二级倒立摆的状态变量合成模糊神经网络控制  

Control of double inverted pendulum by using state varieties synthesis fuzzy neural network

作　　者：杨振强 朴营国 程树康[2] 

机构地区：[1]青岛海信集团技术中心,山东青岛266071 [2]哈尔滨工业大学电气工程系,黑龙江哈尔滨150001

出　　处：《控制与决策》2002年第1期123-125,128,共4页Control and Decision

摘　　要：为解决模糊神经网络在控制多变量系统时的规则组合爆炸问题 ,提出用状态变量合成模糊神经网络控制二级倒立摆。该方法既能解决具有快速、强非线性、绝对不稳定系统的控制问题 ,又能对状态变量可按性质和类型分类的多变量系统 ,大大减少模糊神经网络控制器的规则数 ,有利于利用专家的控制经验。实验结果证实了该方法的控制效果好 。A method of controlling a double inverted pendulum is presented in which state varieties synthesis fuzzy neural network controller is used in order to solve the rule number′s explosion problem in multi-variable. It is helpful to control the systems which have the characteristics of quick speed, serious nonlinear and absolutely unstable systems. It can reduce the rules of fuzzy neural network sharply. Experiment results show the effectivity and robustness of the method.

关 键 词：模糊神经网络 二级倒立摆 状态变量合成 工业控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]