Learning driving behavior for autonomous vehicles using deep learning based methods^*

摘要

There are amounts of uncertainties in the driving environments. Existing autonomous driving technologies focus on the improvements of the sensing abilities to guaranty the safety of the vehicles. Furthermore, decision making methods could be designed. Human drivers could always make right decisions in the case of unexpected situations in driving activities. In this paper, a driving behavior learning method based on deep learning is proposed. In this method, the driving behavior could be learned in different driving environments. Finally, the proposed method is evaluated in the aspects of effectiveness and efficiency in the simulation experiments. The experimental results demonstrate that the proposed methods outperform other machine learning methods.