《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (5): 1472-1479.DOI: 10.11772/j.issn.1001-9081.2024050636

• 人工智能 • 上一篇    

用于交通流量预测的多图扩散注意力网络

王泉1,2(), 陆啟想1, 施珮2   

  1. 1.南京信息工程大学 电子与信息工程学院,南京 210044
    2.无锡学院 物联网工程学院,江苏 无锡 214105
  • 收稿日期:2024-05-17 修回日期:2024-08-01 接受日期:2024-08-01 发布日期:2024-08-20 出版日期:2025-05-10
  • 通讯作者: 王泉
  • 作者简介:王泉(1980—),男,湖北咸宁人,正高级工程师,博士,主要研究方向:车联网、工业物联网
    陆啟想(2000—),男,江苏宿迁人,硕士研究生,主要研究方向:车联网、交通流量预测
    施珮(1988—),女,安徽宣城人,讲师,博士,主要研究方向:无线传感网络、农业物联网。
  • 基金资助:
    国家自然科学基金资助项目(62072216);江苏省高校自然科学研究面上项目(21KJB520020);无锡市“太湖之光”科技攻关计划基础研究项目(K20231021)

Multi-graph diffusion attention network for traffic flow prediction

Quan WANG1,2(), Qixiang LU1, Pei SHI2   

  1. 1.School of Electronics and Information Engineering,Nanjing University of Information Science and Technology,Nanjing Jiangsu 210044,China
    2.School of IoT Engineering,Wuxi University,Wuxi Jiangsu 214105,China
  • Received:2024-05-17 Revised:2024-08-01 Accepted:2024-08-01 Online:2024-08-20 Published:2025-05-10
  • Contact: Quan WANG
  • About author:WANG Quan, born in 1980, Ph. D., professor of engineering. His research interests include internet of vehicles, industrial internet of things.
    LU Qixiang, born in 2000, M. S. candidate. His research interests include internet of vehicles, traffic flow prediction.
    SHI Pei, born in 1988, Ph. D., lecturer. Her research interests include wireless sensor network, agricultural internet of things.
  • Supported by:
    National Natural Science Foundation of China(62072216);Jiangsu Province Natural Science Project of Higher Educational Institutions(21KJB520020);Wuxi Municipal Science and Technology “Taihu Light” Basic Research Program(K20231021)

摘要:

当前基于时空特征提取的交通流量预测方法中存在挖掘全局空间相关性与长期的动态时间依赖关系能力不足的问题,其中空间相关性的挖掘很大程度上取决于图结构的质量,为此提出一种多图扩散注意力网络(MGDAN),主要包括多图扩散注意力模块(MGDAM)和时间注意力模块。首先,使用自适应时空嵌入生成器构建动态的时空信息;其次,采用最大互信息系数(MIC)矩阵与自适应矩阵挖掘细粒度的空间信息,并利用全局空间注意力机制挖掘动态的空间相关性;最后,使用时间注意力模块提取非线性的时间相关性,并通过3个模块的结合实现时空相关性的有效提取。在PEMS08数据集上的实验结果表明,MGDAN在1 h内的平均绝对误差(MAE)相较于时空自编码器(ST_AE)和时空身份信息(STID)模型分别降低了19.34%和5.74%,且整体预测性能均优于9个基线模型,能够精准地进行中长期交通流量预测,为城市交通疏导提供理论依据。

关键词: 交通流量预测, 时空模型, 自适应时空嵌入, 图卷积网络, 注意力网络

Abstract:

Current traffic flow prediction methods based on spatio-temporal feature extraction has problems of insufficient capture of global spatial correlation and dynamic long-term temporal dependency, where spatial correlation mining relies on the quality of graph structure heavily. Therefore, a Multi-Graph Diffusion Attention Network (MGDAN) was proposed, consisting of a Multi-Graph Diffusion Attention Module (MGDAM) and a temporal attention module. Firstly, adaptive spatio-temporal embedding generator was used to construct dynamic spatio-temporal information. Secondly, a Maximal Information Coefficient (MIC) matrix and an adaptive matrix were utilized to explore fine-grained spatial information, and a global spatial attention mechanism was employed to capture dynamic spatial correlation. Finally, the temporal attention module was used to extract nonlinear temporal correlation, and the integration of the three modules was carried out to realize effective extraction of spatio-temporal correlation. Experimental results demonstrate that, on PEMS08 dataset, the Mean Absolute Error (MAE) of MGDAN model within one hour has 19.34% and 5.74% reductions compared to those of Spatio-Temporal AutoEncoder (ST_AE) and Spatial-Temporal IDentity (STID) models, respectively. At the same time, MGDAN model outperforms 9 baseline models in overall prediction performance, and can conduct medium- and long-term traffic flow prediction accurately, providing theoretical basis for urban traffic dispersion.

Key words: traffic flow prediction, spatio-temporal model, adaptive spatio-temporal embedding, Graph Convolutional Network (GCN), attention network

中图分类号: