Neural control of the underwater vehicle motion under conditions of uncertainty on the basis of the predictive controller

Abstract

The automatic control system of the speed of the underwater vehicle motion on the basis of the predictive controller under conditions of uncertainty of the object parameters has been synthesized and studied. The algorithm of the controller operation is described, and its structure is shown. A series of experiments was conducted on the mathematical model of the motion dynamics of the underwater vehicle. The training data sample was obtained and the artificial neural network approximating the object model was trained. The study of the impact of the controller parameters on the control accuracy was conducted. The best parameters were selected. The developed system showed high accuracy and performance. At that, it is not required to conduct complex experiments on obtaining data from the object for its synthesis. The disadvantages of the control system are the high complexity in obtaining a predictive model on the basis of the artificial neural networks and the high demands for hardware performance to compute the control signal.