基于单截面低频检测数据的信号交叉口排队长度估计

Estimation of Queuing Length at Signalized Intersections Using Low-frequency Point Detector Data

针对我国大多数中小城市信号交叉口交通检测数据条件及基于此数据条件下存在的信号交叉口排队长度估计精度不高问题,研究了基于单截面低频定点检测数据的信号交叉口排队长度估计模型。利用时间占有率与流量、速度之间的函数关系对长排队(排队长度超出检测器位置)进行识别。根据信号配时数据切分低频检测器数据,并与信号配时数据匹配。基于交通波理论,通过关键点的判别求取周期最大排队长度。以青岛市山东路-江西路南进口为例进行仿真和实证验证。结果显示,长排队的识别精度达到了90%以上,不同饱和度下(低、中、高)的信号交叉口排队长度估计精度均达到了80%以上,其中,中、低饱和度场景下排队长度平均绝对误差小于20m/cycle,高饱和度场景下排队长度平均绝对误差小于45m/cycle。

Due to the limitation of traffic data collected at signalized intersections in most small and medium sized cities in China,estimation on queuing length at signalized intersections is with low accuracy.In order to solve this problem,a model of estimating queue length based on data from low frequency,fixed-point,and single detectors at signalized intersections is proposed.A long queuing event(the end of a queue exceeds the position of the detector)is identified based on the relationship between occupancy,flow,and speed.Data of detectors is divided into cycles according to and synchronized with signal timing.The maximum length of queuing is calculated based on shock wave theory.A simulation test and a demonstration study are carried out at the south entrance lanes of Shandong road and Jiangxi road in Qingdao.The result shows that identification accuracy of a long queue is above 90%.The average estimated accuracies of queuing length under different saturation(low,medium,and high)are all more than 80%,among which,the average MAE is lower than 20 m per cycle under medium and low saturation,and lower than 45 m per cycle under high saturation.