Accurate forecasting of traffic flow in the future period is very important for planning traffic routes and alleviating traffic congestion. However, traffic flow forecasting still faces serious challenges. Most of the existing traffic flow forecasting methods are static graph convolutional networks based on prior knowledge, ignoring the special spatial–temporal dynamics of spatial–temporal data. Using only adaptive dynamic graphs completely discards the objective and real spatial connectivity information in static graphs. To this end, we propose a novel information enhancement and dynamic-static fusion attention network (IEDSFAN). Firstly, the Multi-Graph Fusion Gating mechanism (MGFG) designed in IEDSFAN effectively fuses dynamic and static graphs to dynamically capture the hidden spatial–temporal correlation. Secondly, we construct a novel Gated Multi-head Self-Attention (GMHSA), which maps the input through the MGFG module to capture the complex spatial–temporal interactions in the features. Finally, we generate adaptive parameters to solve the problem that shared parameters cannot learn multiple traffic patterns, and enhance the expression of sequence information through the peak flag module. We conducted extensive experiments on five real-world traffic datasets, and the experimental results show that the performance of IEDSFAN is significantly better than all baselines.
The main problem within most existing traffic prediction methods is that they capture sequence features using a single scale, such as the time continuity or the direct adjacency relationship. It makes the relevant traffic prediction models still insufficient in capturing the dynamic patterns and spatio-temporal correlations. To fully mine the spatio-temporal correlation of traffic sequence, a multiscale spatio-temporal aware graph recurrent neural network (MSSTA-GRN) is proposed in the paper. Firstly, to capture the internal temporal correlation of traffic sequences at different scales, we decompose the hidden states of GRU into different scales and then update the corresponding hidden states according to different frequencies, the speed changes at different frequencies can be better captured. Secondly, to capture the multiscale spatial correlation of the traffic network more comprehensively, a multiscale spatial feature capture module is constructed by cascading multiple GCNs, the multiscale GCN can better deal with the interaction between different regions. Finally, MSSTA-GRN is designed to fuse multiscale spatio-temporal features, and experimental results indicate that the proposed perception method can improve the accuracy and robustness of traffic prediction.