Research On Unmanned Surface Vehicle Path Following Based On Nonlinear Model Predictive Control

The general trend towards a strong maritime nation strategy and the autonomy of the oceans has promoted the need for and research into unmanned surface vehicles in both academia and industry.The complexity of the operating environment of unmanned vessels leads to a significant increase in the difficulty of controlling them.Nonlinear model predictive control(NMPC)is a modern control method with a high degree of practicality,which natively supports the imposition of various constraints on the controlled system,but is more computationally intensive than other control algorithms.With significant advances in computer hardware,NMPC controllers have been widely used in areas such as autonomous driving.This thesis investigates the application of NMPC control methods to the tracking control of unmanned vessels and is divided into three main parts as follows.(1)A model parameter estimation method based on optimization theory is proposed.The design of NMPC controller requires the system model of the controlled object for the prediction of the system state,based on the three-degree-of-freedom dynamics model of the unmanned ship,and the use of optimization methods to minimise the deviation between the actual system response and the model-derived response,so as to obtain the unknown hydrodynamic parameters in the model.Collect data on the motion of a particular ship model over a sufficient period of time in the simulation environment and reflect as many degrees of freedom as possible to improve the accuracy of parameter estimation.Experimental results show that the model parameters estimated by this method can make the state response of the system model sufficiently close to the actual response of the system.(2)A NMPC unmanned vessels path tracking controller with hot start capability is implemented.The reference path is obtained by a path planning algorithm,which simplifies the reference value of each calculation point in the optimisation objective by coordinating the resolution of the path search and the speed of the unmanned vessel with the discrete step size of the controller reference model,and improves the performance of the controller by adjusting the weights of each system state error.A hot start step is designed to speed up the computation,and the effectiveness of the control algorithm is verified through a physical simulation environment,and the results show that the controller can achieve fast and accurate tracking.(3)A nonlinear moving horizon estimation(NMHE)based method for estimating the effect of environmental disturbances is designed to improve the adaptive capability of the controller.By comparing the effectiveness of the controller with and without the NMHE method in the presence of disturbance effects,it is found that the NMHE algorithm can significantly improve the control effectiveness of the NMPC path tracking controller even though it cannot estimate the environmental disturbance effects accurately enough,compensating to some extent for the shortcomings of the model-based controller approach.