Game Theory-Based Lane Change and Merge Planning for Semi-Trailer Trucks

Lane change and merge maneuvers often involve highly dynamic interactions between the ego vehicle and surrounding traffic, which become even more challenging when the ego vehicle is a semi-trailer truck. Traditional methods frequently struggle with adapting to individual driving styles and managing complex interactions, especially when applied to semi-trailers. This manuscript introduces a game theory-based approach for planning lane changes and merges in semi-trailer trucks. The method employs level-k game theory to model driver behavior and designs a reward function in Monte Carlo simulations tailored to the kinematic characteristics of semi-trailers. By iteratively evaluating action-state pairs, the approach identifies the trajectory with the highest reward. Experimental results demonstrate that our method effectively navigates complex interactions with high accuracy and adaptability, offering novel insights for improving decision-making in autonomous semi-trailer trucks.