Research Of Transportation Subarea Partition Based On Multi-source Data Fusion

With the progress of urbanization,the scale of urban road networks has been expanding,and regional traffic congestion becomes a common phenomenon.In order to cope with traffic congestion,urban regional coordinated control has become an efficient method to manage and control traffic in large-scale transportation networks.Traffic control subarea partition is basis of coordinated control.The partition of traffic subarea refers to the scientific and effective partitioning of urban road networks so as to apply a reasonable strategy according to the characteristics of each subarea,and improve the traffic service.Scholars at home and abroad have done lots of researches on subarea partition and achieved abundant results,but there are still some problems.The partition to large-scale transportation network must be propitious to traffic management.The partition method should be efficient and reliable as well the partition results can response to the change of traffic state swiftly.How to computer traffic performance index with multi-detector data for fast and logical partition is a research hotspot in Intelligent Transportation Systems.A transportation subarea partition method based on multi-source data fusion is proposed in this paper.And the effectiveness of this method has been proved with a simulation study.The main work of this paper are as follows:1.A traffic data fusion method based on association rules is proposed to process multi-source data and get traffic performance index.In view of the FP_Growth algorithm,the confidence degree of traffic data is introduced into the rule confidence degree,and we propose the IFP(the Improved FP_Growth)algorithm to determine the correlation between traffic flow parameters.Then the data fusion method of IFP_DF algorithm is proposed,it generates association relationship of multi-source data with traffic history data and estimates the traffic performance index with real time traffic data.2.In addition,in order to meet the requirements of efficiency and effectiveness in large-scale road network partition,Normalized Cut(Ncut)algorithm is applied to divide traffic subarea in this paper.We analyze and summarizes the purpose and principle of traffic subarea partition,obtain the traffic state index with the data fusion method,and combines the distance factors between intersections to get the correlation matrix of the transportation networks.Moreover,combining with the characteristics of transportation networks,this paper presents Ncut-based traffic partition method.At the same time,with the goal of maximizing total correlation in subarea,the boundary intersections of subareas are adjusted so that the partition result tends to be better.3.We apply these methods to real data,analyze spatial relationship of road networks,computer the state index of intersections,and analyze the data fusion result.Moreover,the road network is divided with Ncut-based traffic partition method and the partition result is analyzed in detail.At last,the superiority of this method is illustrated when compared with other method.