改进主成分和K-均值聚类算法的行驶工况  

Driving Conditions of a Car Based on Improved Principal Component and K-means Clustering Algorithm

作　　者：张玉西 苏小会[1] 高广棵 尚煜 ZHANG Yu-xi;SU Xiao-hui;GAO Guang-ke;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：《科学技术与工程》2021年第8期3199-3205,共7页Science Technology and Engineering

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

摘　　要：为构建行驶工况,消除K-均值算法对初始聚类中心的敏感性及噪声点的干扰,提出一种改进主成分分析和基于密度的改进K-均值聚类组合方法。结合距离优化法和密度法,构建一种数据集密度度量方法。选取距离较大、密度较高的数据点作为初始聚类中心与候选集,优化聚类结果的同时剔除了孤立点,采用较大贡献因子的特征值进行工况合成,最后对行驶工况油耗进行分析。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为1.17%,特征参数平均相对误差较小。可见,合成的行驶工况能够很好地反映某地实际交通道路特征,拟合度较高。The driving condition of a car is also called the operating cycle or driving cycle,which reflects the speed-time variation of the vehicle in a specific environment.To construct driving conditions and eliminate the interference of K-means algorithm on initial clustering center sensitivity and noise points,an improved principal component analysis and an improved density-based K-means clustering method were proposed.Combining distance optimization method and density method,a data set density measurement method was constructed.Data points with a larger distance and a higher density were selected as the initial clustering center and candidate set,and the outliers were eliminated while optimizing the clustering results.The eigenvalues of larger contribution factors were used to synthesize the operating conditions,and finally,the fuel consumption of the driving conditions was analyzed.Results show that the joint difference of speed-acceleration distribution of the driving conditions is 1.17%,and the average relative error of the characteristic parameters is small.Therefore,the synthesized driving conditions can well reflect the characteristics of the actual traffic in a certain place with high degree of fitting.

关 键 词：改进主成分分析 改进K-均值聚类 距离优化法 密度法 

分 类 号：TP391.75[自动化与计算机技术—计算机应用技术]