Dynamic Origin-Destination Matrix Prediction with Line Graph Neural Networks and Kalman Filter
This work addresses real-time demand prediction for intelligent transportation systems, offering an incremental improvement by integrating existing techniques.
The paper tackled the problem of predicting dynamic Origin-Destination demand matrices from traffic flow data by proposing a framework that combines graph neural networks and a Kalman filter to handle spatial and temporal patterns, achieving the best performance in various scenarios on the New Jersey Turnpike network.
Modern intelligent transportation systems provide data that allow real-time dynamic demand prediction, which is essential for planning and operations. The main challenge of prediction of dynamic Origin-Destination (O-D) demand matrices is that demands cannot be directly measured by traffic sensors; instead, they have to be inferred from aggregate traffic flow data on traffic links. Specifically, spatial correlation, congestion and time dependent factors need to be considered in general transportation networks. In this paper we propose a novel O-D prediction framework combining heterogeneous prediction in graph neural networks and Kalman filter to recognize spatial and temporal patterns simultaneously. The underlying road network topology is converted into a corresponding line graph in the newly designed Fusion Line Graph Convolutional Networks (FL-GCNs), which provide a general framework of predicting spatial-temporal O-D flows from link information. Data from New Jersey Turnpike network are used to evaluate the proposed model. The results show that our proposed approach yields the best performance under various prediction scenarios. In addition, the advantage of combining deep neural networks and Kalman filter is demonstrated.