Stackelberg Game-based Secure Tracking Control for an Unmanned Surface Vehicle under Actuator Attacks

During the navigation of an unmanned surface vehicle (USV), the control input is often vulnerable to intelligently malicious behavior, leading to deviations from the intended trajectory. To this end, a Stackelberg game-based secure tracking control (STC) of the USV is investigated, while considering the unknown dynamics, actuator attacks, and external disturbances. The velocity error dynamics, involving control input and actuator attack, are established to mitigate the burden of controller design in the Stackelberg game framework. To address the challenges posed by unknown dynamics and external disturbances, an integral reinforcement learning (IRL) algorithm with the additional neural network (NN) is deployed to directly derive approximate optimal solutions for the Stackelberg game, effectively mitigating instability of the STC decision-making process, and enhancing the robustness and adaptability of the USV. Simulation results confirm the effectiveness of the Stackelberg game-based STC scheme for the USV under actuator attacks.