基于混合图卷积的多通道时空交通流预测模型

doi:10.11959/j.issn.1000-0801.2023173

电信科学 ›› 2023, Vol. 39 ›› Issue (9): 97-110.doi: 10.11959/j.issn.1000-0801.2023173

• 研究与开发 • 上一篇

基于混合图卷积的多通道时空交通流预测模型

张雄涛¹^,², 郑景玉¹^,², 申情¹^,², 孙丹枫³, 蒋云良¹^,²^,⁴

¹ 湖州师范学院信息工程学院，浙江湖州 313000
² 浙江省现代农业资源智慧管理与应用研究重点实验室，浙江湖州 313000
³ 杭州电子科技大学计算机学院，浙江杭州 310018
⁴ 浙江师范大学计算机科学与技术学院，浙江金华 321004

修回日期:2023-09-01 出版日期:2023-08-01 发布日期:2023-08-01
作者简介:张雄涛（1984- ），男，博士，湖州师范学院副教授、硕士生导师，主要研究方向为机器学习、深度学习、模糊系统等
郑景玉（1997- ），女，湖州师范学院硕士生，主要研究方向为深度学习、智慧交通
申情（1982- ），女，湖州师范学院教授、硕士生导师，主要研究方向为智能决策、智能信息处理
孙丹枫（1988- ），男，博士，杭州电子科技大学副教授、硕士生导师，主要研究方向为模式识别与深度学习
蒋云良（1967- ），男，博士，浙江师范大学教授、博士生导师，主要研究方向为智能信息处理、GIS等
基金资助:
国家自然科学基金资助项目(U22A20102);浙江省“尖兵”“领雁”研发攻关计划项目(2023C01150)

Multi-channel spatial-temporal traffic flow prediction based on hybrid static-dynamic graph convolution

Xiongtao ZHANG¹^,², Jingyu ZHENG¹^,², Qing SHEN¹^,², Danfeng SUN³, Yunliang JIANG¹^,²^,⁴

¹ School of Information Engineering, Huzhou University, Huzhou 313000, China
² Zhejiang Province Key Laboratory of Smart Management and Application of Modern Agricultural Resources, Huzhou 313000, China
³ School of Computer, Hangzhou Dianzi University, Hangzhou 310018, China
⁴ School of Computer Science and Technology, Zhejiang Normal University, Jinhua 321004, China

Revised:2023-09-01 Online:2023-08-01 Published:2023-08-01
Supported by:
The National Natural Science Foundation of China(U22A20102);The “Pioneer” and “Leading Goose” Research ＆ Development Program of Zhejiang Province(2023C01150)

摘要/Abstract

摘要：

针对交通流预测模型没有考虑道路上下文相关性和空间依赖关系动态性的问题，提出一种基于混合图卷积的多通道时空交通流预测模型（MHGCN）。该模型采用三明治结构（即中间多通道空间模块，两边时间模块）提取时空特征，多通道空间模块又分为静态图卷积模块和动态图卷积模块。静态图卷积模块同时从拓扑空间结构、语义空间结构及其组合中提取特定和公共的特征；动态图卷积模块对不同的特征分配不同的权重，从未知的图结构中提取动态的空间特征。时间模块中采用多头注意力机制提取全局时间特征，采用时间门控机制提取局部时间特征。该模型从不同的空间结构中提取空间信息，从不同时间间隔提取时间信息，建立全局、全面的时空关系。实验结果表明，MHGCN 模型在 4 个公开数据集上的性能优于现有的交通流预测模型。

关键词: 智能交通, 动态图卷积, 多头注意力, 时空相关性, 多通道

Abstract:

