Optimal FDI and DoS Attacks on CPSs with Energy Constraint

Abstract

To have a good understanding of attackers’ behaviour and seek effective defensive measures, an optimal strategy of false data injection (FDI) and denial-of-service (DoS) attacks is proposed to intrude sensors and transmission channels of cyber-physical systems (CPSs) from the perspective of an attacker. Specifically, all sensors of CPSs make measurements of the state and send them to a fusion center for state estimation. Intruders attempt to modify the measurements and block innovations’ transmission by respectively launching FDI and DoS attacks to maliciously degrade the system estimation performance. A cost function is introduced to quantify the estimation performance. In order to maximize the cost function, it is important for the attacker to decide which sensors and transmission channels to intrude due to energy constraint. A convex optimization problem is formed, by solving which the optimal attack strategy is obtained. Simulations demonstrate the optimality of our proposed attack strategy.