K-means算法在高速公路ETC数据分析中的应用  

Application of K-means Algorithm in Expressway ETC Data Analysis

作　　者：张添翼 杨涵 田俊山 王歆远 ZHANG Tian-yi;YANG Han;TIAN Jun-shan;WANG Xin-yuan(Research Institute of Highway,Ministry of Transport,Beijing 100088,China;Fujian Expressway Science and Technology Innovation Research Institute Co.,Ltd.,Fuzhou,Fujian 350001,China)

机构地区：[1]交通运输部公路科学研究院,北京100088 [2]福建省高速公路科技创新研究院有限公司,福建福州350001

出　　处：《公路交通科技》2024年第6期199-206,共8页Journal of Highway and Transportation Research and Development

摘　　要：为了更高效地利用高速公路ETC数据集并提升数据处理速度,深入分析ETC用户的主要特征和高速公路存在的潜在问题。以我国某省份某高速公路出入口2023年6月的ETC通行数据为例,通过Python编程语言对数据进行清洗,采用环形特征编码处理时间数据,并运用K-means聚类算法对数据进行处理。重点关注入口时间、出口时间、本省通行里程等指标,对用户的收费里程、速度以及行驶时间3个核心特征进行分析,借助聚类中心点和雷达图进行可视化展示。分析结果显示,傍晚时段的通行效率较低,晚间疲劳驾驶和午夜超速问题较为突出。根据通行里程分析,白天主要以短程和中程用户为主,长程用户倾向于在上午进入高速公路,同时,该高速公路存在大量的通勤车辆。在速度分析方面,低速组多为短途车辆。K-means聚类算法的应用使得数据处理过程快速且可靠,结合更多的ETC数据,可以进一步深入了解高速公路通行的主要群体和状况。研究成果可为制定差异化收费政策提供有力依据。例如,通过聚类分析进入高速公路的时间,确定高峰时段和低谷时段,适时提高高峰时段的费用,降低低谷时段的费用,从而提高通行效率、平衡路网流量。这具有重要的现实意义。To utilize the expressway ETC dataset more efficiently and to improve the data processing speed,the main characteristics of ETC users and the potential problems of expressway were analyzed in depth.Taking the ETC passage data of expressway entrance/exit in a province of China in June 2023 for an example,the data were cleaned by Python programming language.The time data were processed by using ring feature coding,and the K-means clustering algorithm was applied to process the data.The indicators(e.g.,entrance time,exit time,and mileage of passage in the province)were mainly focused.The 3 core features(i.e.,users’toll mileage,speed,and driving time)were analyzed,and visualized with the assist of clustering centroids and radar charts.The result indicates that the passage efficiency is lower during evening time.The problems of fatigue driving in the evening and speeding in the midnight are more prominent.According to the mileage analysis,the daytime is mainly dominated by short-distance and medium-distance users,and the long-distance users tend to enter the expressway in the morning,while there is a large number of commuter vehicles on the expressway.In terms of speed analysis,the low-speed group is mostly short-distance vehicles.The application of K-means clustering algorithm makes the data processing process fast and reliable.Combining with more ETC data,it can provide further insights into the main groups and conditions of expressway access.The study result can provide a strong basis for the development of differentiated toll policies,e.g.,analyzing the time of entering expressway through clustering,determining the peak time and trough time,increasing the fee in peak time and reducing in trough time.It can improve the access efficiency and balance the traffic flow of road network.

关 键 词：智能交通 用户聚类 K-MEANS算法 高速公路ETC数据 海量数据 

分 类 号：U495[交通运输工程—交通运输规划与管理]