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

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。

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

In order to solve difficulty in the multi-class classification problem like the shift decision of the construction vehicle,on the basis of introducing basic principle of shift decision for automatic derailleur,authors presented shift decision algorithm which was based on SVM binary tree multi-class classification.It distributed classifier to every node for constructing the multi-class SVM,and the undulation problem of consumed power of engine oil pump was considered when making shift strategy.Test results show that the method can optimize the gear shift position according to operation states,consequently,it can meet the needs of the automatic shift transmission for construction vehicle accurately and in time.