The urban road traffic state identification method based on FCM clustering

Aiming at the fuzzy and uncertainty of traffic condition, on the analysis of traffic flow characteristics and traffic state division, this paper presented a new real-time traffic condition identification method based on the Fuzzy c-means clustering. Firstly, fuzzy c-means clustering technique was used to classify the sampled historical data, and the clustering center of different traffic condition was gotten. Then the real-time traffic data were used to identify which kinds of the states that the current traffic condition belonged to. Flow, speed and occupancy were as feature attribute of sample data, and traffic condition was divided into four states. One road in Ganzhou as an example, the traffic condition of this road was tested and analyzed with the method. The result was same with the results of actual measurement traffic data and the questionnaire survey through drivers. It shows that the division of traffic state was effective, and this method can accurately identify road traffic condition.