Traffic Flow Detection Method Studying Based On Information Fusion

In order to alleviate the urban road congestion, directing traffic for traffic management agencies to divert vehicles to bring great convenience. In order to make response measures, we want to detect traffic flow parameters to get traffic information. The information obtained by the detector is very complicated, if these data are processed together can cause a lot of waste of resources, so we should use this new technology-the information fusion. The paper is comes from project of ’Based on highway traffic sensing network information examination·Excavation· Gathering, issue in auxiliary decision system demonstration’, based on information fusion and efficient solution to the diversion control, induced traffic, the best driving routes, and reduce congestion and other problems.Firstly, excluding abnormal data is a crucial step, because the information fusion results largely depends on the sources of information-the reliability of data, if not addressed all of the data to a unified fusion, it will certainly affect the accuracy of the results and an enormous waste of human and financial resources. Then the weighted average method based on minimum variance and minimum mean speed of detection, multiple detector data integration to get a more accurate valueThe main contribution of this paper are:(1) According to the current multi-detector fusion algorithm based on the weighted average of the Cauchy inequality algorithm and multi-attribute decision making algorithm, simulation results show that these two algorithms to meet the actual demand, and the algorithm itself has low complexity, very suitable for integration of the occasions of less precision.(2) Good or bad according to the detector quality fusion algorithm known good detector and an unknown good detector in a known good detector, based on the comparison of the data, we take the error of4%can be achieved the best results. And complexity of the algorithm is small, the results of integration to fully meet the actual demand.The simulation diagram can be seen using the trichotomy removed the data and then use the multi-attribute decision making for information fusion are the most ideal results.