基于D2STGNN的双向高效多尺度交通流预测

doi:10.16088/j.issn.1001-6600.2025022702

摘要/Abstract

摘要： 由于交通流的复杂性和时空特征提取不足,D2STGNN难以捕捉交通网络的动态变化,限制了预测精度提升。本文提出一种高效的多头自注意力机制(EMHSA)和双向门控循环单元(BiGRU)结合的Bi-EMHGRU模型,该模型通过BiGRU捕捉前后时序依赖,并利用多头自注意力机制动态分配各时间步权重,聚焦关键时序特征。同时,引入多尺度时间特征提取模块,增强对短期波动和长期趋势的建模能力,提升复杂时空动态的建模效果。实验结果表明,Bi-EMHGRU在PeMS04和PeMS08数据集上表现优异,均方根误差值下降约0.55~1.55,平均绝对误差下降约0.89~1.40,平均绝对百分比误差下降约0.86~1.77个百分点,预测步长增加时仍能保持稳定的预测性能,泛化能力强。与现有基准模型相比,Bi-EMHGRU能够更有效地捕捉交通流的动态时空特征,显著提升预测精度和鲁棒性。

关键词: 交通流预测, 多头自注意力机制, Bi-EMHGRU, 动态时空特征, 多尺度时间特征

Abstract: Due to the complexity of traffic flow and the insufficient extraction of spatio-temporal features, it is difficult for D2STGNN to capture the dynamic changes of traffic networks, which limits the improvement of prediction accuracy. In this paper, a Bi-EMHGRU model combining an efficient multi-head self-attention mechanism (EMHSA) and a bidirectional gated recurrent unit (BiGRU) is proposed. This model captures the sequential dependencies of both forward and backward timings through BiGRU and dynamically allocates weights to each time step by using the multi-head self-attention mechanism to focus on key sequential features. Meanwhile, a multi-scale time feature extraction module is introduced, which enhances the modeling ability for short-term fluctuations and long-term trends and improves the modeling effect of complex spatio-temporal dynamics. The experimental results show that Bi-EMHGRU performs excellently on the PEMS04 and PEMS08 datasets. The root mean square error value has decreased by approximately 0.55~1.55, the mean absolute error has decreased by approximately 0.89~1.40, and the mean absolute percentage error has decreased by approximately 0.86~1.77 percentage points. It can still maintain stable prediction performance when the prediction step length increases and has strong generalization ability. Compared with the existing benchmark models, Bi-EMHGRU can capture the dynamic spatio-temporal features of traffic flow more effectively and significantly improves the prediction accuracy and robustness.

Key words: traffic flow prediction, multi-head self-attention, Bi-EMHGRU, dynamic spatio-temporal features, multi-scale temporal features

中图分类号: U49

黄艳国, 肖洁, 吴水清. 基于D2STGNN的双向高效多尺度交通流预测[J]. 广西师范大学学报（自然科学版）, 2026, 44(1): 10-22.

HUANG Yanguo, XIAO Jie, WU Shuiqing. Bidirectional Efficient Multi-scale Traffic Flow Prediction Based on D2STGNN[J]. Journal of Guangxi Normal University(Natural Science Edition), 2026, 44(1): 10-22.

