|
|
基于空间信息迟延感知的时空Transformer交通流预测
|
Abstract:
准确可靠的交通流预测对城市规划、居民出行、公共安全等具有重要现实意义，目前对于交通流预测的方法主要集中在RNN等深度学习模型，仍有以下问题：一是没有考虑长、短距离动态空间依赖，二是忽视了空间信息传播情况。基于以上情况本文提出一种基于空间信息迟延感知的时空Transformer模型。首先，该模型采用多头自注意力机制捕捉城市交通流的时空特性；其次，在编码层使用空间与时间自注意力机制关注长、短距离动态空间影响与时间动态特性，同时也模拟了空间信息传播迟延情况；最后，时空编码层通过跳跃连接处理，从而进一步提高模型的性能。实验在两个真实的交通流数据集上进行，对比主流的交通预测算法，结果表明所提出的DAST-Transformer模型具有更好的性能和预测精度。
Accurate and reliable traffic flow prediction has important practical significance for urban planning, residents’ travel, public safety, etc. At present, the methods for traffic flow prediction are mainly integrated in the RNN and other in DL models, but there are still the following problems: first, the dynamic spatial dependence of long and short distance is not considered, and second, the spatial information dissemination is ignored. Based on the above situation, this paper proposes a spatial- temporal Transformer model based on spatial information delay awareness. First of all, the model uses the multi-attention mechanism to capture the space-time characteristics of urban traffic flow; Secondly, in the encoder layer, the spatial-temporal self-attention mechanism is used to focus on the long and short distance dynamic spatial impact and temporal dynamic characteristics, and the spatial information transmission delay is also simulated; Finally, the spatial-temporal encoder layer can further improve the performance of the model through jump connection processing. The experiment is carried out on two real traffic flow data sets. Compared with the mainstream traffic flow prediction algorithms, the results show that the proposed DAST-Transformer model has better performance and prediction accuracy.
| [1] | Yin, C.T., Xiong, Z., Chen, H., et al. (2015) A Literature Survey on Smart Cities. Science China Information Sciences, 58, 1-18. https://doi.org/10.1007/s11432-015-5397-4 |
| [2] | Nagy, A.M. and Simon, V. (2018) Survey on Traffic Prediction in Smart Cities. Pervasive and Mobile Computing, 50, 148-163. https://doi.org/10.1016/j.pmcj.2018.07.004 |
| [3] | Zivot, E. and Wang, J. (2006) Modeling Financial Time Series with S-PLUS. Springer, New York. |
| [4] | 傅贵, 韩国强, 逯峰, 许子鑫. 基于支持向量机回归的短时交通流预测模型[J]. 华南理工大学学报(自然科学版), 2013, 41(9), 71-76. |
| [5] | Djenouri, Y., Belhadi, A., Lin, J.C.-W. and Cano, A. (2019) Adapted K-Nearest Neighbors for Detecting Anomalies on Spatio-Temporal Traffic Flow. IEEE Access, 7, 10015-10027 https://doi.org/10.1109/ACCESS.2019.2891933 |
| [6] | Ben-Gal, I. (2008) Bayesian Networks. In: Ruggeri, F., Kenett, R.S. and Faltin, F.W., Eds., Encyclopedia of Statistics in Quality and Reliability, Wiley, Hoboken. https://doi.org/10.1002/9780470061572.eqr089 |
| [7] | Vinayakumar, R., Soman, K.P. and Poornachandran, P. (2017) Applying Deep Learning Approaches for Network Traffic Prediction. 2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI), Udupi, 13-16 September 2017, 2353-2358. https://doi.org/10.1109/ICACCI.2017.8126198 |
| [8] | Smith, B.L. and Demetsky, M.J. (1994) Short-Term Traffic Flow Prediction Models—A Comparison of Neural Network and Nonparametric Regression Approaches. Proceedings of IEEE International Conference on Systems, Man and Cybernetics, San Antonio, 2-5 October 1994, 1706-1709. https://doi.org/10.1109/ICSMC.1994.400094 |
| [9] | Ranjan, N., Bhandari, S., Zhao, H.P., Kim, H. and Khan, P. (2020) City-Wide Traffic Congestion Prediction Based on CNN, LSTM and Transpose CNN. IEEE Access, 8, 81606-81620. https://doi.org/10.1109/ACCESS.2020.2991462 |
| [10] | Yu, B., Yin, H. and Zhu, Z. (2018) Spatio-Temporal Graph Convolutional Networks: A Deep Learning Framework for Traffic Forecasting. Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, Stockholm, 13-19 July 2018, 3634-3640. https://doi.org/10.24963/ijcai.2018/505 |
| [11] | Zhao, L., Song, Y., Zhang, C., et al. (2019) T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction. IEEE Transactions on Intelligent Transportation Systems, 21, 3848-3858. https://doi.org/10.1109/TITS.2019.2935152 |
| [12] | Hu, J., Lin, X. and Wang, C. (2022) DSTGCN: Dynamic Spatial-Temporal Graph Convolutional Network for Traffic Prediction. IEEE Sensors Journal, 22, 13116-13124. https://doi.org/10.1109/JSEN.2022.3176016 |
| [13] | Bahdanau, D., Cho, K. and Bengio, Y. (2014) Neural Machine Translation by Jointly Learning to Align and Translate.
https://arxiv.org/abs/1409.0473 |
| [14] | Vaswani, A., Shazeer, N., Parmar, N., et al. (2017) Attention Is All You Need. Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, 4-9 December 2017, 6000-6010. |
| [15] | Li, G.Y., Zhong, S.H., Chan, G., et al. (2021) A Lightweight and Accurate Spatial-Temporal Transformer for Traffic Forecasting. https://arxiv.org/abs/2201.00008 |
| [16] | Li, Y., Yu, R., Shahabi, C. and Liu, Y. (2017) Diffusion Convolutional Recurrent Neural Network: Data-Driven Traffic Forecasting. https://arxiv.org/abs/1707.01926 |
| [17] | Wu, Z., Pan, S., Long, G., et al. (2020) Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Virtual Event, 6-10 July 2020, 753-763. https://doi.org/10.1145/3394486.3403118 |
| [18] | Zheng, C., Fan, X., Wang, C. and Qi, J. (2020) GMAN: A Graph Multi-Attention Network for Traffic Prediction. Proceedings of the AAAI Conference on Artificial Intelligence, 34, 1234-1241. https://doi.org/10.1609/aaai.v34i01.5477 |
| [19] | Yan, H., Ma, X. and Pu, Z. (2021) Learning Dynamic and Hierarchical Traffic Spatiotemporal Features with Transformer. IEEE Transactions on Intelligent Transportation Systems, 23, 22386-22399.
https://doi.org/10.1109/TITS.2021.3102983 |
