基于高斯核层次聚类的汽车工况构建  

Gaussian Kernel Based Hierarchical Clustering for Driving Cycle Construction

作　　者：韩鑫 HAN Xin(School of Computer Science,Xi'an Shiyou University,Xi'an 710065,China)

机构地区：[1]西安石油大学计算机学院,西安710065

出　　处：《智能计算机与应用》2020年第9期65-68,共4页Intelligent Computer and Applications

摘　　要：现有的车况构建主要采用K-means方法对运动学片段进行聚类,该方法需要通过经验确定聚类的个数,然而人工经验在数据量大和情况复杂时很容易带来误差。因此,本文在对不良数据进行处理、定义怠慢区并对运动学片段进行分割之后,构建基于高斯核的层次聚类算法,对片段进行聚类后确定构建工况的候选集,以解决这个难题。本文还引入统计特征、形状特征、熵特征等共14个运动学片段,作为聚类运动学片段的有效特征。根据运动学片段类别及时间比例,构建了1300 s的工况图。实验结果表明,本文构建的工况图具一定的有效性和实用性。The existing vehicle condition construction mainly uses the k-means method to cluster the kinematic segments.This method needs to determine the number of clusters through experience.However,manual experience is easy to bring errors when the amount of data is large and the situation is complex.Therefore,this paper constructs a hierarchical clustering algorithm based on Gaussian kernel after processing the bad data,defining the slack area and segmenting the kinematic segments,and determines the candidate set of construction conditions after clustering the segments to solve this problem.This paper also introduces 14 kinematic segments,such as statistical feature,shape feature and entropy feature,as the effective features of clustering kinematic segments.According to the category and time proportion of kinematic segments,the 1300 s working condition diagram is constructed.The experimental results show that the working condition diagram constructed in this paper is effective and practical.

关 键 词：汽车工况构建 层次聚类 高斯核 核方法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]