Zero-Sum Game-Based Tracking Control for Fully-Coupled Nonaffine AUV With Unknown Affine Disturbance Using Critic Network

In the process of water quality monitoring, autonomous underwater vehicle (AUV) is required to track desired trajectory. And AUV is a very complex nonlinear systems and exhibits fully-coupled and nonaffine properties. These make trajectory-tracking control more challenging. In this paper, the unknown affine disturbance is considered and a two-player zero-sum game (TPZSG) is employed to design tracking control scheme for fully-coupled nonaffine AUV (FCNAUV). Firstly, the mean value theorem is used to transfer the fully-coupled nonaffine system into affine system. Secondly,the error system is constructed based on the reference trajectory. Then, the trajectory-tracking control problem of FCNAUV with unknown affine disturbance is formulated as the TPZSG. One player is the optimal controller which is designed to make states of error system converge to zero. And the other player is the unknown affine disturbance that represents the overall damping effect of the FCNAUV. In order to reduce computing burden, only a single critic network is employed to realize TPZSG based adaptive tracking control scheme. Using the Lyapunov stability theory, the stability of the FCNAUV is analyzed and the system is guaranteed to be uniformly ultimately bounded (UUB). Finally, simulation and comparison results of FCNAUV with unknown affine disturbance verify that the proposed method has superior computational efficiency and accuracy.