The Prediction Of Peak Commuting Traffic Flow Based On Multisource Data

With the accelerated process of urbanization in China,traffic congestion has increasingly become common problem in urban development,especially peak commuting.Accurately prediction of the commuter peak traffic flow is the key to ease traffic congestion.On the other hand,with the development of intelligent transportation and multidisciplinary integration,different types of massive multi-source data are accumulated,such as IC card,taxi GPS positioning data,population household registration data,insured data,social security Data,etc.These data reflect the commuter discipline by taking different modes in a certain extent,and provide a data guarantee for the prediction of commuter peak traffic flow.However,how to effectively match the mass of multi-source data and on this basis to predict commute peak traffic flow still lack a complete set of method system.This paper proposes a new prediction method of commuter peak traffic flow based on multi-source data.According to the vehicle's service radius and travel distance,the relevant classification algorithm,the paper summarizes the best scheme of commuter mode,and provides scientific decision for the optimal design of macro traffic planning.The main contents of this paper include:1.Firstly,the research background and significance of this article were introduced,then the summary analysis about traffic flow prediction progress of domestic and overseas was proposed.Research content of this paper was put forward based on above;2.The characteristics of traffic flow were introduced:dynamic characteristics,temporal similarity characteristics and spatial correlation characteristics based on different sensor acquisition data classification and analyzing the real traffic data of some typical roads in Nanan distract of Chongqing city3.Multi-source data preprocessing:Through the analysis of the basic population data,road network data,bus lines,taxi GPS trajectory data and so on,this paper proposes multi-source data preprocessing method for next preparation of forecasting traffic flow;4.Prediction method and its steps of commuter peak traffic flow were established:non-motorized travel statistical model was established based on using classical Dijkstra algorithm to compute the shortest path;private car traffic prediction model was established based on the rail transit priority algorithm;taxi traffic prediction model was introduced by using the map matching algorithm;then commuter peak traffic flow was predicted based on multi-source data.This paper used Wuhan city as an experimental area.to verify the accuracy of this method.The prediction of commuter peak traffic flow based on multi-source data breakthrough the limitation of relying on historical traffic data,which provides a new method for traffic flow forecasting and new ideas for micro-traffic planning and simulation.