Institute of Transportation, MOTC

       Travel time is one of the most important travel information. If we can accurately predict the travel time, travelers can plan their schedule appropriately to avoid traffic jam, and the traffic management agencies can implement effective traffic management actions. The purpose of this research is to integrate multiple data and propose a machine learning model to predict highway travel time.
       This study adopts the gradient boosting decision tree (GBDT) model as well as two proposed methods of fusing ETC and VD data to predict travel time. The empirical testing results show that the proposed GBDT model and data fusion method can improve the accuracy of travel time prediction by 82% to 97% compared with that of using only ETC data. It shows that the integration of multiple data sources can result in a better prediction accuracy