Traffic Status Classification for Highways Based on FCM Algorithm

In this paper, a real-time method of traffic status estimation for highways is proposed based on the fuzzy C-mean (FCM) clustering algorithm. By using the exponential smoothing method in data preprocessing, the short-term traffic disturbance is eliminated such that the stability of the results is improved. The method is verified by the data collected by ultrasonic detectors installed on state road 107, Fangshan District of Beijing. Traffic flow and time average speed are selected as classification indicators. The false rate cross-evaluation method is used to evaluate the results of traffic status classification. The evaluation result shows that the method established in this paper is feasible and effective for highway traffic status estimation.