Multi-Objective Optimal Stealthy Attack on Cyber-Physical Systems

Abstract

This paper addresses security problems in cyber-physical systems (CPSs). A multi-objective optimal stealthy attack strategy is designed to disrupt the estimation process and degrade the stabilizing performance of the systems simultaneously. First, a feasible region in which the attack sequences remain stealthy is specified. Then, the attack strategy is derived from solving an optimization problem aimed at maximizing the perturbation of CPSs. The paper also outlines an algorithm that demonstrates the implementation of the attack. Finally, a simulation example is presented to illustrate both the stealthiness of the attack and the resulting degradation in system performance.