Improved adaptive resilient control against sensor and actuator attacks

In this paper, an improved adaptive resilient control scheme is proposed for mitigating adversarial attacks in cyber-physical systems (CPSs). By introducing a two-step backstepping approach, an adaptive bound estimation mechanism and a Nussbam function with faster growth rate, the effects of unknown sensor and actuator attacks are successfully mitigated. An exponentially-decaying barrier Lypunov function is used to constrain state variables. Compared with the existing results where only uniform ultimate boundedness is achieved, the developed controller guarantees that the closed-loop system is asymptotically stable and simultaneously the state constraints are not violated even in the presence of the time-varying sensor and actuator attacks. Simulation results are presented to verify the efficacy of the proposed scheme.