Aiming at the problem that the traffic flow prediction model did not consider the correlation of road context and the dynamics of spatial dependency, a multi-channel spatial-temporal traffic flow prediction based on hybrid static-dynamic graph convolution (MHGCN) was proposed.A sandwich structure (i.e.multi-channel spatial module in the middle and temporal module on both sides) was used in the model to extract spatial-temporal features, and the multi-channel spatial module was divided into static graph convolution module and dynamic graph convolution module.The static graph convolution module simultaneously extracted specific and common features from topological spatial structures, semantic spatial structures, and their combinations.The dynamic graph convolution module assigned different weights to different features and extracts dynamic spatial features from unknown graph structures.In the temporal module, the multi-head attention mechanism was used to extract the global temporal features, and the temporal gating mechanism extracted the local temporal features.The model extracted spatial information from different spatial structures and temporal information from different time intervals to establish a global and comprehensive spatial-temporal relationship.The experimental results show that the MHGCN performs better than the existing traffic flow prediction models on four real world traffic flow datasets.

Key words: The National Natural Science Foundation of China, The “Pioneer” and “Leading Goose” Research ＆ Development Program of Zhejiang Province, intelligent transportation, dynamic graph convolution, multi-head attention, spatial-temporal correlation, multi-channel

中图分类号:

TP393

张雄涛, 郑景玉, 申情, 孙丹枫, 蒋云良. 基于混合图卷积的多通道时空交通流预测模型[J]. 电信科学, 2023, 39(9): 97-110.

Xiongtao ZHANG, Jingyu ZHENG, Qing SHEN, Danfeng SUN, Yunliang JIANG. Multi-channel spatial-temporal traffic flow prediction based on hybrid static-dynamic graph convolution[J]. Telecommunications Science, 2023, 39(9): 97-110.

