The Establishment Of Non-linear Ship Model And The Design Of ANFIS Ship Course Keeping Controller

Adaptive-Network-based Fuzzy Inference System (ANFIS) and its application on the ship course keeping control are studied in this thesis. According to the research of non-linear model of ships, the simulation module of the non-linear model with wind and wave interference of ships is realized by writing S-functions. Base on the research of the Theory of Fuzzy Controller, fuzzy ship course control system is designed. Adaptive-Network-based Fuzzy Inference System is studied, and the ANFIS ship course keeping control system is designed.Firstly the non-linear model of ships is analyzed, and the interference of wind and wave and the method of simulation used S-function are introduced. The S-function of non-linear model with wind and wave interference of ships is written, and a compact and convenient platform is offered for the simulate test of the control of ships.Secondly according to Fuzzy Control Theory, the process of realizing Fuzzy controller is introduced and the Fuzzy ship course keeping controller is designed by analyzing the typical Fuzzy ship course keeping control system.Thirdly the arithmetic and principle of ANFIS is introduced particularly, and the process of realizing ANFIS controller by using MATLAB is discussed. The training data of ANFIS ship course keeping control system is gained by designing PID ship course control system. ANFIS ship course keeping control system is designed by taking nonlinear ship model interfered with wind and wave as the object being controlled. At the same time, it has a good simulation effect to simulate full loaded and semi-loaded respectively. Therefore, the robustness of ANFIS system has been verified from two aspects, outsides interference and changes of ship's parameters. It can be found that the control effect of ANFIS system is better than that of PID system and ANFIS system has better self-learning and self-adaptive from the results of simulation.