Precise Docking Control of AUV Based on Neural Network Adaptive Controller

It is a great challenge for the docking of Autonomous Underwater Vehicle (AUV) in presence of model errors and actuator failure. The neural network adaptive controller including the state feedback controller and adaptive compensator is proposed. The state feedback controller is designed based on Linear Quadratic Regular (LQR) theory, which ensures the expected docking performances of AUV under no uncertainties. The neural network adaptive compensator derived based on the Lyapunov stability method helps to compensate the model errors and actuator failure, and Uniform Ultimate Boundedness (UUB) of AUV prediction errors is proved. Simulation shows that, the controller helps AUV to dock precisely with strong robustness to uncertainties and disturbance.