Lugre摩擦模型的模糊神经网络辨识仿真研究  

Simulation of Fuzzy Neural Network Identification with Lugre friction model

作　　者：刘海荣 刘金琨[2] 

机构地区：[1]大同电力技工学校计算机室,山西大同037039 [2]北京航空航天大学自动控制系,北京100083

出　　处：《计算机仿真》2007年第1期80-82,共3页Computer Simulation

基　　金：航空科学基金资助项目(00E51022)

摘　　要：LuG re模型是典型伺服系统的摩擦模型,该模型能够准确地描述伺服系统摩擦过程复杂的动静态特性。模糊神经网络已成为模糊逻辑和神经网络研究最前沿的课题之一。模糊神经网络同时具有神经网络和模糊逻辑的优点。针对伺服系统所面临的摩擦问题,以低速伺服系统为对象,建立伺服系统的LuG re摩擦模型,采用模糊神经网络进行非线性在线辨识。仿真结果表明,采用模糊神经方法建立对非线性对象的辨识器,具有较高的辨识精度。Lugre friction is a typical friction model of servo system, which is used to describe dynamic and static characteristics of friction phenomenon in servo system. Fuzzy neural network has advantages of both neural network and fuzzy logic control, which has become a promising area in these years. The LuGre friction model is set up according to the friction of low speed servo system. Nonlinear online identification is realized by adopting fuzzy neural network. Simulation indicates high precision identification can be obtained by using fuzzy neural identifier.

关 键 词：摩擦模型 模糊控制 神经网络 辨识 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]