图/表 9

图1

图2

图3

图4

表1

图5

表2

表3

图6

参考文献 32

[1]	JIANG Y L , FAN J B , LIU Y ,et al. Deep graph Gaussian processes for short-term traffic flow forecasting from spatiotemporal data[J]. IEEE Transactions on Intelligent Transportation Systems, 2022,23(11): 20177-20186.
[2]	LUO X L , NIU L Y , ZHANG S R . An algorithm for traffic flow prediction based on improved SARIMA and GA[J]. KSCE Journal of Civil Engineering, 2018,22(10): 4107-4115.
[3]	LUO X L , LI D Y , ZHANG S R . Traffic flow prediction during the holidays based on DFT and SVR[J]. Journal of Sensors, 2019: 1-10.
[4]	KNOL D , LEEUW F D , MEIRINK J F ,et al. Deep learning for solar irradiance nowcasting:a comparison of a recurrent neural network and two traditional methods[C]// International Confer-ence on Computational Science. Heidelberg:Springer-Verlag, 2021: 309-322.
[5]	ZHAO Z , CHEN W H , WU X M ,et al. LSTM network:a deep learning approach for short-term traffic forecast[J]. IET Intelligent Transport Systems, 2017,11(2): 68-75.
[6]	GUO S N , LIN Y F , LI S J ,et al. Deep spatial–temporal 3D convolutional neural networks for traffic data forecasting[J]. IEEE Transactions on Intelligent Transportation Systems, 2019,20(10): 3913-3926.
[7]	ZHANG J B , ZHENG Y , QI D K . Deep spatio-temporal residual networks for citywide crowd flows prediction[C]// Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence. New York:ACM Press, 2017: 1655-1661.
[8]	YAO H X , TANG X F , WEI H ,et al. Revisiting spatial-temporal similarity:a deep learning framework for traffic prediction[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019,33(1): 5668-5675.
[9]	廖熙雯, 冷甦鹏, 明昱君 ,等. 基于数字孪生的城市交通流智能预测与导引策略[J]. 电信科学, 2023,39(3): 70-79.
	LIAO X W , LENG S P , MING Y J ,et al. Digital twin based intelligent urban traffic forecasting and guidance strategy[J]. Telecommunications Science, 2023,39(3): 70-79.
[10]	WU Y K , TAN H C , QIN L Q ,et al. A hybrid deep learning based traffic flow prediction method and its understanding[J]. Transportation Research Part C:Emerging Technologies, 2018(90): 166-180.
[11]	BRUNA J , ZAREMBA W , SZLAM A ,et al. Spectral networks and locally connected networks on graphs[EB]. 2013:arXiv:1312.6203.
[12]	ZHAO T X , ZHANG X , WANG S H . GraphSMOTE:imbalanced node classification on graphs with graph neural networks[C]// Proceedings of the 14th ACM International Conference on Web Search and Data Mining. New York:ACM Press, 2021: 833-841.
[13]	LUO X , JU W , QU M ,et al. DualGraph:improving semi-supervised graph classification via dual contrastive learning[C]// Proceedings of 2022 IEEE 38th International Conference on Data Engineering (ICDE). Piscataway:IEEE Press, 2022: 699-712.
[14]	ZENG J Q , XIE P T . Contrastive self-supervised learning for graph classification[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021,35(12): 10824-10832.
[15]	ROSSI A , BARBOSA D , FIRMANI D ,et al. Knowledge graph embedding for link prediction:a comparative analysis[J]. ACM Transactions on Knowledge Discovery from Data, 2021,15(2): 1-49.
[16]	DING K Z , XU Z , TONG H H ,et al. Data augmentation for deep graph learning[J]. ACM SIGKDD Explorations Newsletter, 2022,24(2): 61-77.
[17]	ZHANG J L , CHEN F , GUO Y N ,et al. Multi-graph convolutional network for short-term passenger flow forecasting in urban rail transit[J]. IET Intelligent Transport Systems, 2020,14(10): 1210-1217.
[18]	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.
[19]	YU B , YIN H T , ZHU Z X . Spatio-temporal graph convolutional networks:a deep learning framework for traffic forecasting[C]// Proceedings of the 27th International Joint Conference on Artificial Intelligence. Menlo Park:AAAI Press, 2018: 3634-3640.
[20]	HUANG R Z , HUANG C Y , LIU Y B ,et al. LSGCN:long short-term traffic prediction with graph convolutional networks[C]// Proceedings of the Twenty-Ninth International Conference on International Joint Conferences on Artificial Intelligence. 2021: 2355-2361.
[21]	BERNDT D J , CLIFFORD J . Using dynamic time warping to find patterns in time series[C]// Proceedings of the 3rd International Conference on Knowledge Discovery and Data Mining. Menlo Park:AAAI Press, 1994: 359-370.
[22]	WANG L , KONIUSZ P . Uncertainty-DTW for time series and sequences[C]// Proceedings of European Conference on Computer Vision. Heidelberg:Springer-Verlag, 2022: 176-195.
[23]	SUN H , SUN M S , WENG J ,et al. Analysis of ID sequences similarity using dtw in intrusion detection for CAN bus[J]. IEEE Transactions on Vehicular Technology, 2022,71(10): 10426-10441.
[24]	杜辉, 郑长亮, 苗春雨 ,等. 一种动态特征匹配的部分重叠点云配准方法[J]. 电信科学, 2021,37(4): 97-107.
	DU H , ZHENG C L , MIAO C Y ,et al. A partial overlapping point cloud registration method based on dynamic feature matching[J]. Telecommunications Science, 2021,37(4): 97-107.
[25]	WANG X , ZHU M Q , BO D Y ,et al. AM-GCN:adaptive multi-channel graph convolutional networks[C]// Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery ＆ Data Mining. New York:ACM Press, 2020: 1243-1253.
[26]	TIAN C Y , CHAN W K , VICTOR C . Spatial‐temporal atten-tion wavenet:a deep learning framework for traffic prediction considering spatial‐temporal dependencies[J]. IET Intelligent Transport Systems, 2021,15(4): 549-561.
[27]	SUTSKEVER I , VINYALS O , LE Q V . Sequence to sequence learning with neural networks[C]// Proceedings of the 27th International Conference on Neural Information Processing Systems. Cambridge:MIT Press, 2014: 3104-3112.
[28]	LI Y G , YU R , SHAHABI C ,et al. Diffusion convolutional recurrent neural network:data-driven traffic forecasting[C]// Proceedings of International Conference on Learning Representations.[S.l.:s.n.], 2018: 1-15.
[29]	GUO S N , LIN Y F , FENG N ,et al. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting[C]// Proceedings of the AAAI conference on artificial intelligence. Menlo Park:AAAI Press, 2019: 922-929.
[30]	WU Z H , PAN S R , LONG G D ,et al. Graph WaveNet for deep spatial-temporal graph modeling[C]// Proceedings of the 28th International Joint Conference on Artificial Intelligence. Menlo Park:AAAI Press, 2019: 1907-1913.
[31]	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.
[32]	FANG Z , LONG Q Q , SONG G J ,et al. Spatial-temporal graph ODE networks for traffic flow forecasting[C]// Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery＆ Data Mining. New York:ACM Press, 2021: 364-373.

