基于SVM的二叉树多类分类算法在工程车辆挡位决策中的应用  

Application to Shift Decision for Construction Vehicle Based on SVM Binary Tree Multi-class Classification Algorithm

作　　者：韩顺杰[1] 赵丁选[1] 

机构地区：[1]吉林大学机械科学与工程学院

出　　处：《中国公路学报》2007年第5期122-126,共5页China Journal of Highway and Transport

基　　金：国家自然科学基金项目(59705005)

摘　　要：为了解决工程车辆挡位决策这种典型的多类分类问题,在介绍自动变速器挡位决策基本原理的基础上,提出基于SVM的二叉树多类分类的挡位决策算法,将分类器分布在各个节点上,从而构成了多类SVM,并在制定换挡策略时,考虑了工作油泵所消耗发动机功率的波动问题。试验结果表明:该方法可根据车辆运行状态确定最佳挡位,从而及时准确地满足工程车辆自动换挡要求。In order to solve difficulty in the multi-class classification problem like the shift decision of the construction vehicle, automatic derailleur, authors tree multi-class classification. class SVM, and the undulati when making shift strategy. position accordin transmission for on the basis of introducing basic pri presented shift decision algorithm whic It distributed classifier to every node on problem of consumed power of engi Test results show that the method g to operation states, consequently, it can meet construction vehicle accurately and in time. nciple of shift decision for h was based on SVM binary for constructing the multine oil pump was considered can optimize the gear shift needs of the automatic shift

关 键 词：机械工程 工程车辆 试验研究 挡位决策算法 支持向量机 二叉树 自动变速 

分 类 号：U415.5[交通运输工程—道路与铁道工程]