Introductionmentioning

confidence: 98%

Event Counting of Partially-Observed Discrete-Event Systems with Uniformly and Nonuniformly Bounded Diagnosis Delays

Yoo

García

2008

We present an approach dealing with repeated fault events in the framework of model-based monitoring of discrete-event systems (DES). Various notions of diagnosability reported in the literature deal with uniformly bounded finite detection of counting delays over all faulty behaviors (uniform delays for brevity). The situation where the diagnosability notion of interest fails to hold under a given observation configuration leads typically to the deployment of more observational devices (e.g., sensors), which may be costly or infeasible. As an alternative to the additional deployment of observational devices, one might want to relax the uniformity of delays, while delays remain finite. To this end, we introduce a notion of diagnosability characterized with nonuniformly bounded finite counting delays (nonuniform counting delays for brevity), where finite delay bounds can vary on faulty behaviors. To evaluate the introduced notion of diagnosability with nonuniform counting delays, a polynomial-time verification algorithm is developed. Notably, the developed verification technique can readily be modified to construct a computationally superior verification algorithm for diagnosability under uniformly bounded finite counting delays (uniform counting delays for brevity) as compared to an algorithm previously reported in the literature. We also develop a novel online event counting algorithm that improves the time and space complexities of the currently available algorithms for the counting of special events.