首页 | 本学科首页   官方微博 | 高级检索  
     检索      

多层前向神经网络在电动转向系统中的应用
引用本文:张海华,许镇琳,王豪,尚喆.多层前向神经网络在电动转向系统中的应用[J].公路交通科技,2004,21(8):121-123.
作者姓名:张海华  许镇琳  王豪  尚喆
作者单位:天津大学电气与自动化工程学院,天津,300072
摘    要:针对电动转向系统助力特性存在非线性的特点,应用多层前向神经网络,对电动转向系统全车速下的助力特性进行拟合。文中设计了多层前向神经网络结构,并采用基于综合目标函数的二阶学习算法对其进行离线学习,最后在电动转向综合试验台上实现在线控制。系统实现了全车速范围的非线性转向助力,克服了转向助力盲区。

关 键 词:电动转向  助力特性  多层前向神经网络
文章编号:1002-0268(2004)08-0121-03
修稿时间:2003年6月15日

Application of Multilayer Foreword Neural Network on the Electric Power Steering System
ZHANG Hai-hua,XU Zhen-lin,WANG Hao,SHANG Zhe.Application of Multilayer Foreword Neural Network on the Electric Power Steering System[J].Journal of Highway and Transportation Research and Development,2004,21(8):121-123.
Authors:ZHANG Hai-hua  XU Zhen-lin  WANG Hao  SHANG Zhe
Abstract:Considering the nonlinear characteristic of electric power steering (EPS) system's power control characteristic curve,Multilayer Foreword Neural Network is applied to fitting the power conlrol characteristic curve at any speed in this paper.The Multilayer Foreword Neural Network structure is designed.And the second-order learning algorithm,which is based on generalized cost criterion,is adopted to do off-line learning.Finally,the on-line control is realized on the EPS experiment rig.Then the full-speed nonlinear steering power control curves are generated and the blind zone is overcome.
Keywords:Electric power steering (EPS)  Power control characteristic  Multilayer foreword neural network
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号