Method For Estimation Queue Length At Intersection Based On Data Fusion

At present,the mainstream research on queue length estimation at home and abroad mostly obtains the maximum queue length and average queue length in the signal period of the road intersection according to the traffic flow state.Existing research relies on a single source of traffic data,such as GPS floating car data,traffic data for cross-section detection,and so on.The stability and reliability of the queue length estimation algorithm that relies on a single data source is difficult to guarantee due to detection errors,signal loss,interference,and coverage limitations of the detection device during data detection and transmission.Therefore,the algorithm research of merging the multi-source data for queuing length estimation has good research value in data complementation,precision correction and improvement of the comprehensive performance of the estimation model.This paper focuses on traffic data fusion and intersection length estimation method,which is divided into three parts: traffic data fusion and queue length estimation method;data fusion intersection length estimation method accuracy analysis;based on queue length signal timing Explore.Firstly,this paper analyzes the characteristics of different traffic data,and summarizes the attribution types of traffic data,the classification of traffic data collection technology and the characteristics of different types of detection equipment.The applicable conditions,advantages and disadvantages of various methods are analyzed and compared.Then based on the statistical induction theory and traffic flow theory,combined with the queuing law of the floating car and the characteristics of the cross-section detection flow data,the queuing length estimation model is established respectively.Considering the difference in accuracy between the two types of models in different data acquisition environments,in order to enhance the universality and stability of the estimation model,a combined estimation method is proposed to perform error fusion correction,and the cross-section detection flow data and mobile GPS data are used.Effective fusion,and finally construct an estimation model based on data fusion that can estimate the queue length of intersection vehicles in real time.Secondly,using the fixed detector detection data of the intersection of Zhongshan Road and Bayi Road in Donghu District of Nanchang City,the VISSIM software is used for queuing simulation analysis to obtain the real-time floating car data under the same flow and the same according to the mobile GPS data.And the cross-section detection flow data respectively estimate the queue length of the intersections in each time interval,according to the combined estimation theory of the entropy method,weighted combination of the queuing length estimation result based on the mobile GPS data and the queuing length estimationresult based on the cross-section detection flow data.It is estimated that the actual queue length result of the simulation output is used to analyze and evaluate the error index of the model calculation accuracy.Finally,the queuing length estimation model based on data fusion is compared with the queuing length calculation method of single data source.The overall estimation stability is better and the calculation result is more accurate.Finally,based on the estimated queue length,the method of determining the timing of intersection signal based on the estimated queue length is discussed.The basic process of signal timing under the continuous data acquisition conditions at intersection is analyzed,and the signal based on queue length equalization as control target is summarized.The main idea of the timing strategy.