| With the continuous development of the economy,the number of the vehicles has increased dramatically,and the occurrence of urban traffic congestion has become more frequently.Urban traffic congestion has become one of the important reasons that hinder the continued development of major cities,and it is a common social problem which the world faces today.In order to solve the problem of urban traffic congestion better,it is more urgent to build an urban intelligent transportation system However,the traffic conditions estimation is the premise of build the intelligent transportation system.Therefore,the real-time estimation of urban traffic conditions has strong theoretical value and practical significance.This paper studies the urban traffic conditions estimation based on the compressed sensing methods.The main research contents are summarized as follows:Firstly,this paper systematically introduces the research background and significance of urban traffic conditions estimation,and outlines the research and characteristics of the technology at home and abroad.Next,from the basic theory of compressed sensing,the related principles of sparse representation matrix,measurement matrix and signal reconstruction are introduced.And then,the deep learning method based on sparse coding and its basic theory are introduced,followed by the basic principles of map matching and related algorithms.Secondly,aiming at how to use less vehicle trajectory data to achieve accurate estimation of urban traffic conditions,this paper proposes a traffic conditions estimation method based on compressed sensing.This method makes sparse coding techniques into compressive sensing methods to achieve more efficient sparse representation of traffic conditions information.In this paper,through the matched vehicles trajectory data,two indicators to measure the traffic conditions are calculated—the average speed of the road section and the congestion rate of the road sections.The sparse coding method is used to train the above two indicators,thereby obtaining a corresponding sparse representation matrix.In addition,the corresponding distance-type measurement matrix and time-type measurement matrix are also designed to realize the traffic conditions estimation method based on compressed sensing.The experimental results show that the sparse coding method can achieve a more effective sparse representation of traffic conditions information at adjacent moments,and it can obtain loower estimation error.Thirdly,in order to further reduce the estimation error of compressed sensing method for urban traffic conditions estimation,this paper combines the topology of road network and Gaussian joint distribution model,and a sparse representation matrix based on Gaussian kernel function to improve the sparse performance of traffic conditions information is proposed.The method structures the measurement matrix based on a sparse type by using a part of a probe vehicle,thereby reducing the number of probe vehicles.Finally,the sparse representation matrix and measurement of this method are analyzed,and it satisfies the double non-incoherence principle in compressed sensing theory.The simulation results of real data show that this method can reduce the reconstruction error of traffic conditio ns information and achieve more effective urban traffic state estimation. |
