Traffic Flow Data Fusion Of Floating Cars And Fixed Detectors

Traffic flow data detection is the basic problem of ITS, and obtaining the accurate traffic parameters like flow rate, vehicle speed by data detection is the important basis of carrying on the efficacious traffic management i.e. traffic induced, traffic control, etc. Traditionally, employing the single detector in the method of detection has been used. However, because the single detector may be influenced by the condition of itself, the data accuracy and detection range may be restricted. More accurate traffic information can be obtained by the fusion of multi-source traffic flow data, compared to that obtained by employing the single detector, to achieve more efficacious control and management of traffic.This paper studies the fusion of traffic flow data in connection with the characteristic of two detection technique i.e. floating car and fixed detectors. Providing the general frame of data fusion and matching the two kinds of data by compatibly selecting the fusion algorithm improve the accuracy and efficiency of the traffic flow data. The paper executes the experiment based on MATLAB to verify the effectiveness of the selected algorithm. The main work is summarized as follows:1,Designing the framework of data fusion, pretreatment, correlate before data fusion, providing data basis of the fusion of traffic flow data.2,Selecting the exponential smoothing and BP neural network two fusion algorithms on the basis of data fusion framework, execute the fusion respectively to aim at the two conditions of less fully floating car data and fully floating car data.3,Executing the fusion simulation in connection with average speed, employing exponential smoothing and BP neural network on the basis of simulation platform of MATLAB, and analyzing the experiment result.