Lane change decision-making with active interactions in dense highway traffic: A Bayesian game approach

Abstract

It is foreseeable that automated and human-driven vehicles will coexist in a mixed traffic environment in the predictable future. One of the crucial tasks of automated vehicles is to make human-like driving decisions when interacting with human drivers to ensure safety and efficiency. To this end, this paper proposes a lane change decision-making method based on the Bayesian game approach, focusing on dealing with the strong interactions in dense highway traffic. Firstly, the driver aggressiveness is estimated using Gaussian Mixture Model (GMM) based on the naturalistic driving data. Then the driver aggressiveness is integrated into the Bayesian game model as a critical factor to affect the decisions of the autonomous vehicle. Finally, the tentative behavior is generated to actively gather surrounding vehicles' responses when the recognized driver aggressiveness does not help the ego vehicle to make safe and efficient decisions in the uncertain driving environment. The results show that the proposed model can generate more human-like decisions after being validated and compared in the multiple interactive lane change scenarios.