This research addresses the prediction needs in urban traffic development under resilient transportation scenarios. The model focuses on various sudden disturbance events in complex urban traffic environments, using deep learning models to predict traffic pattern evolution before and after emergencies.
The project combines Graph Convolutional Networks and attention mechanisms, using deep knowledge distillation to identify and select the most influential features and variables for traffic pattern prediction, building a reliable traffic prediction model for resilient transportation scenarios.
This research also includes a visualization platform to provide scientific decision support for traffic management departments, offering significant theoretical value and practical application prospects.