Adaptive Neural Network Optimal Event‐Triggered Control for Unmanned Surface Vehicle Based on Zero‐Sum Game

ABSTRACT This article develops an adaptive neural network (NN) optimal event‐triggered control strategy for an unmanned surface vehicle (USV) system with uncertain dynamics and unknown external disturbances. A new event‐triggered NN optimal control policy is developed, which is composed of the NN adaptive feed‐forward controller and an optimal error feedback control policy. The former is constructed by adopting the backstepping control design technique, and the latter is designed by using adaptive dynamic programming theory and a zero‐sum game. Event‐triggered mechanisms are designed to reduce the updating times of actuators. It is demonstrated that the developed optimal control strategy can ensure the controlled USV system is uniformly ultimately bounded (UUB) and achieve optimal control performance. Simulation and comparison results are given to verify the validity of the developed optimal control scheme.