UUV Trajectory Tracking Control Based on Gaussian Process Model Predictive Control

Aiming at the problem of Unmanned Underwater Vehicle (UUV) trajectory tracking control in three-dimension (3D) underwater environment, a trajectory tracking control method based on Gaussian Process Model Predictive Control (GP-MPC) method is proposed. By integrating the theory and methods of Gaussian Process (GP) with the framework of model predictive control to achieve modelling predictions of system uncertainty. Firstly, the mathematical model of UUV under five degrees of freedom is established and linearly discretized; Then, the model predictive controller is designed to achieve the tracking control of the desired trajectory and predict the unmodelled part of the UUV by using GP; Finally, the three-dimensional trajectory tracking simulation experiments are conducted to validate that the method has a better controlling effect compared with the traditional MPC method during the tracking process with uncertain model parameters.