Prediction of Following Vehicle Trajectory Considering Operation Characteristics of a Human Driver


In this paper, we propose a novel method to predict the trajectory of a following vehicle, based on the operation characteristics of a driver. If a lead vehicle suddenly decelerates to avoid colliding with interrupting vehicles, it may lead to an accident with the following vehicle. To prevent such accidents, it would be beneficial to predict the future positions of surrounding vehicles. Previous studies have proposed similar prediction methods; however, these studies have not considered the operation characteristics of drivers, even though the prediction performance largely depends on these characteristics. In this research, we assumed a driving scene wherein a human driver follows an autonomous vehicle. The proposed method is implemented in the autonomous vehicle. Consequently, the method is able to predict the trajectory of the following vehicle operated by a human driver. The contribution of this paper is to estimate the operation characteristics of the following driver and to apply the estimated result to obtain the trajectory prediction. It is demonstrated that the proposed method shows high prediction accuracy as compared to the previous methods.

Proceedings of the 2020 IEEE/SICE International Symposium on System Integration, pp.712-717