基于改进主成分和全局k均值聚类的汽车行驶工况构建  

Driving condition construction based on improved principal component and global k-means clustering

作　　者：张玉西 苏小会[1] 高广棵 尚煜 ZHANG Yuxi;SU Xiaohui;GAO Guangke;SHANG Yu(School of Computer Science, Xi’an University of Technology, Xi’an 710032, China)

机构地区：[1]西安工业大学计算机科学与工程学院,西安710032

出　　处：《中国科技论文》2020年第11期1253-1259,共7页China Sciencepaper

基　　金：国家地方联合工程实验室基金资助项目(GSYSJ2018012);陕西省教育厅专项科学研究计划项目(17JK0381)。

摘　　要：为解决传统聚类算法构建工况初始中心易陷入局部最优、执行时耗长的问题,提出了一种改进全局k均值聚类(improved global k-means clustering,IGKM)算法,以缩小作为候选下一簇的初始中心点集,减少算法的迭代次数;采用小波分层阈值降噪和小波分解域量化压缩对原始数据进行预处理,结合改进主成分分析(improved principal component analysis,IPCA)对片段进行降维和分类;最后,合成汽车行驶工况。实验结果表明,所提方法构建行驶工况的速度-加速度联合分布差异值仅为0.87%,聚类平均耗时仅为83.35 s,行驶工况拟合度较高,更能综合反映实际车辆的运行状况。In order to solve the problem that the initial center of the traditional clustering algorithm was easy to fall into the local optimum and execution time was too long,an improved global k-means clustering(IGKM)algorithm is proposed to reduce the initial center of the next cluster as a candidate point set,reducing the number of iterations of the algorithm.The original data was preprocessed by wavelet hierarchical threshold denoising and wavelet decomposition domain quantization compression,and the dimension reduction and classification of fragments were performed by improved principal component analysis(IPCA).Finally,car driving process condition was synthesized.The experimental results show that the combined speed-acceleration distribution difference value of the proposed method for constructing driving conditions is only 0.87%,the average time for clustering was only 83.35 s,and the driving conditions has a higher fitting degree,which can more comprehensively reflect the actual running status of vehicle.

关 键 词：行驶工况 改进主成分分析 改进全局k均值聚类 运动学片段 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]