Study On Traffic Flow Reverse Of Guiyang Urban Road Intersections

In recent years, most cities in China have done lots of transport planning in different scales, therefore, the traffic surveys become an important aspect in many investigations. Because the intersection as the bottleneck constraining for a smooth traffic flow, the investigation for flow and flow characteristics is required a higher accuracy. In view of the complicated and inefficiency for artificial intersection traffic survey, in this study, we rely on the projects of "traffic capacity building in Guiyang" and make use of the data of Guiyang traffic survey to set up the anti-push model of traffic flow in intersection of Guiyang. Try to find a set of efficient, practical and accurate methods to get the data of road traffic flow and then realize the purpose of estimating intersection traffic flow only through the traffic survey of a particular section of the regional road.In this study, we take the traffic flow of some intersections within the downtown of Guiyang city and the observed distribution of traffic as input/output factor of the artificial neural network training to construct anti-push model of intersection, eventually, realizing calculate the other traffic flow in road network by a simple traffic survey (including the survey of road network structure and the traffic flow in some sections).Firstly this paper outlines some relevant content and described the methods of traffic investigation in Guiyang; meanwhile, the survey data were analyzed. Second, in view of the heavy traffic problems caused by unreasonable division, we introduced some basic principles and methods of cell division and the OD matrix. Then we completed the re-division of Guiyang traffic zone and the re-count of OD matrix. Thirdly, under a clear analysis of the principle of back stepping OD traffic matrix we complete the inverse prediction of Guiyang Transport Area OD matrix, Obtained a more realistic OD matrix after the re-zoning district, and then redistribution of the OD matrix to obtained the statistics of the road network in Guiyang. Finally, take traffic flow in the downtown of Guiyang city and observed distribution of traffic as input/ output factors to establish the back stepping traffic forecasting model of intersection that is based on BP neural network. Then testing the precision of the model and put it into application.