Numerical and experimental analysis of motion control of offshore fishing unmanned underwater vehicle in ocean environment

A dual closed -loop motion control system is proposed to address the requirements of control systems for offshore fishing by unmanned underwater vehicles (UUVs) during autonomous operation. First, a reinforcement learning -based model predictive controller (RL-MPC) is represented as an outer -loop kinematic controller to plan for obtaining the expected optimal velocity commands and transferring them to the inner -loop controller. Second, a dynamic sliding mode controller (DSMC) is proposed as the inner -loop dynamic controller to obtain the expected optimal thrust input commands, and the aggregate disturbance to the control system is compensated by designing a nonlinear disturbance observer (NDO). In addition, the asymptotic stability of the control system is verified by stability analysis based on the Lyapunov method. Finally, the effectiveness and robustness of the proposed motion control system, which can solve the problems of uncertainty in the weights of the MPC objective function as well as the problems of singularity, velocity, and actuator chattering induced by the standard SMC, are verified by sufficient numerical simulations and offshore experimental analysis based on the combined sensing system.