Automatic Identification And Visualization System Of Urban Traffic Congestion Based On The Big Floating Car Data

The problem of traffic congestion is more serious by the increase of vehicle population, which severely affects people’s travel experience and daily life, even gradually become a social problem that the transportation department need to solve. After years of planning and implementation, several cities of China has established the monitoring and scheduling information platforms via GPS data of floating car, which not only bring convenience for travel but also ease the traffic congestion. However, they didn’t deeply mining the existing data or analyze the distribution law of traffic congestion.Based on scientific research project in communications department of Shanxi province, the paper studies key algorithms and technology involved in the automatic identification for traffic congestion state with floating car data, and provide strong support for decision-making of traffic management by analyzing the distribution law of traffic congestion. Firstly, clean speed data by the distance-based weighted average method, which can ensure the high quality of instant speed data. Then increase efficiency of shortest path searching by constructing topology for digital map data based on breadth-first traversal to and raise speed of matching information updating by infer direction. Finally, identify traffic congestion state automatically using congestion identifying Method, statistical and analysis the time-varying laws about the proportion of traffic jam section and time varying regularity of main road congestion status.The paper also build and test visualization system of urban traffic congestion, which can supply visual result about the distribution of the traffic congestion for traffic managers. Firstly, put forward an outline about visualization of urban traffic congestion based on the WebGIS on the basis of study about the key technology includes map display, show of realtime congestion state, present of historical congestion state and spatial data processing. Secondly, design the database based on demand analysis and overall design, conduct detailed design based on module partition of system. Finally, implement visualization system of urban traffic congestion using the parallel technology of multi-nodes, then test and analyzing result.The results shows, the cleaning method of speed data can effectively remove towering change, building topology data can obviously raise the efficiency of map matching algorithm. The improved algorithm has higher computational efficiency than the existing map matching algorithm. The implemented System of the paper has basically met the demand of real-time processing, and can realize the scale out of the system by adding nodes.