A Highly Efficient Traffic Prediction Model for Continuous Time-series Graph Attention Networks

doi:10.19678/j.issn.1000-3428.0069059

Abstract

Abstract:

Traffic prediction faces three primary challenges: traditional spatiotemporal modeling methods struggle to capture long-range dependencies effectively, fixed time-window mechanisms cannot adapt to dynamic temporal patterns, and conventional statistical-based models exhibit limitations in modeling complex topological relationships. To address these issues, this study proposes a Temporal-enhanced Efficient Graph Attention Network (T-EGT). First, an Efficient Multi-head Self-Attention (EMSA) mechanism is designed, employing parameter sharing and sparse computation strategies to reduce the computational complexity of attention heads from O(N²) to O(Nlog_aN). Second, a linear temporal extension module is developed, extending the temporal perception range from fixed K steps to an elastic window of K+Δ through learnable temporal convolution kernels, where Δ∈ serves as an adaptive adjustment parameter. Finally, a dynamic graph inference architecture is constructed by utilizing the neighborhood aggregation characteristics of Graph Neural Networks (GNNs) to automatically generate topological relationship matrixes containing 83 traffic elements at each time step. Experiments on five benchmark datasets, including PeMSD4 and METR-LA, demonstrate that T-EGAT significantly outperforms 16 baseline models (including Diffusion Convolutional Recurrent Neural Network (DCRNN), GraphWaveNet, and Attention Based Spatial-Temporal Graph Convolutional Network (ASTGCN)), achieving a 2.77%-5.97% reduction in Mean Absolute Error (MAE), 3.12%-6.44% improvement in Root Mean Square Error (RMSE), and 1.41%-2.3% decrease in single-step prediction time. Ablation studies quantify the module contributions: EMSA accounts for a 42% accuracy improvement, the temporal extension module reduces long-term prediction errors by 17%, and the dynamic graph generation mechanism enhances the topological modeling accuracy by 29%. The model demonstrates enhanced robustness in sudden traffic accident scenarios, achieving an anomaly detection F1 value of 0.873, indicating a 21.5% improvement over conventional methods. These findings provide a new technical framework for real-time traffic management systems with an elastic temporal modeling mechanism and efficient attention architecture, offering universal solutions for spatiotemporal prediction tasks.

Key words: intelligent traffic, traffic prediction model, Graph Neural Network (GNN), traffic flow, multi-head self-attention mechanism, artificial intelligence decision-making

摘要：

交通预测领域面临传统时空建模方法难以有效捕获长程依赖关系、固定时间窗口机制无法适应动态时序模式以及基于统计学的传统模型在复杂拓扑关系建模方面存在局限性3个主要挑战。针对上述问题, 提出基于连续时序的高效图注意力网络(T-EGAT)。首先设计高效多头自注意力机制(EMSA), 采用参数共享和稀疏计算策略, 将注意力头的计算复杂度从O(N)降低到O(Nlog_aN); 其次开发线性时序扩展模块, 通过可学习的时序卷积核将时间感知范围从固定K步扩展到K+Δ步的弹性窗口, 其中Δ∈为自适应调整参数; 最后构建动态图推理架构, 利用图神经网络(GNNs)的邻域聚合特性, 在每个时间步自动生成包含83个交通要素的拓扑关系矩阵。实验结果表明, 在PeMSD4、METR-LA等5个基准数据集上, T-EGAT相较16种基线模型(包括DCRNN、GraphWaveNet、ASTGCN等)取得显著提升, 平均绝对误差(MAE)降低了2.77%~5.97%, 均方根误差(RMSE)改善了3.12%~6.44%, 单步预测时间缩短了1.41%~2.3%。消融实验结果表明, EMSA带来42%的精度提升, 时序扩展模块减少了17%的长时预测误差, 动态图生成机制提高了29%的拓扑建模准确率。该模型在突发交通事故场景下表现出更强的鲁棒性, 异常事件检测F1值达到0.873, 较传统方法提升了21.5%。该方案为实时交通管理系统提供了新的技术框架, 其弹性时序建模机制和高效注意力架构为时空预测任务提供了普适性解决方案。

关键词: 智能交通, 交通预测模型, 图神经网络, 交通流, 多头自注意力机制, 人工智能决策

LIU Yunxiang, LIANG Zhichao. A Highly Efficient Traffic Prediction Model for Continuous Time-series Graph Attention Networks[J]. Computer Engineering, 2025, 51(4): 350-359.

