Neural-Network-based Finite-Set Model Predictive Control of an Autonomous Surface Vehicle Powered by an Electrical Motor

摘要

This paper is dedicated to a problem of surge speed control for an autonomous surface vehicle (ASV) powered by an electrical motor. The surge dynamics, propeller model, and motor model are unknown. A surge speed controller is developed based on a finite-set model predictive control and a neural predictor design. Firstly, two predictors based on neural networks are developed to estimate the unknown nonlinear functions existing in electrical motor model, propeller model and surge dynamics. The stability analysis is provided on the basis of input-to-state stability (ISS) theory. Then, a model predictive control scheme is proposed for surge speed control with a possible finite voltage control set and a predefined cost function. The simulation results are provided to illustrate the validity of the proposed surge velocity controller for the ASV.