Spatio-Temporal attention-based decomposition and reconstruction network for urban traffic flow prediction

Abstract

Short-term traffic flow prediction plays a critical role in alleviating traffic congestion and improving road utilization efficiency. However, due to the intricate spatiotemporal variation patterns in traffic flow data, many existing methods struggle to effectively capture multi-scale temporal features and the influence of spatial locations on flow changes. To tackle this issue, we propose a decomposition and reconstruction network model based on spatiotemporal attention mechanisms, called ST-ADRNet. ST-ADRNet uses trigonometric encoding to generate time position vectors for capturing multi-scale temporal patterns and introduces trend offset vectors to improve stability and long-term prediction accuracy. In the decomposition module, time position and flow feature vectors are processed using a hybrid attention mechanism to combine flow data with time information, while geographical and similarity spatial attention balance the effects of distance and flow similarity. In the reconstruction module, these learned vectors are integrated and processed through temporal and spatial attention modules to further extract multi-scale temporal and spatial features. The experimental results demonstrate that the ST-ADRNet model significantly outperforms the baseline models in predicting traffic flow for the next 60 min. On the PeMSD4 dataset, ST-ADRNet achieves improvements of 16.84, 27.45, and 10.23% in MAE, RMSE, and MAPE, respectively. On the PeMSD8 dataset, ST-ADRNet achieves improvements of 12.08, 20.89, and 8.23% in the same metrics. On the METR-LA dataset, ST-ADRNet achieves improvements of 3.10, 6.78, and 9.02% in the same metrics.

Keywords

traffic flow prediction spatiotemporal attention mechanisms time position vectors trend offset vectors hybrid attention mechanism

Get full access to this article

View all access options for this article.

References

Jin

. Forecasting wholesale prices of yellow corn through the Gaussian process regression. Neur Comput Appl 2024; 36: 8693–8710.

Jin

. Price forecasting through neural networks for crude oil, heating oil, and natural gas. Meas Energy 2024; 1: 100001.

Gkioka

Dominguez

Mentzas

. An AutoML-based approach for automatic traffic incident detection in smart cities. Intell Decis Technol 2024; 18: 1101–1122.

Mourad

Shen

, et al. ASTIR: spatio-temporal data mining for crowd flow prediction. IEEE Access 2019; 7: 175159–175165.

Zhang

Zheng

. Deep spatio-temporal residual networks for citywide crowd flows prediction. Proc AAAI Conf Artif Intell 2017; 31.

Chen

Teo

, et al. Gated residual recurrent graph neural networks for traffic prediction. Proc AAAI Conf Artif Intell 2019; 33: 485–492.

Zhao

Song

Zhang

, et al. T-GCN: a temporal graph convolutional network for traffic prediction. IEEE Trans Intell Transp Syst 2020; 21: 3848–3858.

Bai

Zhu

Song

, et al. A3t-gcn: attention temporal graph convolutional network for traffic forecasting. ISPRS Int J Geo-Inform 2021; 10: 485.

Bai

Yao

, et al. Adaptive graph convolutional recurrent network for traffic forecasting. Adv Neur Inform Process Syst 2020; 33: 17804–17815.

10.

Pan

Long

, et al. Connecting the dots: multivariate time series forecasting with graph neural networks. Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining, 2020, pp. 753–763.

11.

Bai

Xia

Huang

, et al. Spatial-temporal graph neural network based on gated convolution and topological attention for traffic flow prediction. Appl Intell 2023; 53: 30843–30864.

12.

Wen

, et al. RPConvformer: a novel transformer-based deep neural networks for traffic flow prediction. Expert Syst Appl 2023; 218: 119587.

13.

Cai

Janowicz

Mai

, et al. Traffic transformer: capturing the continuity and periodicity of time series for traffic forecasting. Trans GIS 2020; 24: 736–755.

14.

Boh

GEP

Jenkins

Reinsel

, et al. Time series analysis: forecasting and control. John Wiley & Sons, 2015.

15.

Ahmed

Cook

. Analysis of freeway traffic time-series data by using Box-Jenkins techniques, 1979.

16.

Kamarianakis

Prastacos

. Spatial time series modeling: a review of the proposed methodologies. The Regional Economics Applications Laboratory, 2003.

17.

Mehdi

Pooranian

Vinueza Naranjo

. Cloud traffic prediction based on fuzzy ARIMA model with low dependence on historical data. Trans Emerg Telecommun Technol 2022; 33: e3731.

