Research On Adaptive Motion Control System For Unmanned Surface Vehicle Based On Characteristic Model

The unmanned surface vehicle(USV)is a new kind of intelligent equipment on water.With the advancement of artificial intelligence and automation technology,USVs and other unmanned equipment are rapidly being widely used in military and civil fields.The working environment of USVs is complex and often disturbed by external factors such as wind,waves and currents in rivers and lakes,and as the functionality of USVs increases,the physical parameters of USVs themselves will change under different tasks,which makes control more difficult.As a key aspect in the design of unmanned boat systems,high precision motion control of USV has become the focus of research by scholars at home and abroad.In this thesis,we investigate unmanned boat motion control under changing external environment and uncertain model conditions,and build an adaptive control software and hardware architecture on the self-developed "CS-Ⅱ" unmanned boat.The main contents of this thesis are as follows:1.Develop an on-board hardware and software control system for the "CS-Ⅱ" USV,which is different from the traditional control platform and contains a recognition module and a control module that can match the adaptive control algorithm;analyse the kinematic and dynamical characteristics of the "CS-Ⅱ" USV,and establish an adaptive control system for the "CS-Ⅱ" USV.The kinematic and dynamic characteristics of the "CS-Ⅱ" USV are analysed,a three-degree-of-freedom model is established,and the model parameters are obtained through hydrodynamic parameter identification to establish the basis for controller design and verification.2.To address the tracking control problem of USV under uncertain model parameters,an adaptive control method based on characteristic modelling is proposed for the speed and heading control of USV.The virtual control quantities of forward velocity and bow rocking angle velocity are obtained through trajectory error;the second-order characteristic model of velocity and heading channels are constructed and the characteristic model parameters are obtained through online identification;on this basis,the adaptive controller with the combination of golden split control,feed-forward control and integral control is designed.The accuracy and adaptability are verified through outfield lake experiments.3.In order to be able to more accurately approximate the non-linear model of the USV,an adaptive internal mode control method for USV based on recurrent neural networks is proposed,in which the model of the unmanned boat is identified through recurrent neural networks and the output is predicted through this model,and the controller is designed according to the principle of internal mode control,which effectively solves the problem of the implementation of the forward and inverse models,and finally sets up a parameter update based on the model error threshold.Finally,a parameter update mechanism based on the model error threshold is set up,which effectively reduces the system computing pressure and improves the control accuracy.