Nowadays, as a crucial component of intelligent transportation systems, traffic flow prediction has received extensive concern. However, most of the existing studies extracted spatial-temporal features with modules that do not differentiate with time and space, and failed to consider spatial-temporal heterogeneities. Furthermore, although previous works have achieved synchronous modeling of spatial-temporal dependencies, the consideration of temporal causality is still lacking in their graph structures. To address these shortcomings, a spatial-temporal heterogeneous and synchronous graph convolution network (STHSGCN) is proposed for traffic flow prediction. To be specific, separate dilated causal spatial-temporal synchronous graph convolutional networks (DCSTSGCNs) for various node clusters are designed to reflect spatial heterogeneity, different dilated causal spatial-temporal synchronous graph convolutional modules (DCSTSGCMs) for diverse time steps are deployed to take account of temporal heterogeneity. In addition, causal spatial-temporal synchronous graph (CSTSG) is proposed to capture temporal causality in spatial-temporal synchronous learning. We further conducted extensive experiments on four real-world datasets, and the results verified the consistent superiority of our proposed approach compared with various existing baselines.

The recent proposed Spatial-Temporal Residual Network (ST-ResNet) model is an effective tool to extract both spatial and temporal characteristics and has been successfully applied to urban traffic status prediction. However, the ST-ResNet model only extracts the local spatial characteristics and ignores the very important global spatial characteristics. In this paper, a novel Global-Local Spatial-Temporal Residual Correlation Network (GL-STRCN) model is proposed for urban traffic status prediction to further improve the prediction accuracy of the existing ST-ResNet model. The GL-STRCN model firstly applies Pearson’s correlation coefficient method to extract high correlation series. Then, considering both global and local spatial properties, two components consisting of 2D convolution and residual operation are used to capture spatial features. After that, based on Long Short-Term Memory (LSTM) or Gated Recurrent Unit (GRU), a novel long-term temporal feature extraction component is proposed to capture temporal features. Finally, the spatial and temporal features are aggregated together in a weighted way for final prediction. Experiments have also been performed using two datasets from TaxiCD and PEMS-BAY. The results indicated that the proposed model produces a better prediction performance compared with the results based on other baseline solutions, e.g., CNN, ST-ResNet, GL-TCN, and DGLSTNet.