A Spatiotemporal Apriori Approach to Capture Dynamic Associations of Regional Traffic Congestion

Abstract

Due to the interactions among adjacent roads in urban road networks, traffic congestion gradually propagates to neigboring roads, resulting in regional congestion. To develop advanced regional traffic control strategies, it is necessary to clearly understand the characteristics of regional congestion evolution. To this end, this paper proposes a data-driven approach to mine the spatiotemporal associations of regional traffic congestion. By introducing both time and space attributes, the intra-transaction spatiotemporal Apriori (IntraT-ST-Apriori) algorithm is developed to address the static features of regional traffic congestion; while the inter-transaction spatiotemporal Apriori (InterT-ST-Apriori) algorithm is developed to capture the dynamic characteristics of regional traffic congestion. Case studies are carried out for the urban road network in Tianjin, China, based on empirical data. The results indicate that the Intra-ST-Apriori algorithm can excavate the underlying associations of regional traffic congestion. Furthermore, the congestion propagation trajectories can be clearly revealed based on the InterT-ST-Apriori algorithm. It is expected that the proposed approach can support the regional traffic management and control, significantly relieving traffic congestion.