Topological Conditions for Automated Diagnostic Software to Estimate the Diagnosability of Digital Devices at the Structural Level

Abstract

Diagnosability is the property of a partially observable system with a given set of possible faults; these faults can be detected with certainty with a finite observation. Usually, the definition and the verification methods of diagnosability ignore the nature of controllable and uncontrollable events of the system. This paper shows the influence of controllability of system’s events on the definition and the verification, also shows that the classical diagnosability is a special case where we consider the whole system as controllable. The definition of diagnosability had been generalized using model structure on topological spaces by mean of strategies. Alternating-time Temporal Logic and Model Checking are used to check diagnosability of uncontrollable events to build a whole framework which is suitable for both isolated and interacting systems.