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.

/ Recommend / Download Citations

URL: https://www.ecice06.com/EN/10.19678/j.issn.1000-3428.0069059

https://www.ecice06.com/EN/Y2025/V51/I4/350

Figures/Tables 9

Fig.1 The model structure of graph convolutional network

Fig.2 Efficient multi-headed self-attentive model

Fig.3 T-GMAT model structure

Fig.4 Comparison of predictive performance of various models in different ranges

References 31

1	陆文琦, 芮一康, 冉斌, 等. 智能网联环境下基于混合深度学习的交通流预测模型. 交通运输系统工程与信息, 2020, 20 (3): 47- 53.
	LU W Q , RUI Y K , RAN B , et al. Traffic flow prediction based on hybrid deep learning under connected and automated vehicle environment. Journal of Transportation Systems Engineering and Information Technology, 2020, 20 (3): 47- 53.
2	蒋欣彤, 冯慧芳. 基于时序图卷积网络的交通速度预测. 计算机科学与应用, 2022, 12 (6): 1544- 1552.
	JIANG X T , FENG H F . Traffic speed prediction based on temporal graph convolution network. Computer Science and Application, 2022, 12 (6): 1544- 1552.
3	王祥雪, 许伦辉. 基于深度学习的短时交通流预测研究. 交通运输系统工程与信息, 2018, 18 (1): 81- 88.
	WANG X X , XU L H . Short-term traffic flow prediction based on deep learning. Journal of Transportation Systems Engineering and Information Technology, 2018, 18 (1): 81- 88.
4	户佐安, 邓锦程, 韩金丽, 等. 图神经网络在交通预测中的应用综述. 交通运输工程学报, 2023, 23 (5): 39- 61.
	HU Z A , DENG J C , HAN J L , et al. Review on application of graph neural network in traffic prediction. Journal of Traffic and Transportation Engineering, 2023, 23 (5): 39- 61.
5	李翠锦, 瞿中. 复杂交通环境下多层交叉融合多目标检测. 电讯技术, 2023, 63 (9): 1291- 1299.
	LI C J , QU Z . Multi-layer intersection fusion and multi-target detection in complex traffic environment. Telecommunication Engineering, 2023, 63 (9): 1291- 1299.
6	XU X Y , LIU J , LI H Y , et al. Analysis of subway station capacity with the use of queueing theory. Transportation Research Part C: Emerging Technologies, 2014, 38, 28- 43. doi: 10.1016/j.trc.2013.10.010
7	CASCETTA E . Transportation systems engineering: theory and methods. Berlin, Germany: Springer, 2001.
8	LI C , XU P . Application on traffic flow prediction of machine learning in intelligent transportation. Neural Computing and Applications, 2021, 33 (2): 613- 624. doi: 10.1007/s00521-020-05002-6
9	LIU Q X , SUN S , LIU M , et al. Online spatio-temporal correlation-based federated learning for traffic flow forecasting. IEEE Transactions on Intelligent Transportation Systems, 2023, 25 (10): 13027- 13039.
10	WU Z H , PAN S R , CHEN F W , et al. A comprehensive survey on graph neural networks. IEEE Transactions on Neural Networks and Learning Systems, 2023, 32 (1): 4- 24.
11	WELLING M, KIPF T N. Semi-supervised classification with graph convolutional networks[C]//Proceedings of International Conference on Learning Representations. Washington D. C., USA: IEEE Press, 2016: 356-367.
12	MONTI F, BOSCAINI D, MASCI J, et al. Geometric deep learning on graphs and manifolds using mixture model CNNs[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, USA: IEEE Press, 2017: 5115-5124.
13	LI R Y , WANG S , ZHU F Y , et al. Adaptive graph convolutional neural networks. Artificial Intelligence, 2018, 32 (1): 1156- 1166.
14	LI Y G, YU R, SHAHABI C, et al. Diffusion convolutional recurrent neural network: data-driven traffic forecasting[EB/OL]. [2023-11-10]. https://arxiv.org/abs/1707.01926v3.
