Authors: Jing Li*, University of Denver, Xuantong Wang, University of Denver, Tong Zhang, Wuhan University
Topics: Geographic Information Science and Systems
Keywords: Traffic flow prediction; Deep learning; Spatiotemporal modeling; Extreme weather
Session Type: Paper
Presentation File: No File Uploaded
Predicting traffic flow is critical in efficient maritime transportation management, coordination and planning. Scientists have proposed many prediction methods, most of which are designed for specific locations or for short-term prediction. For the purpose of management, methods that enable long-term prediction on large areas are highly desirable. Therefore, we propose to develop a spatiotemporal approach that can describe and predict traffic flows within a region. We design the model based on a multi-hexagonal Convolutional Network Model (mh-CNN) model that takes both flow dynamics and environmental conditions into account. This model is highly flexible in that it predicts zonal traffic flow within variable time windows. We have applied the method to measure and predict the daily and hourly traffic flows in the South Atlantic States rRegion by taking the impacts of extreme weather events into consideration. Results show that our method outperforms other methods in daily prediction during normal days and hourly prediction during the hurricane events. Based on the results, we also make recommendations regarding the future usages and customization of the model.