K-corruption Intermittent Attacks for Violating the Diagnosability

Abstract

This work addresses the problem of diagnosability analysis of a networked discrete event system under external malicious attacks, in which the considered system is modeled by labeled Petri net, and is attacked by a given replacement structure (i.e., the original transition label is replaced by a different label or empty string). In addition, we assume that the attack is stealthy and intermittent, which requires that the corrupted observations should belong to the normal behavior of the system, and the replaced transition labels must be recovered within a bounded of consecutive corrupted observations (called as K-corruption intermittent attack), respectively. The aim of this work is, from an attacker viewpoint, to design a K-corruption intermittent attacks for violating the diagnosability of systems. To this end, an attack automaton and complete unfolded verifier, i.e., a special automaton, are established to analyze all the potential attacked paths. Then an algorithm is proposed to obtain the K-corruption intermittent attacks, and examples are used to illustrate the proposed attack method.