18.

Szeto

Gazis

. Application of Kalman filtering to the surveillance and control of traffic systems. Transp Sci 1972; 6: 419–439.

19.

Barimani

Rahimi Kian

Moshiri

. Real time adaptive non-linear estimator/predictor design for traffic systems with inadequate detectors. IET Intell Transp Syst 2014; 8: 308–321.

20.

Jin

. Predictions of steel price indices through machine learning for the regional northeast Chinese market. Neur Comput Appl 2024; 36: 20863–20882.

21.

Bernaś

Płaczek

Porwik

, et al. Segmentation of vehicle detector data for improved k-nearest neighbours-based traffic flow prediction. IET Intell Transp Syst 2015; 9: 264–274.

22.

Deshpande

Bajaj

. Performance analysis of support vector machine for traffic flow prediction. 2016 international conference on global trends in signal processing, information computing and communication (ICGTSPICC), 2016, pp. 126–129. IEEE.

23.

Hong

Kang

. Urban traffic flow forecasting using Gauss-SVR with cat mapping, cloud model and PSO hybrid algorithm. Neurocomputing 2013; 99: 230–240.

24.

Jin

. Wholesale price forecasts of green grams using the neural network. Asian J Econ Bank 2024.

25.

Shi

Chen

Wang

, et al. Convolutional LSTM network: a machine learning approach for precipitation nowcasting. Adv Neur Inform Process Syst 2015; 28.

26.

Liu

Zhen

, et al. Dynamic spatial-temporal representation learning for traffic flow prediction. IEEE Trans Intell Transp Syst 2021; 22: 7169–7183.

27.

Shahabi

, et al. Diffusion convolutional recurrent neural network: data-driven traffic forecasting. arXiv preprint arXiv:1707.01926, 2017.

28.

Pan

Long

, et al. Graph wavenet for deep spatial-temporal graph modeling. arXiv preprint arXiv:1906.00121, 2019.

29.

Yin

Zhu

. Spatio-temporal graph convolutional networks: a deep learning framework for traffic forecasting. arXiv preprint arXiv:1709.04875, 2017.

30.

Song

Lin

Guo

, et al. Spatial-temporal synchronous graph convolutional networks: a new framework for spatial-temporal network data forecasting. Proc AAAI Conf Artif Intell 2020; 34: 914–921.

31.

Wang

Huang

, et al. Spatio-Temporal self-supervised learning for traffic flow prediction. Proc AAAI Conf Artif Intell 2023; 37: 4356–4364.

32.

Shao

Zhang

Wei

, et al. Decoupled dynamic spatial-temporal graph neural network for traffic forecasting. arXiv preprint arXiv:2206.09112, 2022.

33.

Xing

Huang

, et al. Spatial linear transformer and temporal convolution network for traffic flow prediction. Sci Rep 2024; 14: 4040.

34.

Bahdanau

. Neural machine translation by jointly learning to align and translate. arXiv preprint arXiv:1409.0473, 2014.

35.

Kong

Xing

Wei

, et al. STGAT: spatial-temporal graph attention networks for traffic flow forecasting. IEEE Access 2020; 8: 134363–134372.

36.

Vaswani

. Attention is all you need. Advances in Neural Information Processing Systems, 2017.

37.

Jiang

Han

Zhao W

, et al. Pdformer: propagation delay-aware dynamic long-range transformer for traffic flow prediction. Proc AAAI Conf Artif Intell 2023; 37: 4365–4373.

38.

Liu

Dong

Jiang

, et al. Spatio-temporal adaptive embedding makes vanilla transformer sota for traffic forecasting. Proceedings of the 32nd ACM international conference on information and knowledge management, 2023, pp. 4125–4129.

39.

Wang

Shi

. Short-term traffic speed forecasting hybrid model based on chaos-wavelet analysis-support vector machine theory. Transp Res Part C Emerg Technol 2013; 27: 219–232.

40.

Cao

VOK

Chan

VWS

. A CNN-LSTM model for traffic speed prediction. 2020 IEEE 91st vehicular technology conference (VTC2020-Spring), 2020, pp. 1–5. IEEE.

41.

Zhu

Wang

Tao

, et al. AST-GCN: attribute-augmented spatiotemporal graph convolutional network for traffic forecasting. IEEE Access 2021; 9: 35973–35983.