Research on Clustering Simulation and Feature Recognition of Urban Traffic Flow
Urban traffic flow feature recognition is the key to traffic management and optimization.Based on this,a method of extracting dynamic traffic flow features and performance evaluation using different clustering algorithms is proposed.Firstly,the signal timing and vehicle distribution scene at the multi-intersection of the road network are de-fined,and the traffic flow and queuing time are output based on the VISSIM simulation environment.Then,the cluste-ring algorithm was used to extract the features of the traffic flow scene,and study the traffic flow clustering feature comparison corresponding to different signal timings,through the analysis of traffic flow clustering features,the optimi-zation target of road vehicle expected speed distribution was extracted to improve road traffic efficiency.Finally,the clustering performance of traffic flow scenes at different intersections was evaluated according to the contour coefficient method,the differences in traffic flow feature recognition by K-means,Agnes and DBSCAN algorithms were compared,and their scene adaptability was analyzed.The research results of traffic flow feature recognition provide a theoretical basis for evaluating the traffic efficiency of regional traffic and analyzing its influencing factors,thereby im-proving the urban traffic service level.
NETL
NSTL
万方数据