数据集	指标	FC-LSTM	STGCN	DCRNN	ASTGCN	Graph WaveNet	STSGCN	STGODE	MHGCN （本文模型）
PEMS03	RMSE	35.11±0.50	30.12±0.70	30.31±0.25	29.66±1.68	32.94±0.18	29.21±0.56	27.69±0.73	27.36±0.22
	MAE	21.33±0.24	17.49±0.46	18.18±0.15	17.69±1.43	19.85±0.03	17.48±0.15	16.51±0.06	16.17±0.10
	MAPE	23.33±4.23%	17.15±0.45%	18.91±0.82%	19.40±2.24%	19.93±0.49%	16.78±0.20%	18.38±1.10%	17.52±0.26%
PEMS04	RMSE	44.56±0.01	35.55±0.75	38.12±0.26	35.22±1.90	39.70±0.04	33.65±0.20	32.45±0.27	32.23±0.11
	MAE	28.70±0.01	22.70±0.64	24.70±0.22	22.93±1.29	25.45±0.03	21.19±0.10	20.87±0.11	20.55±0.07
	MAPE	19.20±0.01%	14.59±0.21%	17.12±0.37%	16.56±1.36%	17.29±0.24%	13.90±0.05%	14.98±0.42%	14.44±0.21%
PEMS07	RMSE	45.84±0.57	38.78±0.58	38.58±0.70	42.57±3.31	42.78±0.07	39.03±0.27	36.75±0.49	34.82±0.06
	MAE	29.98±0.42	25.38±0.49	25.30±0.52	28.05±2.34	26.85±0.05	24.26±0.14	23.95±0.60	22.05±0.05
	MAPE	13.20±0.53%	11.08±0.18%	11.66±0.33%	13.92±1.65%	12.12±0.41%	10.21±1.65%	10.98±0.75%	9.82±0.08%
PEMS08	RMSE	34.06±0.32	27.83±0.20	27.83±0.05	28.16±0.48	31.05±0.07	26.80±0.18	25.97±0.19	25.89±0.15
	MAE	22.20±0.18	18.02±0.14	17.86±0.03	18.61±0.40	19.13±0.08	17.13±0.09	16.79±0.13	16.62±0.07
	MAPE	14.20±0.59%	11.40±0.10%	11.45±0.03%	13.08±1.00%	12.68±0.57%	10.96±0.07%	11.84±0.52%	11.52±0.10%

数据集	指标		n₁=64			n₁=128			n₁=256
数据集	指标	n₂=64	n₂=128	n₂=256	n₂=64	n₂=128	n₂=256	n₂=64	n₂=128	n₂=256
PEMS03	RMSE	27.30	27.47	27.90	27.73	27.98	27.97	28.51	27.30	27.60
	MAE	16.17	16.43	16.52	16.27	16.37	16.53	16.73	16.14	16.24
	MAPE	17.33%	18.15%	18.26%	18.35%	17.39%	17.20%	21.21%	17.33%	17.77%
PEMS04	RMSE	32.32	32.61	32.65	32.34	32.51	33.32	32.40	32.23	32.67
	MAE	20.72	20.77	20.69	20.71	20.67	21.04	20.79	20.55	20.65
	MAPE	14.50%	14.55%	14.58%	14.59%	14.57%	14.42%	14.58%	14.34%	14.36%
PEMS07	RMSE	35.11	35.08	34.94	35.00	34.73	34.77	35.00	34.71	34.94
	MAE	22.61	22.46	22.44	22.43	22.15	22.18	22.45	22.13	22.38
	MAPE	10.59%	10.08%	10.11%	10.10%	9.92%	10.05%	10.11%	9.84%	10.13%
PEMS08	RMSE	26.25	25.97	26.48	26.24	26.14	26.16	26.13	25.89	26.25
	MAE	16.93	16.83	17.14	16.83	16.84	16.81	16.71	16.67	16.86
	MAPE	11.64%	11.52%	12.00%	11.55%	11.70%	11.75%	11.90%	11.51%	11.46%

