基于车辆行驶轨迹的信号交叉口排队长度估计  

Estimation of Queue Length at Signalized Intersection Based on Vehicle Trajectory

作　　者：姚佼[1] 戴亚轩 倪屹聆 韦钰 YAO Jiao;DAI Ya-xuan;NI Yi-ling;WEI Yu(School of Business,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学管理学院,上海200093

出　　处：《公路交通科技》2020年第5期123-130,共8页Journal of Highway and Transportation Research and Development

基　　金：教育部人文社会科学研究青年基金项目(17YJCZH225);上海市青年科技英才扬帆计划项目(19YF1435100)。

摘　　要：为了研究如何结合移动检测数据来确定交叉口排队长度,并以此来衡量交通拥堵程度的问题,利用车辆行驶轨迹,分析了通过交叉口车辆的排队特点。根据车辆在队列中的不同排队位置,分车辆通过交叉口时所存在的A,B,C这3种位置,建立了面向延误最小的排队长度估计模型。其中,通过虚拟线圈检测器后开始减速停止在停车线前的A位置车辆排队估计模型基于基本延误模型;减速进入虚拟线圈检测区域停车的B位置车辆排队估计模型基于简化车辆跟驰模型,对可获得车辆行驶轨迹的网联车减速过程进行了重建;减速停止在虚拟线圈检测器前的C位置车辆排队估计模型基于LWR消散模型以及交通流理论算法,并利用网联车车辆行驶轨迹数据进行了加速过程的重建。在此基础上,根据不同位置车辆与队尾网联车的距离不同,对其到达率赋予不同的权重,计算总的排队长度。最后,通过图新地球地图软件投影并筛选车辆在案例交叉口的车辆行驶轨迹,利用微观交通仿真软件VISSIM对本研究的模型进行仿真验证。结果表明,排队长度估计模型与真值的最大误差为12.4%,最小为2.2%,平均误差为8.75%,方差为12.595%~2,绝对与相对误差均保持在可接受范围以内,说明基于车辆行驶轨迹的信号交叉口排队长度估计模型能够较为有效地估计城市道路交叉口的排队长度。In order to study how to combine movement detection data to determine the queue length at intersection and to measure the degree of traffic congestion, the vehicle queue characteristics at intersection is analysed by using vehicle trajectory. According to the different queue positions of the vehicle in the queue and the stop positions of A, B and C when vehicle passes the intersection, the model of queue length estimation with minimum delay is established. The vehicle queuing estimation model of position A, which starts to decelerate and stops in front of the parking line after passing the virtual loop detector, is based on the basic delay model. The vehicle queuing estimation model of position B, which decelerates into the virtual loop detection area and stops, is based on the simplified car following model, and the deceleration process of the networked vehicle that can obtain the vehicle trajectory is reconstructed. The vehicle queuing estimation model of position C, which decelerates to stop in front of the virtual loop detector, is based on the LWR dissipation model and traffic flow theory algorithm, and the acceleration process is reconstructed by using the networked vehicle trajectory data. On this basis, according to the distance between vehicles at different positions and connected vehicle at the queue end, different weights are given to their arrival rates to calculate the total queue length. Finally, the vehicles are projected by map software and the vehicle trajectories at the case intersection are screened, and the proposed model is verified by the microscopic traffic simulation software VISSIM. The result shows that the maximum error between the value obtained by the queuing length estimation model and the true value is 12.4%, the minimum error is 2.2%, the average error is 8.75%, and the variance is 12.595%~2. The absolute and relative errors are kept within the acceptable ranges, indicating that the estimation model of queue length at signalized intersection based on vehicle trajectory can estimate

关 键 词：城市交通 排队长度估计模型 交通流理论算法 车辆排队位置 车辆行驶轨迹 虚拟线圈检测器 

分 类 号：U491.54[交通运输工程—交通运输规划与管理]