TY  - JOUR
PY  - 2022//
TI  - MetaDrive: composing diverse driving scenarios for generalizable reinforcement learning
JO  - IEEE transactions on pattern analysis and machine intelligence
A1  - Li, Quanyi
A1  - Peng, Zhenghao
A1  - Feng, Lan
A1  - Zhang, Qihang
A1  - Xue, Zhenghai
A1  - Zhou, Bolei
SP  - ePub
EP  - ePub
VL  - ePub
IS  - ePub
N2  - Driving safely requires multiple capabilities from human and intelligent agents, such as the generalizability to unseen environments, the safety awareness of the surrounding traffic, and the decision-making in complex multi-agent settings. Despite the great success of Reinforcement Learning (RL), most of the RL research works investigate each capability separately due to the lack of integrated environments. In this work, we develop a new driving simulation platform called MetaDrive to support the research of generalizable reinforcement learning algorithms for machine autonomy. MetaDrive is highly compositional, which can generate an infinite number of diverse driving scenarios from both the procedural generation and the real data import ing. Based on MetaDrive, we construct a variety of RL tasks and baselines in both single-agent and multi-agent settings, including benchmarking generalizability across unseen scenes, safe exploration, and learning multi-agent traffic. The generalization experiments conducted on both procedurally generated scenarios and real-world scenarios show that increasing the diversity and the size of the training set leads to the improvement of the RL agent's generalizability. We further evaluate various safe reinforcement learning and multi-agent reinforcement learning algorithms in MetaDrive environments and provide the benchmarks. Source code, documentation, and demo video are available at https://metadriverse.github.io/metadrive.<p />  <p>Language: en</p>
LA  - en
SN  - 0162-8828
UR  - http://dx.doi.org/10.1109/TPAMI.2022.3190471
ID  - ref1
ER  -