Urban traffic condition recognition and accessibility prediction based on big data

Traffic congestion, frequent traffic accidents, and air pollution problems in the cities have brought opportunities for the development of intelligent transportation technologies and the data mining of massive traffic spatiotemporal data. Few existing studies have talked about the feature layer data fusion that reflects the main characteristics of the big data of traffic flow time series, and the prediction methods of temporal and spatial accessibility from alternative starting points to destinations in urban traffic spatiotemporal network are pending further research. For these reasons, this paper focused on traffic condition recognition and accessibility prediction based on multi-source big data of urban traffic flow. At first, this paper proposed a method for extracting the features of urban traffic spatiotemporal data, and realized preprocessing and fusion of the obtained data of the traffic flow and average vehicle speed of the road network. Then, a combinatorial optimization algorithm of simulated annealing (SA) and particle swarm optimization (PSO), and the Fuzzy C-Means (FCM) algorithm were employed to perform fuzzy classification on urban traffic conditions, and the urban traffic spatiotemporal network was used to describe the optimization objective function of urban traffic accessibility. Finally, this paper used real cases to verify the effectiveness of the proposed algorithm and the constructed model.