Secure State Estimation for Cyber-Physical Systems with Actuator Attacks

This manuscript investigates the problem of secure state estimation for cyber-physical systems subjected to actuator attacks. Using multi-input multi-output continuous-time linear system modeling, two different state observers are proposed for attacked system state and unattacked system state. Firstly, a Luenberger state observer is proposed, which can effectively observe the unattacked state of the system. Then, based on the multivariable super-twisting sliding mode algorithm, a sliding mode state observer is designed to estimate the system state and attack signal under actuator attacks, and the method can make the system achieve bounded stability. Finally, simulation experiments verify the effectiveness of this method.