Analysis Of Urban Regional Traffic States Based On Floating Car GPS Data

With the rapid development of social economy,motor vehicles have become one of the main means of transportation in tourism.However,the rapid increase in motor vehicles also puts tremendous pressure on urban transport system and leads to more frequent traffic congestion.Intelligent transportation system(ITS)is an effective way to reduce traffic jams.The GPS floating car can provide real-time traffic condition information for intelligent traffic control and traffic guidance.Through real-time traffic GPS data collection and analysis,we can not only obtain real-time traffic conditions of the road network,but also accurately find the congestion area and obtain the road network situation,thus provide guidance and control strategies for the urban traffic system.The purpose of this research is to discover and master the changes of each sub-area in the urban traffic network in time through the processing of GPS data.Analysis of urban road network conditions requires real-time traffic flow data of all road segments.Due to the sparse and incomplete data caused by different real-time coverage of GPS data,we propose a feasible solution to deal with these problems.At the same time,the paper also gives a new solution to the problem of initial parameters sensitivity of the road network subareas division under traditional network classification method,as well as the instability and inaccuracy of the division effect.We run several experiments validating our method.The main work is as follows:(1)Firstly,the paper analyzes and studies the sample size of urban taxi GPS data.By using the minimum sample analysis method,we found that the number of existing taxi samples in Xi’an can fully reflect the overall traffic conditions in Xi’an.The paper also discusses the feasibility of analyzing traffic jams with floating car data.The preprocessing of eliminating redundant data,invalid data,and noise data for GPS data error problems are also discussed.(2)Due to the lack of road real-time network coverage strength,the available GPS data are still problematic and incomplete.According to the propagation law of traffic flow,the paper uses an improved weighted PageRank algorithm to smooth and correct the road network weight matrix of incomplete information.This method effectively solves the data sparsity problem and improves the accuracy of the data.(3)Aim to solve the sensitivity and instability of the initial value of the traditional network partition algorithm,a new road network segmentation algorithm based on AP clustering algorithm is proposed.The AP clustering algorithm can not only solves the problems of the effectiveness of traditional clustering algorithms,but also makes it easy for large-scale data processing problems.By using the AP clustering algorithm,it is possible to obtain an effective road network area classification,and further analyze the status of each sub-area to detect congestion zone.