Adaptive control for nonlinear cyber-physical systems in the presence of actuator attacks

In this paper, we consider time-varying multiplicative and additive actuator attacks for nonlinear cyber-physical systems and present an adaptive control architecture to suppress the effect of these attacks. The control architecture consists of a nominal controller and an adaptive corrective block that is used to modify the output of the nominal controller in the presence of the actuator attacks. It is shown that the closed-loop system remains uniformly ultimately bounded in the face of time-varying multiplicative and additive actuator attacks avoiding the need for a separate attack detection unit. In the case when the attacks are constant our adaptive control architecture guarantees partial asymptotic stability of the closed-loop system. Simulation results corresponding to the controlled dynamics of a flexible link robot, the controlled lateral directional dynamics of an aircraft, as well as a controlled axial flow compression system subjected to actuator attacks are provided to demonstrate the efficacy of the proposed approach.