Hierarchical Spatio-Temporal Graph Convolutional Networks and Transformer Network for Traffic Flow Forecasting

• Published in: IEEE transactions on intelligent transportation systems 2023-04, Vol.24 (4), p.1-13
• Main Authors: Huo, Guangyu, Zhang, Yong, Wang, Boyue, Gao, Junbin, Hu, Yongli, Yin, Baocai
• Format: Article
• Language: English
• Subjects: Artificial neural networks; Convolution; Forecasting; Graph convolutional networks; Network topology; Networks; Predictive models; Roads; Smoothing; Task analysis; Topology; traffic data forecasting; Traffic flow; transformer; Transformers
• ISSN: 1524-9050; 1558-0016
• DOI: 10.1109/TITS.2023.3234512

Graph convolutional networks (GCN) have been applied in the traffic flow forecasting tasks with the graph capability in describing the irregular topology structures of road networks. However, GCN based traffic flow forecasting methods often fail to simultaneously capture the short-term and long-term temporal relations carried by the traffic flow data, and also suffer the over-smoothing problem. To overcome the problems, we propose a hierarchical traffic flow forecasting network by merging newly designed the long-term temporal Transformer network (LTT) and the spatio-temporal graph convolutional networks (STGC). Specifically, LTT aims to learn the long-term temporal relations among the traffic flow data, while the STGC module aims to capture the short-term temporal relations and spatial relations among the traffic flow data, respectively, via cascading between the one-dimensional convolution and the graph convolution. In addition, an attention fusion mechanism is proposed to combine the long-term with the short-term temporal relations as the input of the graph convolution layer in STGC, in order to mitigate the over-smoothing problem of GCN. Experimental results on three public traffic flow datasets prove the effectiveness and robustness of the proposed method.