参考文献

[1] 刘艺, 游克思, 孙培翔. 智慧城市道路体系架构与发展路径研究[J]. 城市交通, 2024, 22(5): 93-100. DOI: 10.13813/j.cn11-5141/u.2024.0029.
[2] 李玉莲. 基于数据驱动的城市道路交通状态预测研究[D]. 重庆: 重庆交通大学, 2021. DOI: 10.27671/d.cnki.gcjtc.2021.000132.
[3] BOX G E P, JENKINS G M, REINSEL G C, et al. Timeseries analysis: forecasting and control[M]. 5th ed. New York: Wiley, 2008.
[4] 张阳, 王梓良, 姚芳钰, 等.融合滞后极限学习机的IDBiLSTM短时交通流预测[J]. 重庆交通大学学报(自然科学版), 2024, 43(6): 39-46. DOI: 10.3969/j.issn.1674-0696.2024.06.06.
[5] RUMELHART D E, HINTON G E, WILLIAMS R J. Learning representations by back-propagating errors[J]. Nature, 1986, 323(6088): 533-536. DOI: 10.1038/323533a0.
[6] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. DOI: 10.1162/neco.1997.9.8.1735.
[7] CHO K, VAN MERRIENBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[C]//Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Doha, Qatar. Stroudsburg, PA: ACL, 2014: 1724-1734. DOI: 10.3115/v1/d14-1179.
[8] KIPF T N,WELLING M.Semi-supervised classification with graph convolutional networks[EB/OL].(2017-02-22)[2025-10-1]. https://arxiv.org/abs/1609.02907.DOI: 10.48550/arXiv.1609.02907.
[9] GORI M, MONFARDINI G, SCARSELLI F. A new model for learning in graph domains[C]//Proceedings. 2005 IEEE International Joint Conference on Neural Networks,Piscataway,NJ:IEEE,2005: 729-734. DOI: 10.1109/ijcnn.2005.1555942.
[10] 夏英, 刘敏. 基于时空注意力卷积神经网络的交通流量预测[J]. 西南交通大学学报, 2023, 58(2): 340-347. DOI: 10.3969/j.issn.0258-2724.20210526.
[11] 袁华, 陈泽濠. 基于时间卷积神经网络的短时交通流预测算法[J]. 华南理工大学学报(自然科学版), 2020, 48(11): 107-113, 122. DOI: 10.12141/j.issn.1000-565X.190276.
[12] 王博文, 王景升, 朱茵, 等. 基于ARMA-SVR的短时交通流量预测模型研究[J]. 公路交通科技, 2021, 38(11): 126-133. DOI: 10.3969/j.issn.1002-0268.2021.11.015.
[13] 赵文竹, 袁冠, 张艳梅, 等. 多视角融合的时空动态GCN城市交通流量预测[J]. 软件学报, 2024, 35(4): 1751-1773. DOI: 10.13328/j.cnki.jos.007018.
[14] 罗向龙, 郭凰, 廖聪, 等. 时空相关的短时交通流宽度学习预测模型[J]. 计算机工程与应用, 2022, 58(9): 181-186. DOI: 10.3778/j.issn.1002-8331.2011-0204.
[15] 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. California: IJCAI, 2017:3634-3640.DOI: 10.24963/ijcai.2018/505.
[16] LI Y G, YU R, SHAHABI C, et al.Diffusion convolutional recurrent neural network: data-driven traffic forecasting[EB/OL].(2019-09-15)[2025-03-08]. http://wds.ac.cn/portal/summer2019/paper/13.pdf.
[17] SHAO Z Z, ZHANG Z, WEI W, et al. Decoupled dynamic spatial-temporal graph neural network for traffic forecasting[J]. Proceedings of the VLDB Endowment, 2022, 15(11): 2733-2746. DOI: 10.14778/3551793.3551827.
[18] SCHUSTER M, PALIWAL K K. Bidirectional recurrent neural networks[J]. IEEE Transactions on Signal Processing, 1997, 45(11): 2673-2681. DOI: 10.1109/78.650093.
[19] 刘云翔, 梁智超. 一种高效的连续时序图注意力网络的交通预测模型[J]. 计算机工程, 2025, 51(4): 350-359. DOI: 10.19678/j.issn.1000-3428.0069059.
[20] 张伟, 朱凤华, 陈圆圆, 等. 基于注意力机制和分时图卷积的公交客流预测[J]. 模式识别与人工智能, 2021, 34(2): 167-175. DOI: 10.16451/j.cnki.issn1003-6059.202102008.
[21] 王小博, 张轩, 高张浩, 等. 基于边缘智能的高速公路交通流预测方法[J]. 计算机技术与发展, 2025, 35(4): 193-201. DOI: 10.20165/j.cnki.ISSN1673-629X.2025.0017.
[22] GUO S N, LIN Y F, FENG N, et al. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting[C]//The Thirty-Third AAAI Conference on Artificial Intelligence (AAAI-19). Washington, DC:AAAI, 2019: 922-929. DOI: 10.1609/aaai.v33i01.3301922.
[23] ZHENG C P, FAN X L, WANG C, et al. GMAN: a graph multi-attention network for traffic prediction[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(1): 1234-1241. DOI: 10.1609/aaai.v34i01.5477.
[24] 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. California:IJCAI,2019:1907-1913. DOI: 10.24963/ijcai.2019/264.
[25] SONG C, LIN Y F, GUO S N, et al. Spatial-temporal synchronous graph convolutional networks: a new framework for spatial-temporal network data forecasting[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2020, 34(1): 914-921. DOI: 10.1609/aaai.v34i01.5438.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed