| 0 | 0 | 21 |
| 下载次数 | 被引频次 | 阅读次数 |
针对节假日、春运暑运、恶劣天气及高温等事件扰动下高速公路交通流量预测中时空关联刻画不足、扰动适应能力较弱的问题，提出一种融合事件扰动感知与动态时空图学习的短时交通流量预测方法。该方法融合历史流量、出入口流量、分车型流量、OD 流向、周期时间及事件标签等多源数据，结合物理拓扑、OD 关联、历史相似和事件扰动构建动态图，并通过时空特征学习、常态预测与扰动残差修正实现自适应预测。基于约两年的高速公路实际业务数据开展实验，在 15 min 预测时距 下 ，所 提 ED‑DSTG 模 型 的 MAE、RMSE 和 MAPE 分 别 为 17.436、24.512 和 5.382%，R2 达 到0.984。相较 Graph WaveNet，MAE、RMSE 和 MAPE 分别降低 12.21%、12.28% 和 16.18%。在节假日、春运暑运、恶劣天气和高温场景下，MAE 相较 Graph WaveNet 降低 14.39%~15.48%。结果表明，所提方法能够提高事件扰动条件下高速公路短时交通流量预测精度及场景适应能力。
Abstract：To address the insufficient characterization of spatiotemporal correlations and limited adaptability to disturbances in expressway traffic flow prediction under holidays， Spring Festival and summer travel periods， adverse weather， and high-temperature conditions， an event disturbance-aware dynamic spatiotemporal graph learning method is proposed. The method integrates multi-source data， including historical traffic flow， entrance and exit flow， vehicle-type-specific flow， OD flow， periodic temporal features， and event labels. A dynamic graph is constructed by jointly considering physical topology， OD correlations， historical similarities， and event disturbances， followed by spatiotemporal feature learning， baseline prediction， and disturbance residual correction. Experiments on approximately two years of real-world expressway data show that， for 15-min-ahead prediction， ED-DSTG achieves an MAE of 17.436， an RMSE of 24.512， a MAPE of 5.382%， and an R2of 0.984. Compared with Graph WaveNet， the MAE，RMSE， and MAPE are reduced by 12.21%12.28%， and 16.18%， respectively. Under holiday， Spring Festival and summer travel， adverse weather， and high-temperature scenarios， the MAE reductions range from 14.39% to 15.48%. The results demonstrate that the proposed method improves short-term expressway traffic flow prediction accuracy and adaptability under event disturbances.
［ 1］ JIANG W W， LUO J Y. Graph neural network for traffic forecasting： a survey［J］. Expert Systems with Applications，2022，207：117921.
［ 2］ LI G P， KNOOP V L， VAN LINT H. Multistep traffic forecasting by dynamic graph convolution： interpretations of real-time spatial correlations［J］. Transportation Research Part C： Emerging Technologies，2021，128：
103185.
［ 3］ XU Y B， CAI X， WANG E， et al. Dynamic traffic correlations based spatio-temporal graph convolutional network for urban traffic prediction［J］. Information Sciences，2023，621：580-595.
［ 4］ QIU Z K， ZHU T Y， JIN Y H， et al. A graph attention fusion network for event-driven traffic speed prediction［J］. Information Sciences，2023，622：405-423.
［ 5］ QI X Y， MEI G， TU J Z， et al. A deep learning approach for long-term traffic flow prediction with multifactor fusion using spatiotemporal graph convolutional network［J］. IEEE Transactions on Intelligent Transportation Systems，2023，24（8）：8687-8700.
［ 6］ FANG S， PRINET V， CHANG J L， et al. MS-net：multi-source spatio-temporal network for traffic flow prediction［J］. IEEE Transactions on Intelligent Transportation Systems，2022，23（7）：7142-7155.
［ 7］ WENG W C， FAN J， WU H F， et al. A Decomposition Dynamic graph convolutional recurrent network for traffic forecasting［J］. Pattern Recognition，2023，142：109670.
［ 8］ AHMED S F， KULDEEP S A， RAFA S J， et al. Enhancement of traffic forecasting through graph neural network-based information fusion techniques［J］. Information Fusion，2024，110：102466.
［ 9］ LI F X， FENG J， YAN H， et al. Dynamic graph convolutional recurrent network for traffic prediction：benchmark and solution［J］. ACM Transactions on
Knowledge Discovery from Data，2023，17（1）：1-21.
［10］ XU Y， HAN L Z， ZHU T Y， et al. Generic Dynamic Graph Convolutional Network for traffic flow forecasting［J］. Information Fusion，2023，100：101946.［11］ LIU A Y， ZHANG Y Y. Spatial‑temporal dynamic graph convolutional network with interactive learning for traffic forecasting［J］. IEEE Transactions on Intelligent Transportation Systems，2024，25（7）：7645-7660.
［12］ GAN R， AN B C， LI L H， et al. A freeway traffic flow prediction model based on a generalized dynamic spatio-emporal graph convolutional network［J］. IEEE Transactions on Intelligent Transportation Systems，2024，25（10）：13682-13693.
［13］ LI Q， XU P， HE D Q， et al. Multi-source information fusion graph convolution network for traffic flow prediction［J］. Expert Systems with Applications， 2024，252：124288.
［14］ LI Y G， YU R， SHAHABI C， et al. Diffusion convolutional recurrent neural network： data-driven traffic forecasting［C］//Proceedings of the International Conference on Learning Representations. 2018.
［15］ VELIČKOVIĆ P， CUCURULL G， CASANOVA A，et al. Graph attention networks［C］//Proceedings of the International Conference on Learning Representations.2018.
［16］ BAI S J， KOLTER J Z， KOLTUN V. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling［EB/OL］.（2018-03-04）.https：//arXiv.org/abs/1803.01271.
［17］ VASWANI A， SHAZEER N， PARMAR N， et al. Attention is all you need［C］//Advances in Neural Information Processing Systems. 2017，30：5998-6008.
［18］ KONG J L， FAN X M， ZUO M， et al. ADCT-Net：adaptive traffic forecasting neural network via dualgraphic cross-fused transformer［J］. Information Fusion，2024，103：102122.
［19］ XU X， HU X Q， ZHAO Y， et al. Urban short-term traffic speed prediction with complicated information fusion on accidents［J］. Expert Systems with Applications，2023，224：119887.
［20］ YU B， YIN H T， ZHU Z X. Spatio-temporal graph convolutional networks： a deep learning framework for traffic forecasting［C］//Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence. International Joint Conferences on Artificial Intelligence Organization，2018：3634-3640.
［21］ GUO S N， LIN Y F， FENG N， et al. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting［J］. Proceedings of the AAAI Conference on Artificial Intelligence，2019，33（1）：922-929.
［22］ WU Z H， PAN S R， LONG G D， et al. Graph WaveNet for deep spatial-temporal graph modeling［C］//Proceedings of the Twenty-Eighth International
Joint Conference on Artificial Intelligence. International Joint Conferences on Artificial Intelligence Organization，2019：1907-1913.
基本信息:
DOI：10.3969/j.issn.2097-017X.2026.05.004
引用信息:
[1]李正言,江薇,戈晓峰,等.融合事件扰动感知与动态时空图学习的高速公路短时交通流量预测方法[J],2026,(05):28-46.DOI:10.3969/j.issn.2097-017X.2026.05.004.
基金信息:
国家自然科学基金面上项目(52278486)
2026-07-21
2026
2026-09-15