Traffic Pattern Recognition of Intersection Based on Rough Fuzzy Neural Network
Mengmeng Zhang, Fengrui Sun, Xintong Han · 2014
Identifying traffic patterns of intersections accurately can provide technical support to optimize signal control schemes and realize intelligent control. The traffic patterns of intersections were defined from a qualitative point, and then the identification parameters were reduced based on the rough set theory to search the minimum set of parameters used to characterize the feature of traffic patterns. Through combining fuzzy inference system and artificial network organically, a traffic pattern recognition model of intersections was constructed based on the fuzzy neural network, which has been used as a fuzzy recognizer in this paper. Finally, the paper simulated an intersection in Jinan by the simulation platform of VISSIM as the research object to analyze the recognition accuracy of the model comparatively. The results of experiments show that the model can recognize traffic patterns precisely, such as free traffic flow, steady traffic flow (tolerable delay), close to unstable traffic