我所殷辰堃老师课题组完成的论文《Traffic Flow Prediction of Highway Service Area Based on Temporal Convolution Network and Graph Attention Network》于2023-12在《International Conference on Robotics, Intelligent Control and Artificial Intelligence (RICAI)》正式发表。论文摘要如下：
Abstract: Accurately predicting traffic volume in highway service area can improve management efficiency. On the basis of complex spatiotemporal correlations from road network topology and traffic data, a TeN-GAT-T model is proposed for predicting traffic in the service area, capturing complex spatiotemporal correlations in the traffic sequence. Firstly, the extracted data is preprocessed according to the rules, and the signal matrix of the diagram is constructed according to the road network structure. This model utilizes Transformer with attention mechanism as its foundation and incorporates time convolution and graph attention network to extract various features. During the experiment, the model's efficiency is evaluated based on RMSE, MAE, MAPE, and other relevant metrics. Finally, the proposed model is compared with several commonly used models, using the real data of UK motorway service area as an example.
引用信息：ZeRen Zhao, Shangtai Jin, Chenkun Yin, Yanqiu Wang. Traffic Flow Prediction of Highway Service Area Based on Temporal Convolution Network and Graph Attention Network. International Conference on Robotics, Intelligent Control and Artificial Intelligence (RICAI), 2023-12. DOI: 10.1109/ricai60863.2023.10488990