NEURAL GENERALIZED PREDICTIVE CONTROL FOR AN AUTONOMOUS UNDERWATER VEHICLE

摘要

This paper investigates the application of neural generalized predictive control algorithm to an autonomous underwater vehicle of which the dynamics are highly nonlinear, coupled, and time-varying. The accuracy and robustness of this control strategy relies on the quality of the nonlinear vehicle model, in particular its ability to predict the vehicle's response accurately multiple-steps ahead, a practical plant modeling method in order to obtain a model of an autonomous underwater vehicle using dynamic recurrent neural network is presented. The minimization algorithm of neural generalized predictive control cost function is realized with the error between the reference trajectory and the predictive output of the dynamical recurrent neural networks. Simulation results are presented to illustrate the effectiveness of the proposed neural generalized predictive control strategy through the application to the vehicle surge velocity control of an autonomous underwater vehicle.