Real-time Queue Length Detection On Freeways Using Autonomous Vehicle Data

With the acceleration of urbanization in China,the number of cars in cities is increasing year by year and the traffic congestion is becoming more and more serious.With limited transportation infrastructure,congestion has greatly increased the burden of road and caused a series of social problems.Queue length,is the main phenomenon of traffic congestion.Therefore,it is of great importance to estimate the queue length accurately and extract traffic status information in real time,in order to solve the problem of traffic congestion and improve traffic operation efficiency.In recent years,the study on autonomous vehicles(AVs)has been well developed.As an intelligent tool,autonomous vehicle has the ability of environmental perception,real-time information transmission,etc.Therefore,the investigation on the capacity of such intelligent vehicles in detecting traffic congestion will be helpful to explore the methods of congestion mitigation in future traffic modes and improve traffic efficiency.In order to investigate the potential value of using autonomous vehicles in the detection of congestion,this study builds a real-time queue profile estimation methodology on freeways based on data collected by autonomous vehicles.The main contributions are concluded as follow:1.This study introduced the concept,classification and driving technology of AVs in detail.Based on the process of environmental perception,we analyzed the characteristics of data collected by AVs and proved that NGSIMs data set could be used as an alternative data when lacking the real-world data collected by AVs.2.Considering the measurement accuracy of radar,we used a wavelet based denoising algorithm to process the noised vehicle trajectory data and found out the appropriate mother wavelet.We restored the real characteristics of data collected by AVs.3.This study built a real-time shockwave detection and queue profile estimation methodology on freeways based on data collected by autonomous vehicles.The proposed framework consists of four stages:(1)local shockwave(LSW)position detection,(2)LSW speed estimation,(3)grouping of LSWs into a whole shockwave(WSW),(4)WSW speed estimation,(5)queue length estimation in temporal space area based on the shockwave speed.Particular,four methods are proposed to estimate the shockwave speed according to the characteristics of data collected by autonomous vehicles and traffic flow theory with different penetration rate of AVs,which is then used for the measurement of shockwave speed and queue profile.The results demonstrate that the proposed system is applicable for real-time queue profile estimation.The system has the advantages of real-time,high efficiency and accuracy,and is of great significance for analyzing the congestion propagation process of urban traffic.