基于核主元约简与半监督核模糊聚类的车辆行驶工况判别  

Driving Cycle Distinguishing Based on the Kernel Principal Component and Semi-supervised Kernel Fuzzy C Means Clustering Algorithm

作　　者：张袅娜[1,2] 丁海涛[2] 于海芳[1] 刘姝阳[1] 

机构地区：[1]长春工业大学电气与电子工程学院,长春130012 [2]吉林大学汽车仿真与控制国家重点实验室,长春130012

出　　处：《机械工程学报》2015年第2期96-102,共7页Journal of Mechanical Engineering

基　　金：国家重点基础研究发展计划(973计划;2011CB711205);国家高技术研究发展计划(863计划;2011AA11A221);吉林省科技支撑计划重大专项(20126008)资助项目

摘　　要：为有效利用汽车行驶工况中与类别属性间的统计特征,提高汽车行驶工况判别的准确性与快速性,首先选择车速和踏板信号的数据信息构建特征集,利用相关性分析和核主元分析对特征集中能敏感反映工况类别的特征数据信息进行二次特征提取,按累计贡献率大于90%的标准进行主要特征量的选择,实现输入变量的二次约简;利用小波核函数的非线性映射能力构建半监督核模糊C均值聚类方法进行车辆行驶工况的判别。通过长春某混合动力公交车试验结果表明,该方法更全面准确地反映了工况特性,有效降低了输入特征参数的维数,更加准确有效地提取了不同工况条件下汽车行驶状态的数据特征,通过半监督核模糊C均值聚类算法中加入少量的训练样本来引导聚类过程,聚类精度可达到98.75%。In order to utilize the statistic characteristics among the information that associated with categories in vehicle driving cycle effectively, the speed and the pedal signal data information are selected to construct feature set to improve the accuracy and rapidity of distinguishing the vehicle driving cycle firstly. Then, to realize the second reduction of the input variables, second feature which sensitively reflects characteristic data information of driving cycle category is extracted through the correlation analysis and the kernel principal component analysis. According to the standard of accumulation contribution rate that greater than 90%, the main feature is chosen to achieve secondary reduction of input variables. The nonlinear mapping ability of the wavelet kernel function is adopted to establish semi-supervised kernel fuzzy C means clustering algorithm for the working condition judgments. The experimental results show that this algorithm can reflect the driving cycle characteristics more comprehensively and accurately. In addition, the dimension of the input characteristic parameters is reduced effectively. And the data characteristics of vehicle driving condition are extracted under different working conditions more accurately and availably. Simultaneously, the clustering accuracy can reach 98.75% by adding a small amount of training samples in semi-supervised kernel fuzzy C means clustering algorithm to guide the clustering process.

关 键 词：核主元 半监督 核模糊C均值聚类 行驶工况判别 

分 类 号：U469[机械工程—车辆工程]