Tag
The paper proposes REARL, a closed-loop framework using real traffic data and large language models to enhance autonomous driving simulation by continuously monitoring and adjusting vehicle behavior for greater realism.