路段行程时间估计的浮动车数据挖掘方法

发布时间：2018-04-09 16:04

本文选题：智能交通系统　切入点：路段行程时间估计　出处：《交通运输工程学报》2014年06期

【摘要】：基于浮动车数据,提出一种信号配时信息缺失下的路段行程时间估计方法,由交叉口范围动态划分、路段影响范围划分、浮动车数据提取与路段行程时间估计4个模块组成,每个模块的实现均需借助于前一模块的输出。根据交叉口信号控制下的车辆行驶状态,在交叉口范围动态划分与路段影响范围划分2个模块中,利用密度法将单元路段划分为不同区域。根据路段行程时间估计原理,利用浮动车数据提取模块过滤掉受信号控制影响较大的浮动车数据,提取路段行程时间估计的目标数据。利用路段行程时间估计模块挖掘历史浮动车数据,根据浮动车目标数据点存在区域的不同,将浮动车数据分为3类,并对不同类型数据采取相应的断面通过时刻估计方法,建立基于不同数据条件下的行程时间估计模型。利用VISSIM软件对路段行程时间估计方法进行仿真验证,并与直接法和间接法进行对比分析。分析结果表明:对于粗粒度浮动车数据,路段行程时间估计方法的平均绝对误差和平均相对误差分别为12s和8.67%,优于传统的直接法与间接法。
[Abstract]:Based on the floating vehicle data, this paper presents a method for estimating the travel time of the road under the absence of signal timing information. The method consists of four modules: the dynamic division of the intersection range, the division of the section of influence, the extraction of the floating vehicle data and the estimation of the travel time of the section.Each module needs to be implemented with the output of the previous module.According to the vehicle driving state under the control of intersection signal, the unit sections are divided into different areas by density method in the two modules of dynamic division of intersections and impact areas of road sections.According to the principle of segment travel time estimation, the floating vehicle data extraction module is used to filter out the floating vehicle data which is greatly affected by the signal control, and the target data of the section travel time estimation is extracted.Using the section travel time estimation module to mine the historical floating vehicle data, according to the different region of the floating vehicle target data point, the floating vehicle data is divided into three categories, and the corresponding section passing time estimation method is adopted for the different type data.A travel time estimation model based on different data is established.The method of estimating the travel time of road section is simulated by VISSIM software, and compared with the direct method and indirect method.The results show that the average absolute error and average relative error of the method are 12s and 8.67s respectively for the coarse particle floating vehicle data, which is superior to the traditional direct method and indirect method.
【作者单位】：上海海事大学交通运输学院;同济大学交通运输工程学院;
【基金】：国家自然科学基金项目(61304203) 上海市科研计划项目(12ZR1444800)
【分类号】：U491.1

【参考文献】