Modeling lane-changing behavior of connected autonomous vehicles based on molecular force fields

Connected autonomous vehicles (CAVs) play a pivotal role in intelligent transportation systems, where accurate modeling of lane-changing behavior is essential for ensuring traffic safety and efficiency. To overcome the limitations of existing approaches in capturing dynamic multi-vehicle interactions during lane-change maneuvers, this study introduces a novel lane-changing decision-making model grounded in molecular force field theory. The study begins by analyzing the dynamic characteristics of lane-changing behavior and derives the lane-change expected safety distance. Drawing inspiration from molecular interactions, vehicle behaviors are analogized as attractive and repulsive forces, and a decision-making framework is formulated using the Buckingham potential function. Simulation experiments conducted in the SUMO environment demonstrate that the proposed model reduces risky lane changes by 45 %, increases average vehicle speed by 3.38 %, and improves the average number of successful lane changes by 5.12 %. Compared to traditional models, the proposed approach significantly enhances the safety, efficiency, and comfort of lane-changing behavior. These findings offer a new theoretical foundation for cooperative lane-changing strategies among CAVs in intelligent connected traffic environments.