考虑交通状态影响的服务区凝聚层次聚类入区判别模型  

An Identification Model of Entering the Service Area of Agglomerative Hierarchical Clustering Considering the Influence of Traffic Status

作　　者：章玉 张婷婷 姚成北 曹鹏超 ZHANG Yu;ZHANG Tingting;YAO Chengbei;CAO Pengchao(China Railway Changjiang Traffic Design Group Co.,Ltd.,Chongqing 401121,China;School of Traffic&Transportation,Chongqing Jiaotong University,Chongqing 400074,China)

机构地区：[1]中铁长江交通设计集团有限公司,重庆401121 [2]重庆交通大学交通运输学院,重庆400074

出　　处：《重庆交通大学学报（自然科学版）》2024年第3期84-91,共8页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：重庆市科技局面上项目(CSTB2022TIAD-GPX0024)。

摘　　要：为了精准掌握高速公路服务区入区车辆特征、提升服务区运营管理水平,基于高速公路ETC门架通行和收费数据,在分析服务区路段和邻近服务区路段车辆行程时间和速度分布特征基础上,考虑路段交通运行状态影响,提出了基于凝聚层次聚类的运行状态识别方法和服务区分车型入区判别模型。以G65包茂高速大观服务区为例,通过关联上、下游门架路段交通运行状态,明确了服务区路段车辆在4种不同运行状态下的速度概率分布特性,结合聚类给出了各个运行状态下车流密度和速度变化的入区判定条件,并利用服务区视频卡口数据进行验证分析。结果表明:判别误差主要分布在拥堵时段,全日客车和货车在考虑运行状态下的相对误差分别为1.5%、7.0%,与不考虑路段运行状态情况相比分别提高了2.9%、4.1%,验证了模型的有效性,为获取高速公路服务区入区车辆特征提供了一种新的思路。To accurately grasp the characteristics of vehicles entering the expressway service area and improve the operation management level of the service area,the vehicle travel time and speed distribution characteristics of the sections in the service area and adjacent service area were analyzed based on the highway ETC gantry traffic and toll data.Considering the influence of traffic operation state in section,a method for identifying the operation state based on the agglomeration hierarchical clustering and an identification model for judging vehicles entering service area were proposed.Taking the Daguan service area on the G65 Bao-Mao highway as an example,firstly,the speed probability distribution characteristics of vehicles in the service area section under four different operation states were clarified by correlating the traffic states of the upstream and downstream gantry sections.Then,the judging criteria of entering the service area based on the change of traffic density and speed in various operating states were given by clustering algorithm.Finally,the verification and analysis were carried out by using the video bayonet data in the service area.The results show that the identification errors are mainly distributed in the congestion period,and the relative errors of all-day passenger cars and trucks with considering the operation states are 1.5%and 7%respectively,which are 2.9%and 4.1%higher than those without considering the operation states.The effectiveness of the proposed model is verified,which provides a new way to obtain the characteristics of vehicles entering the expressway service area.

关 键 词：交通工程 高速公路服务区 入区车辆判别 凝聚层次聚类 ETC数据 

分 类 号：U412.366[交通运输工程—道路与铁道工程]