Research On The Urban Road Traffic State Identification Based On Support Vector Machine

Road traffic state identification is the basis of intelligent traffic managementsystem to release road traffic information and traffic guidance. Currently, theconstruction of the traffic information collection system and distribution system havebegun to take shape in China’s major cities. However, there are still some problemsto be resolved in data processing and real-time road traffic status discrimination.Road traffic state identification is a very important issue. It`s a reference for travelersto make a reasonable path selection by providing accurate real-time traffic statusinformation. Therefore, we need to make an accurate judgment and predictions forthe traffic state of the road network. This paper focuses on the traffic stateidentification problems.For the traffic state identification problems, traditional traffic statusclassification based on the national transportation department is not adapt to differentroads. In practice, it is difficult to meet the requirements of accuracy. Therefore, thispaper adopts fuzzy clustering to divide the traffic state for different sections of theroad. Then, we use MSVM（Multi-category Support Vector Machine）to classify thetraffic state in future time. MSVM based on the traditional support vector machinesis suitable for traffic state identification.This paper presents a model of real-time traffic status discrimination system.The System can determines accurately and release the traffic status in the future. Itcan also achieve traffic guidance smoothly and improve the service quality oftransport system.