北京大学学报（自然科学版）2026，Vol.62Issue(2)：253-265,中插1-中插3,16.DOI:10.13209/j.0479-8023.2025.094
DrivingGym:面向自动驾驶的跨仿真强化学习代理构建方法
DrivingGym:Building Cross-Simulation Reinforcement Learning Agent for Autonomous Driving
摘要
Abstract
Reinforcement learning(RL)for autonomous driving faces challenges such as low sample efficiency and convergence difficulties when directly trained in complex scenarios.To address this issue,we propose a cross-simulation agent construction method based on unified data representation and implement the DrivingGym training environment.This method abstracts the input state into three layers:sensor data,vehicle states,and road network information.The control interface unification is achieved across different simulation environments through action adapters.Experiments on common simulation platforms such as CARLA and Metadrive demonstrate that the propo-sed method can support training with mainstream reinforcement learning frameworks like RLlib and Stable-Baseli-nes3,and enable cross-simulation application of autonomous driving policies from simple to complex scenarios.关键词
自动驾驶/强化学习(RL)/跨仿真环境/代理构建Key words
autonomous driving/reinforcement learning(RL)/cross-simulation environment/agent building引用本文复制引用
聂子力,李俊泽,陈敬宇,董乾,薛云志..DrivingGym:面向自动驾驶的跨仿真强化学习代理构建方法[J].北京大学学报（自然科学版）,2026,62(2):253-265,中插1-中插3,16.基金项目
中国科学院青年创新促进会项目资助 （）
