基于车载数据k均值聚类的特种车辆行驶工况识别  

Identification of onboard data driving condition for special vehicles based on k-means clustering algorithm

作　　者：赵津 王立勇[1] 张金乐[2] ZHAO Jin;WANG Liyong;ZHANG Jinle(Mechanical Electrical Engineering School,Beijing Information Science&Technology University,Beijing 100192,China;China North Vehicle Research Institute,Beijing 100072,China)

机构地区：[1]北京信息科技大学机电工程学院,北京100192 [2]中国北方车辆研究所,北京100072

出　　处：《北京信息科技大学学报（自然科学版）》2023年第2期39-46,共8页Journal of Beijing Information Science and Technology University

基　　金：国家“173”计划项目(MKF20210009)。

摘　　要：运行工况的合理划分是车辆可靠性和耐久性研究的基础。为了综合考虑特种车辆动力性能和操纵性能,针对换挡、运行、爬坡和转向4种工况进行研究,并提出各工况特征值计算方法。为提高车辆行驶工况识别的准确性,通过小波阈值滤波算法和自顶向下分段线性表示算法对原始数据进行去噪和短行程划分,再利用k均值聚类算法对车辆行驶工况进行识别。通过某型特种车辆试验数据验证,该方法能够对行驶工况有效识别,工况聚类精度可达92.75%。The reasonable division of operating conditions is the basis for the study of vehicle reliability and durability.To comprehensively consider the power performance and handling performance of special vehicles,four operating conditions including gear shifting,operation,climbing and steering were studied,and the calculation method of eigenvalue of each operating condition was proposed.In order to improve the accuracy of vehicle driving condition identification,wavelet threshold filtering algorithm and top-down piecewise linear representation algorithm were used to denoise and divide the original data into short journeys,and k-means clustering algorithm was used to identify vehicle driving conditions.The test results of a special vehicle show that the method can effectively identify driving conditions,and the clustering accuracy can reach 92.75%.

关 键 词：特征计算 分段线性表示算法 K均值聚类 行驶工况识别 

分 类 号：U469.74[机械工程—车辆工程]