状态变量合成三级倒立摆模糊神经网络控制  

Control of Triple Inverted Pendulum by Using State Varieties Synthesis Fuzzy Neural Network

作　　者：洪晓英[1] 崔桂梅[1] 秦力舒[1] 王建良[1] 

机构地区：[1]包头钢铁学院自动化系,内蒙古包头014010

出　　处：《控制工程》2004年第1期36-38,共3页Control Engineering of China

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,提出了用状态变量合成模糊神经网络控制三级倒立摆的方法。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,又能适用对状态变量可按性质和类型分类的多变量系统的控制,大大减少了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,鲁棒性强。A method of controlling a triple inverted pendulum is presented in which the state varieties synthesis fuzzy neural network controller is used to solve the rule number explosion problem in multi-variable systems. The method is helpful to control the systems　that have the characteristics of quick speed, serious nonlinearity and absolutely unstable systems. It can reduce the number of rules of fuzzy neural network sharply. Experiment results show the efficiency and robustness of the method.

关 键 词：模糊神经网络 状态变量合成 三级倒立摆 鲁棒性 模糊控制 

分 类 号：TP273.4[自动化与计算机技术—检测技术与自动化装置] TP183[自动化与计算机技术—控制科学与工程]