Identifying and Tracking Network-Wide Traffic Congestion Based on Mapping-to-Cells Vehicle Trajectory Data

To mitigate urban traffic congestion, it is a prerequisite to accurately identify bottlenecks where traffic congestion usually occurs and track the source of traffic flow at bottlenecks. With the development of mobile sensing technologies, this study employed a mapping-to-cells method to explore massive vehicle trajectory data in urban road network. First, the study area is divided into homogeneous square cells and vehicle trajectory data is mapped into each cell to estimate the information of average travel speed. Then, the congested cells, i.e., bottlenecks, can be easily identified by referring to the determined speed threshold. Next, the source of congestion, i.e., where the congested traffic flow at bottlenecks come from, is tracked in the network via mining both real and historical vehicle trajectory data. The experiment results based on real data in Baoding City help verify the effectiveness of the proposed method in both congestion identification and tracking analyses.