Prescribed time performance control of underwater vehicles based on time-varying disturbance

To enhance the robustness and control efficiency of underwater robot systems, this paper presents a controller that integrates model predictive control (MPC) with a prescribed time control, and introduces a novel time controller capable of more intuitively and effectively determining the convergence time of system performance. A radial basis function neural network (RBFNN) coupled with an evaluation neural network has been engineered to bolster the system's capacity to address ocean current disturbances. The introduction of a new error function ensures the effective realization of the neural network's stability. By incorporating a MPC outer loop controller with system state parameters, the system state can be transitioned smoothly and kept within bounds, aligning with the actual performance characteristics of underwater vehicles. The system's stability is verified through theoretical analysis and further substantiated through simulation experiments. Additionally, the performance enhancement of the robot system achieved by the proposed algorithm is emphasised through a comparative analysis with the traditional PID control algorithm. Ultimately, in light of the complex, time-varying external disturbance environment delineated in this paper, a sea trial experiment is conducted to ascertain the effectiveness of the proposed algorithm.