数据集	指标	MHGCN-two	MHGCN-S	MHGCN-T	MHGCN-D	MHGCN
PEMS03	RMSE	28.99	27.71	27.68	38.25	27.30
	MAE	17.12	16.80	16.33	22.23	16.14
	MAPE	19.10%	18.98%	16.76%	31.77%	17.33%
PEMS04	RMSE	33.68	32.81	32.98	38.54	32.23
	MAE	20.85	21.08	20.97	25.22	20.55
	MAPE	14.46%	14.69%	14.23%	18.33%	14.34%
PEMS07	RMSE	36.51	35.58	35.57	56.86	34.71
	MAE	22.70	22.88	22.90	33.90	22.13
	MAPE	10.12%	10.39%	10.45%	15.75%	9.84%
PEMS08	RMSE	27.65	26.74	26.70	32.29	25.89
	MAE	17.64	17.12	17.16	21.15	16.67
	MAPE	11.43%	12.50%	11.37%	14.40%	11.51%

基于混合图卷积的多通道时空交通流预测模型

Multi-channel spatial-temporal traffic flow prediction based on hybrid static-dynamic graph convolution

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 32

相关文章 11

Metrics

推荐阅读 0

[1]	杨帅, 王瑞琴, 马辉. 基于多通道的边学习图卷积网络[J]. 电信科学, 2022, 38(9): 95-104.
[2]	陈山枝. 蜂窝车联网（C-V2X）及其赋能智能网联汽车发展的辩思与建议[J]. 电信科学, 2022, 38(7): 1-17.
[3]	陈燕雷, 孟俊仙, 刘玮, 王勋. 5G室内新型覆盖方案对比与分析[J]. 电信科学, 2021, 37(4): 151-158.
[4]	潘毅,李晖晖,曾磊,蔡伟文. 5G室内场景多通道联合收发技术性能与关键问题[J]. 电信科学, 2020, 36(7): 168-174.
[5]	周颖,杨丽花,杨龙祥,倪梦. 基于卡尔曼预测与压缩感知的WSN中高能效数据收集方法[J]. 电信科学, 2019, 35(1): 74-80.
[6]	陈正文. 基于公众移动通信网的V2X运营需求[J]. 电信科学, 2016, 32(8): 28-33.
[7]	方箭,朱颖,郑娜,李芃芃. 从WRC-19议题看5G与未来移动通信发展趋势[J]. 电信科学, 2016, 32(6): 65-72.
[8]	张婷婷,张武雄,裴冬,赵铖,俞涵. 基于自回归模型与神经网络模型的车流量预测对比[J]. 电信科学, 2016, 32(2): 55-59.
[9]	陈康,黄晓宇,王爱宝,陶彩霞,关迎晖,李磊. 基于位置信息的用户行为轨迹分析与应用综述[J]. 电信科学, 2013, 29(4): 118-124.
[10]	肖恒辉,李炯城,丁胜培,李桂愉. 面向公路智能交通系统的无线物联网技术研究[J]. 电信科学, 2013, 29(1): 89-96.
[11]	范庆彬,孙健凇,牟红光. DSRC技术及其通信机制的研究[J]. 电信科学, 2010, 26(8): 99-101.