Study On Queue Length Estimation At Signalized Intersection Using Bus GPS Data

The queue length at signalized intersection is the key to optimize the signal timing parameters and to evaluate the traffic efficiency of the intersection.Constrained by traditional detection methods,the existing queue length estimation researches based on traffic wave theory do not analyze the queuing process of vehicles in single signal cycle,so,it is difficult to characterize the real-time queue length of vehicles at signalized intersections.With the development of modern traffic information collection technology,to analyze and study on the queuing process and departure process at the signalized intersection based on the new information collection method,and then to optimize the real-time queue length estimation,is of theoretical and practical significance.This paper analyzes the queuing process and departure process of vehicles at signalized intersection.Based on the estimation of bus penetration rate and vehicle arrival rate at signalized intersections,and combined with the characteristics of bus GPS data,a queue length estimation method based on bus GPS data is established to estimate the real-time queue length,average queue length and maximum queue length at signalized intersection.The main contents are as following.(1)The estimation of bus penetration rate at road section using bus GPS data.The existing floating car penetration estimation methods assume that the arrival of vehicles obeys a specific distribution,and the simulation experiment conditions are too ideal to apply to the actual traffic environment.So,this article analyzes the location of vehicles in the queue at signalized intersection and combines the characteristics of the bus GPS data,to propose a bus penetration estimation method using bus GPS data.In the simulation environment,different flow rates and penetration rate were set to carry out simulation tests,and typical road sections in Chongqing were selected for field test.The results show that the estimation method in this paper is applicable and effective.(2)Vehicle arrival rate estimation at signalized intersection.Considering that there may be differences in the arrival process of different types of signalized intersections,this paper divided the signalized intersections into independent signalized intersections and associated signalized intersections according to the degree of correlation between the upstream and downstream signalized intersections.Based on the analysis of the arrivals process of vehicles at the two types of intersections,two models of vehicle arrival rate estimation at signalized intersections are established.Finally,a simulation experiment was carried out on the VISSIM platform.The results show that the vehicle arrival rate estimation models for both types of signalized intersections have better applicability.(3)Queue length estimation method at signalized intersections.By analyzing the departure process of vehicles at the signalized intersection,the departure characteristics of the vehicles at the signalized intersection are obtained.On the basis,the vehicle's saturation flow rate and delay time of vehicles at the start time of the green light are calculated.Combined with the estimation results of penetration rate and vehicle arrival rate,the real-time queue length,average queue length and maximum queue length estimation model at signalized intersections are established.The field test results at a typical signalized intersection in Chongqing show that the proposed method has a better performance than the comparison estimation method and has better applicability and effectiveness.In all,the queue length estimation method at signalized intersections using bus GPS data proposed in this paper has a better performance,and can provide services to signalized intersection timing optimization,traffic induction,and route decision making.