长安大学中央高校基本科研业务费专项资金资助(No. 300102321504, 300102321501, 300102321503)；西安市智慧高速公路信息融合与控制重点实验室(No. ZD13CG46)；陕西省重点研发计划(No. 2021GY-098).
Special Funds for Fundamental Research Expenses of Central Universities of Chang"an University(No. 300102321504, 300102321501, 300102321503); Key Laboratory of Information Fusion and Control of Xi"an Smart Expressway(No. ZD13CG46); Shaanxi Province Key Research and Development Plan (No. 2021GY-098).
Accurate and real-time short-term traffic flow prediction is critical for the construction of modern traffic management service system. In order to fully exploit and utilize the spatial-temporal characteristics of traffic flow interaction in different road sections, a two-level screening mechanism composed of autocorrelation function, cross-correlation function and KNN algorithm is constructed to evaluate the correlation between the target road section and optimize the combination of road sections, and realize deep mining of spatial information. One of the GCN-GRU combination forecasting model is proposed. The spatial characteristics of short-term traffic flow are captured by using the advantage of Graph Convolutional Network (GCN) in the global processing of section topology information, and the time characteristics are extracted by using the long-term memory ability of Gated Recurrent Unit (GRU) for time information. They are verified by the measured short-term traffic flow data of expressway. The results show that using the two-level screening mechanism to effectively screen the road sections and introducing a deep learning combination model, the prediction performance will be significantly improved, which is better than the commonly used models such as Stacked Autoencoders network (SAEs) and Temporal Convolutional Network (TCN).