Network-Based Modeling and Sampling Guaranteed Cost Control for Unmanned Surface Vehicle Systems Under Stochastic Cyber-Attacks

The future extensive usage of continuous measurable communication network for output feedback control in unmanned surface vehicle (USV) will inevitably increase the control cost, waste energy, and even reduce the system performance under stochastic cyber attacks. However, the existing study largely ignores the guaranteed cost control of USV under stochastic cyber attacks. To fill the knowledge gap, this paper attempts to focus on the guarantee cost control of USV systems by taking into account the finite measurable output sampling information and stochastic cyber attacks. Initially, a point output sampling control strategy based on finite measurable output sampling information is proposed, which has a strong application background and important theoretical research value in the field of power grid security. On this basis, a new stochastic composite system containing both system state and output sampling information is developed. Subsequently, a sampling-time-dependent Lyapunov function (LYFU) is exploited to ensure the workable criteria of the cost sampling controller, which can not only stabilize the stochastic composite system by means of the mean square exponential stability, but also reveal the upper bound (UPBO) of the quadratic cost function (COFU). Specially, in the strict sensu of minimizing the UPBO of the COFU, a meaningful design method for a suboptimal guaranteed cost output sampling controller is designed, which is a modulus-related LMI optimization algorithm. Finally, numerical simulation results of USV and additional vehicle lateral motion model are developed to demonstrate the validity and rationality of the proposed design method.