15	ZHANG J N, SHI X J, XIE J Y, et al. GaAN: gated attention networks for learning on large and spatiotemporal graphs[EB/OL]. [2023-11-10]. https://arxiv.org/abs/1803.07294v1.
16	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. Artificial Intelligence, 2020, 34 (1): 914- 921.
17	LI M Z, ZHU Z X. Spatial-temporal fusion graph neural networks for traffic flow forecasting[C]//Proceedings of the 35th AAAI Conference on Artificial Intelligence. [S. 1. ]: AAAI Press, 2021: 323-335.
18	GUO S N , LIN Y F , FENG N , et al. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting. Artificial Intelligence, 2019, 33 (1): 922- 929.
19	CHEN Y, SEGOVIA I, GEL Y R. Z-gcnets: time zigzags at graph convolutional networks for time series forecasting[C]//Proceedings of International Conference on Machine Learning. Washington D. C., USA: IEEE Press, 2021: 1684-1694.
20	CHOI J , CHOI H , HWANG J , et al. Graph neural controlled differential equations for traffic forecasting. Artificial Intelligence, 2022, 36 (6): 6367- 6374.
21	DUAN Y J , LV Y S , LIU Y L , et al. An efficient realization of deep learning for traffic data imputation. Transportation Research Part C: Emerging Technologies, 2016, 72, 168- 181.
22	DISSANAYAKE B, HEMACHANDRA O, LAHSHITHA N, et al. A comparison of arimax, VAR and LSTM on multivariate short-term traffic volume forecasting[C]//Proceedings of International Conference of Open Innovations Association. Washington D. C., USA: IEEE Press, 2021: 564-570.
23	LEI H , LIU L L X , WANG H . Survey on research and application of support vector machines in intelligent transportation system. Journal of Frontiers of Computer Science & Technology, 2020, 14 (6): 901.
24	BAI S J, KOLTER J Z, KOLTUN V, et al. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[EB/OL]. [2023-11-10]. https://arxiv.org/abs/1803.01271v2.
25	ZHANG K , LIU Z , ZHENG L . Short-term prediction of passenger demand in multi-zone level: temporal convolutional neural network with multi-task learning. IEEE Transactions on Intelligent Transportation systems, 2019, 21 (4): 1480- 1490.
26	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of Advances in Neural Information Processing Systems. Cambridge, USA: MIT Press, 2017: 146-157.
27	XU M X, DAI W R, LIU C M, et al. Spatial-temporal transformer networks for traffic flow forecasting[EB/OL]. [2023-11-10]. https://arxiv.org/abs/2001.02908v2.
28	ZHOU H Y, ZHANG S H, PENG J Q, et al. Informer: beyond efficient transformer for long sequence time-series forecasting[C]//Proceedings of AAAI Technical Track on Machine Learning. Washington D. C., USA: IEEE Press, 2021: 563-572.
29	ZHAO X , ASKARI H , CHEN J . Nanogenerators for smart cities in the era of 5G and Internet of things. Joule, 2021, 5 (6): 1391- 1431.
30	FENG A S, TASSIULAS L. Adaptive graph spatial-temporal transformer network for traffic forecasting[C]//Proceedings of the 31st ACM International Conference on Information and Knowledge Management. New York, USA: ACM Press, 2022: 3933-3937.
31	LUO X L, ZHU C J, ZHANG D T, et al. Dynamic graph convolutional network with attention fusion for traffic flow prediction[EB/OL]. [2023-11-10]. https://arxiv.org/abs/2302.12598.

