Dual-branch graph neural network with spatio-temporal fusion for traffic flow forecasting
Proceedings of the Institution of Civil Engineers - Transport · 05.11.2026 · 1-20
DOI: 10.1680/jtran.26.00084 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the Institution of Civil Engineers - Transport, online erschienen am 05.11.2026. Verfasst von Jing Quan und 2 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Gefördert von: Natural Science Foundation of Chongqing, Scientific and Technological Research Programme of Chongqing Municipal Education Commission.
Abstract (Originalsprache)
Reliable traffic flow forecasting is fundamental to intelligent transportation systems. Existing spatio-temporal models often emphasise either localised variation or long-range dependency, rely on spatial relations that respond too slowly to traffic changes, and fuse multi-branch features in a relatively shallow way. To address these issues, this paper proposes a dual-branch spatio-temporal fusion graph neural network (DB-STFGNN). The first branch captures localised coupled dynamics through multi-scale temporal convolution and a hybrid diffusion dynamic graph convolution (HDGCN) module with iterative interaction updates. The second branch employs spatio-temporal multi-head attention to model broader dependencies across nodes and time steps. To strengthen collaboration between the two branches, the authors further introduce a bidirectional cross-attention fusion module (CAFM), which performs cross-branch interaction and feature refinement before prediction. A non-autoregressive prediction head then maps the fused representation to future traffic flow values. Experiments on real-world benchmarks, including PeMS03, PeMS04, PeMS07 and PeMS08, show that DB-STFGNN achieves competitive performance under mean absolute error, root mean square error and mean absolute percentage error compared with traditional spatio-temporal models and mainstream graph-based baselines.
Förderung
- Natural Science Foundation of Chongqing
- Scientific and Technological Research Programme of Chongqing Municipal Education Commission
Zitieren
APA
Quan, J., Zeng, Y., & Su, L. (2026). Dual-branch graph neural network with spatio-temporal fusion for traffic flow forecasting. Proceedings of the Institution of Civil Engineers - Transport, 1-20. https://doi.org/10.1680/jtran.26.00084
Vancouver
Quan J, Zeng Y, Su L. Dual-branch graph neural network with spatio-temporal fusion for traffic flow forecasting. Proceedings of the Institution of Civil Engineers - Transport. 2026:1-20. doi:10.1680/jtran.26.00084
BibTeX
@article{Quan2026,
author = {Quan, Jing and Zeng, Yufan and Su, Liyun},
title = {Dual-branch graph neural network with spatio-temporal fusion for traffic flow forecasting},
journal = {Proceedings of the Institution of Civil Engineers - Transport},
year = {2026},
pages = {1--20},
doi = {10.1680/jtran.26.00084}
}
EUVIMED