Secure estimation for cyber-physical systems under adversarial actuator attacks

摘要

Almost all of the available literature on secure state estimation focus on sensor attacks. Unlike the existing results, this study investigates the attack-resilient state estimation problem for continuous-time linear systems with sparse actuator attacks and process noises. It is assumed that, the attacker has limited resources and can only manipulate a certain number of actuators. Specifically, a novel switched observer is proposed with milder design conditions which are derived in terms of linear matrix (in)equalities. It is shown that, the observer not only provides an attack-resilient state estimation, but also guarantees that the resulting observer error system is with noise attenuation level which can be optimised. Finally, a simulation example of a linearised reduced-order aircraft system is provided to show the effectiveness of the proposed approach.