Research On The Traffic Information Fusion With The Multi-Source Detectors

Recently, with rapid economic development, the surge in car ownership, traffic congestion and traffic accidents occurred frequently. Traffic Problem has become one of the major problems in the world's governments including China. People urgently need to know the actual operating status of the road network traffic so that it can guide people commute travel, and it can provide the data basis about various traffic management measures for vehicle supervision department. The development of traffic detection technique provides the possibility for people to obtain information.Based on Multi-sources traffic detector, obtaining a large number of traffic flow basic data, they include the microwave detector, loop detector, license plate recognition detector and GPS taxi data. Especially, the GPS taxi data, daily 28 million data uploaded to the database, provides a wealth of historical traffic data in Beijing, and this paper analyzes the road network in Beijing and the distribution of traffic flow detector. Based on the road network in Beijing and the distribution of traffic flow detector, this paper has the two-step data fusion:the first step is to use multiple sources detectors and GPS in Beijing to basically cover the large lows of backbone network, and make use of basic data collection to make up the other road traffic status using the GPS data. The second step is to fuse road velocity among unit road and main road, thus making the value of the fusion is closer to the true value of the velocity and improve the accuracy of the basic data. This paper uses the BP Neural Network Model for the fusion of road velocity, makes the model training from the following part which divided into:units road and the main road, the expressways and the major trunk roads, the peak hour and the usual hour, completes eight kinds of network models and obtains expected results through validity of judgments. The paper also conducts an extensive survey of the velocity of GPS accuracy and finds that the value of the GPS velocity has the different levels according to road level. When the taxi is on the express instead of on the major trunk road, the value of the GPS velocity has higher accuracy.In relative with discontinued traffic flow, the value of velocity from the GPS is more accurate in the continued traffic flow.