Traffic Congestion Source Back-Tracing Framework for Urban Road Networks Using Traffic Volume Data

Abstract

In a road network, traffic congestion always exhibits strong spatiotemporal dependence on its upstream links, even those distant from the congested area. Implementing optimal management strategies for these upstream source links can efficiently prevent downstream congestion. However, owing to the limitations of aggregated and historical data, real-time tracing the exact upstream source links of future congestion remains challenging. To address this issue, this study designed a novel framework to trace the source of future traffic congestion in urban road networks using real-time detected link volume data. The framework comprises a prediction module and a back-tracing module. The prediction module predicts the traffic states of all links and identifies the congested link in the future. The back-tracing module determines the upstream source links and their contributions in proportion to the potentially congested link. Both modules were developed based on explicit models, thus ensuring the interpretability and scalability in road networks with traffic signals. The proposed framework was validated on two different size networks in China, using the simulation of urban mobility (SUMO) platform. The case study results demonstrate that the proposed framework precisely identifies the upstream source links of the future congested link. Sensitivity analyses revealed that the proposed framework can accurately back-trace congestion under different demand levels. These findings highlight the significance of the proposed traffic congestion source back-tracing framework in real-world applications for mitigating congestion in advance.