Dynamic lane-changing game model in mixed autonomous and human-driven vehicle environment

Abstract

Lane changing in mixed environments is a critical task that requires autonomous vehicles(AV) to respond to constantly changing traffic conditions in real time, while dealing with the uncertainties introduced by human-driven vehicles(HV). This paper proposes a dynamic lane-changing game model that captures the dynamic interaction among vehicles during lane-changing. To address the uncertainty in human drivers' intentions in the mixed environment, uncertain parameter is introduced that capture their diverse payoff preferences. When tackling the lack of information exchange between AV and HV, a strategy prediction algorithm based on the introduced parameter is proposed, enabling AV to predict the action of HV. The simulation is conducted using Next Generation Simulation (NGSIM) data to verify the effectiveness of the algorithm. The results indicate that the presence of algorithmic strategy assistance enables AV to consistently perform safe and efficient lane-changing maneuvers in the presence of parameter variations in human drivers.

Enhanced Safety in Multi-Lane Automated Driving Through Semantic Features

“…Researchers have explored lane-changing models, strategies, and decision-making algorithms to ensure safe and efficient lane transitions for autonomous vehicles (Liu et al, 2019). Understanding the dynamics of lane-changing behaviors in mixed traffic environments, where autonomous and humandriven vehicles coexist, is essential for developing effective lane-changing strategies (Cao et al, 2023).…”