An Event-Based Stealthy Attack on Remote State Estimation

Security issues in cyber-physical systems (CPSs) have gained increasing attention in recent years due to the importance and unavoidable vulnerability of CPSs. This article focuses on designing an intelligent online attack, which can compromise a sensor, eavesdrop measurements, and inject false feedback information, against remote state estimation. From the viewpoint of the attacker, we design an event-based attack strategy to degrade the estimation quality with an arbitrary communication rate stealthy constraint. The approximate minimum mean-squared error estimation algorithm from the viewpoint of the attacker is derived under a Gaussian assumption. Furthermore, the relation between the attack threshold and the scheduling threshold is obtained in a closed form. We show that the mean-squared stability condition of the estimation is weakened under the attack. Two examples are provided to demonstrate the main results.