基于收费数据的高速公路收费站通行能力分析  

Analysis on Capacity of Expressway Toll Station Based on Toll Data

作　　者：李君羡[1] 周一晨 高志波 沈宙彪 吴志周[1] LI Jun-xian;ZHOU Yi-chen;GAO Zhi-bo;SHEN Zhou-biao;WU Zhi-zhou(Key Laboratory of Road and Traffic Engineering of Ministry of Education,Tongji University,Shanghai 201804,China;Shanghai Naite Expressway Toll Settlement Co.,Ltd.,Shanghai 200063,China;Shanghai Urban Construction Design and Research Institute(Group)Co.,Ltd.,Shanghai 200125,China)

机构地区：[1]同济大学道路与交通工程教育部重点实验室,上海201804 [2]上海耐特高速公路收费结算有限公司,上海200063 [3]上海市城市建设设计研究总院(集团)有限公司,上海200125

出　　处：《公路交通科技》2021年第11期106-116,125,共12页Journal of Highway and Transportation Research and Development

基　　金：国家重点研发计划项目(2018YFB1600805);河北省交通运输厅科技项目(QG2018-15)。

摘　　要：为了研究混合交通流收费站中不同类型收费车道的出入通行能力,采用真实收费过车数据驱动的方法,以克服仿真模型参数标定工作复杂且适用性受限、理论建模涉及诸多理想假设与实际情况存在差距等弊端。首先将通行能力估计问题转化为服务时长分布及其影响因素分析问题;通过合并车辆类型数据保证样本量,再提取服务饱和期间的观测样本群以开展统计研究。确定了样本收费站两种收费模式入、出车道的通行能力;进一步对比了周期复现性、时间序列与分布拟合3种方式,确定收费车道过车类型占比的描述方法。结果表明:不同于收费车道单车服务时长服从负指数分布或正态分布的固有认识,无论是过车总体样本还是按过车类型划分的样本子集,对数正态分布对饱和服务时长分布的拟合效果最优;因收费站功能、驾驶员出行偏好等差异,不同收费站的同类收费车道服务水平明显不同,在实际应用中不宜套用统一标准值,应根据车道实际数据分别统计测算;同一收费站同类收费车道的过车类型占比对通行能力有影响,相对于周期复现性、时间序列以及正态分布,伽马分布对该占比变化描述能力较好,可作为数据不足时按照标准值调整不同车道通行能力的依据。In order to study the capacities of different types of entrance and exit toll lanes in mixed traffic flow toll stations,real toll data-driven are applied to overcome the disadvantages of complex parameter calibration and limited applicability of simulation model as well as theoretical modeling involves many shortcomings such as gaps between ideal assumptions and actual conditions.First,capacity estimation is transformed into the analysis of service duration distribution and its influencing factors.With data of different vehicle types combined to maintain the sample size,and the observation sample group during the service saturation period is extracted for statistical research.The capacities of the entrance and exit lanes of 2 toll modes for sample stations are determined,respectively.Then,the methods of cycle reproducibility,time series and distribution fitting are further compared,and the description method of the proportion of traffic types in toll lanes is determined.The result indicates that(1)Unlike the common understanding that the toll lane's single vehicle service duration follows negative exponential distribution or normal distribution,regardless of whether it is the overall sample of passing vehicles or the subset of samples divided by the vehicle types,the log-normal distribution has the best fitting effect on the saturated service duration distribution.(2)Different toll stations see the same type of toll lanes with different service levels due to the differences in stations’functions and drivers’preferences.Thus,it is inappropriate to apply the unified capacity value suggested by codes but should statistically measure it in practical application.(3)The proportion of vehicles in the same toll lane at the same toll station has an impact on the capacity.Compared with periodic reproducibility,time series and normal distribution,the gamma distribution has a better ability to describe the change in this proportion,and can be used as a basis for adjusting the capacities of different lanes according to th

关 键 词：交通工程 通行能力分析 数据驱动 高速公路收费站 ETC MTC 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]