基于PCA和FCM的汽车行驶工况研究与构建

【作者】 高建平； 丁伟； 孙中博； 郗建国；

【机构】 河南科技大学；

【摘要】 运用K-means聚类构建工况时,K-means聚类是一种硬划分,但是运动学片段之间的联系是模糊的,故本研究采用模糊聚类对运动学片段进行聚类分析。以郑州市乘用车实际行驶工况为研究对象,运用MATLAB编程对共采集到的613792条行驶数据进行了运动学片段划分、主成分分析(PCA)和模糊C均值聚类(FCM)分析,构建了郑州市乘用车行驶工况,从特征参数误差、速度-加速度联合分布、两个方面验证了构建的工况的合理性,并与国内外相关标准工况进行了特征参数的对比,结果表明:郑州工况有自己的特点,其他工况不能完全反应郑州市实际交通,建议该地区使用此工况作为测试乘用车污染物排放法规。更多还原

【Abstract】 Constructing conditions by using k-means clustering and K-means clustering is a hard division,but kinematics sequences of different characteristic parameters between different,has the characteristics of fuzziness.Therefore this study using fuzzy clustering,clustering analysis of kinematics sequences.Using MATLAB programming to carry on the kinematics analysis of the 613792 driving cycle which are collected,the principal component analysis(PCA) and the fuzzy C mean clustering(FCM) analysis,Construction of Zhengzhou City passenger car driving cycle and the rationality of the construction condition is verified from two aspects:characteristic parameter error,velocity acceleration joint distribution,the characteristic parameters are compared with the relevant standard condition at home and abroad,the results show that Zhengzhou driving cycle has its own characteristics,the other conditions can not fully reflect the actual traffic in Zhengzhou,it is recommended that the area use this condition as a test of passenger vehicle emission regulations.更多还原