Title: On the impact of prior experiences in car following models: model development, computational efficiency, comparative analyses and extensive applications
Speaker: Prof. Qu Xiaobo (Chalmers University of Technology)
Venue: Room 604, Transportation Building, Wushan Campus
Time: June 29, Saturday, 9:30
A major shortcoming of conventional car following models is that these models only consider current spacing and velocities of the target vehicle and its immediate leading vehicle, without taking into account prior driving actions, even for those from the same driver. In other words, the prior experiences have no influence at all in predicting vehicular movements for the next time step. In this study, we propose a machine-learning-based methodology that is able to take advantage of the high-resolution historical traffic data in the current data-rich era, to predict vehicular movements in an accurate manner with a high computational efficiency for high-resolution, real-time traffic flow prediction, modelling, and vehicle control. The new model has a simple model structure based on a Fixed Radius Near Neighbours (FRNN) search algorithm. We further conduct a comprehensive performance comparison among the proposed model, another similar machine-learning-based model, and a conventional model. The results indicate that the FRNN search algorithm based car following model is superior to the other models in terms of prediction accuracy and is more computational efficient compared to its machine-learning-based counterpart. Some extensive applications of the proposed model are discussed at the end of this paper.
Announced by the School of Civil Engineering and Transportation