Research On Ship Course-Keeping Control Based On RBF-ARX Model

Ship motion is characterized by]arge inertia,the time delay are usually relatively large,even up to 100 seconds for million tons of oil tanker,and response to rudder is slow;there are some nonlinear factors such as delay and relay characteristics in the operation of the servo subsystem;the speed change of the ship and ship loading will lead to perturbation of the parameters and structure of ship motion model;meanwhile,the ship running environments is complicated.All these factors make ship course keeping control a complex nonlinear control problem,it is difficult to achieve the desired effect of the traditional course keeping controller.In this paper,the nonlinear control problem in ship course control is discussed,based on the RBF-ARX model,the ship model is modeled by the statistical modeling method.Finally,it is applied to the design of ship keeping controller.The main research work is as follows:Firstly,the structure of ARX model and RBF-ARX model are introduced.The RBF-ARX model is a global model obtained by approximating the function coefficients in ARX model by RBF neural network,this model can describe the smooth nonlinear process of a class of operating points with time.RBF-ARX model can be locally linearized at any operating point as a linear ARX model,which makes it easy to use the RBF-ARX model in control.The RBF-ARX and ARX models are established for the ship course keeping system,and the long-term prediction performace compared with the ship's physical model,it is verified that the superiority of the RBF-ARX model in describing the global dynamic characteristics of nonlinear systems.Then,the fuzzy PID heading control strategy and LQR ship course-keeping controller based on RBF-ARX model are designed.Fuzzy PID not only has the advantages of strong adaptability of fuzzy control,but also has the characteristics of high precision of traditional PID control,the LQR control is a typical control method in the optimal control,which can achieve the desired control effect through the optimal state feedback control.Finally,in the MATLAB/Simulink simulation platform,the classical PID control,fuzzy PID and LQR control based on the RBF-ARX model are compared in the simulation experiment of heading control,the effectiveness of LQR course-keeping strategy based on RBF-ARX model proposed in this paper is verified.