基于聚类算法的车辆数据挖掘及可视化研究

Research on Vehicle Data Mining and Visualization Based on Clustering Algorithm

车辆运行工况决定了承载人员的生命安全,影响了整车工作效率,同时也是分析驾驶人员行为及构建UBI保险业务的数据基础。基于车联网技术,为构建车辆及周边环境的评价系统,以50辆不同型号私家车作为研究对象,在固定路段及时间段内,利用车载OBD终端采样车辆行驶数据。对各车辆的行驶速度、进气量、瞬时油耗、发动机转速等运行数据进行汇总与清洗,利用k-means聚类算法对样本数据进行了分布计算,并通过最小二乘拟合获得发动机各数据间运行曲线,与理论模型相一致,实现对车、人、路况的客观评价。进一步开发了车辆数据信息管理系统平台,经TCP/IP协议获取车载终端GPS数据,实现了行驶轨迹及车况数据的动态显示,方便了对车辆的跟踪与查询。该研究工作可为构建驾驶行为及UBI保险营销模型提供理论依据,同时为车联网的功能完善提供参考。

The operating condition of the vehicle determines the life safety of the carrying personnel,affect the work efficiency of the vehicle,and serve as the data basis for analyzing the behavior of the driving personnel and constructing UBI insurance business.Based on the Internet of vehicles technology,in order to construct the evaluation system of vehicles and surrounding environment,fifty private cars of the different model are taken as research objects,and vehicle condition data obtained from vehicle OBD terminal are sampled in selected road sections and fixed periods.We summarize and clean the driving speed,air intake,instantaneous fuel consumption,engine speed and other operation data of each vehicle,use the k-means clustering algorithm to calculate the distribution of sample data and the least square fitting to obtain the running curve between the engine data,which is consistent with the theoretical model,so as to realize the evaluation of vehicle,human and road conditions.Furthermore,the upper computer platform of vehicle information management system is developed,and the vehicle terminal GPS positioning is obtained through TCP/IP protocol to realize the dynamic display of vehicle trajectory and vehicle condition data,which facilitates the tracking and query of vehicles.This study can provide theoretical basis for the construction of driving behavior and UBI insurance marketing model,as well as provide reference for the improvement of functions of Internet of vehicles.