[1]	LI Siyuan, ZHONG Xingyu, LI Kaiyin, XU Qingzhen. Strategy Teaching Research Based on Multilayer Graph Relationship and Reinforcement Learning [J]. Computer Engineering, 2025, 51(3): 122-130.
[2]	CHEN Depin, ZHAO Shen, JIAO Yiping, WANG Xiangxue, LÜ Hong, XU Jun. Graph Construction on Whole Image Slides Based on Attention and Learnable Threshold [J]. Computer Engineering, 2025, 51(3): 229-240.
[3]	JIA Shuo, LIN Shiyang, YANG Miaohui, SUN Teng. Short-Time Traffic Flow Prediction on Narrow Roads Based on Improved Whale-Optimized GRU [J]. Computer Engineering, 2025, 51(2): 111-125.
[4]	Zhengkang ZHANG, Dan YANG, Tiezheng NIE, Yue KOU. Self-Supervised Learning Based on Graph Structural Clustering for Disease Diagnosis Method [J]. Computer Engineering, 2024, 50(7): 360-371.
[5]	GAO Jiahao, HU Chuangye, DING Nan, LIU Zhandong. Validity Analysis of Traffic-Flow Data for Combined Driving Style in Intelligent Connected Vehicle [J]. Computer Engineering, 2024, 50(6): 367-376.
[6]	SHI Xin, CAO Fengteng, JI Yi, MA Junyan. Traffic Flow Prediction Method Based on Multi-Scale Spatio-Temporal Features and Soft Attention Mechanism [J]. Computer Engineering, 2024, 50(12): 346-357.
[7]	ZHU Yanbin, WANG Runmin, CHEN Hua, CAO Xiaofei, ZHU Zhenlin, DING Yajun. Traffic Text Detection Method Based on Multi-Granularity Feature Enhancement Network [J]. Computer Engineering, 2024, 50(11): 80-88.
[8]	Yueyue CHU, Fei YAN, Pu LI. Urban Traffic Signal Predictive Control Method with Iterative Learning Compensation for Disturbances [J]. Computer Engineering, 2023, 49(7): 305-312.
[9]	GE Yuran, FU Qiang. Short-Time Prediction Model for Urban Traffic Flow Based on Joint Spatio-Temporal Learning [J]. Computer Engineering, 2023, 49(1): 270-278.
[10]	SUN Xiufang, LI Jianbo, LÜ Zhiqiang, DONG Chuanhao. Lane Occupancy Prediction Model Based on Multi-Component Fusion and Dilated Graph Convolution [J]. Computer Engineering, 2021, 47(11): 62-68,76.
[11]	ZHAO Yanan, LIU Yuan, SONG She. Financial News Polarity Analysis Fusing with Multi-Head Self-Attention Mechanism [J]. Computer Engineering, 2020, 46(8): 85-92.
[12]	CAI Yanguang, LE Bing, CAI Hao, LI Xuyang. Short-Term Traffic Flow Forecast of Expressway Under Heavy Rain [J]. Computer Engineering, 2020, 46(6): 34-39.
[13]	YAN Yang, SUN Lijun, ZHU Lanting. Short-Term Traffic Flow Prediction Method Based on Spatiotemporal Relativity [J]. Computer Engineering, 2020, 46(1): 31-37.
[14]	YANG Miaomiao,FENG Jianhu,CHENG Xiaohan,FENG Juanjuan. Low dissipation central-upwind scheme for solving complex traffic flow model [J]. Computer Engineering, 2019, 45(6): 37-44.
[15]	Yuelong LI, Dehua TANG, Guiyuan JIANG, Zhitao XIAO, Lei GENG, Fang ZHANG, Jun WU. Short Term Traffic Flow Forecasting Based on Dimension Weighted Residual LSTM [J]. Computer Engineering, 2019, 45(6): 1-5.

Please choose a citation manager

Content to export