Petri Based Recurrent Fuzzy Neural Control for SY-II Remote Operated Vehicle

作     者：Huang Hai Wan Lei Zhang Guocheng Pang Yongjie 

作者单位：Key Laboratory of Science and Technology for National Defense of Autonomous Underwater Vehicle Harbin Engineering University 

会议名称：《2012年计算机应用与系统建模国际会议》

会议日期：2012年

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 080202[工学-机械电子工程] 081104[工学-模式识别与智能系统] 08[工学] 0804[工学-仪器科学与技术] 0835[工学-软件工程] 0802[工学-机械工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by National High Technology Research and Development Program of China(2008AA092301) Postdoctral Science Foundation of Chinese Hei Long-jing province(323630295) 

关 键 词：Remote Operated Vehicle Petri Network(PN) Recurrent Fuzzy Neural Network(RFNN) 

摘      要：Recurrent fuzzy neural network is widely applied in many areas because it combines the advantages of low level learning and high level reasoning. In considering the complicated factors and requirements for the Remote Operated Vehicle(ROV) control, petri network has been introduced to design a dynamic controller for underwater robot. It intends to reduce the computation burdens during network parameters learning. The gradient descent method has been used for online training. In order to guarantee its convergence, we have used the discrete Lyapunov function to determine its learning rate. The tank experiments have proved that the controller can adjust control quantity to reduce caculation and present strong advantages in the ROV robustness control.