MetaDrive: Composing Diverse Driving Scenarios for Generalizable Reinforcement Learning

By Quanyi Li, Zhenghao Peng, Zhenghai Xue, Qihang Zhang, Bolei Zhou. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021

433 citations recorded by the source corpus.

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 importing. 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.

Subfields: Artificial Intelligence, Automotive Engineering, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/9dae8d598f9dd9d8c96f771699f5e37a27e71df3

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Peng Wu
2. Bernhard Jaeger
3. Li Chen
4. Kashyap Chitta
5. Hongyang Li
6. Andreas Geiger
7. Simon Lacoste-Julien
8. Sebastian Nowozin

In Automotive Engineering

1. Kashyap Chitta
2. Andreas Geiger
3. Hongyang Li
4. Li Chen
5. Peng Wu
6. Bernhard Jaeger
7. Pravendra Singh
8. Pranav Singh Chib

In Control and Systems Engineering

1. Peng Wu
2. Li Chen
3. Kashyap Chitta
4. Hongyang Li
5. Andreas Geiger

Citations in the corpus

• End-to-End Autonomous Driving: Challenges and Frontiers (2023, background): “To address this challenge, simulation is a viable alternative [14], [88], [89], [90], [91], [92].”
• NAVSIM: Data-Driven Non-Reactive Autonomous Vehicle Simulation and Benchmarking (2024, neutral): “A broad body of research conducts evaluations in simulators, such as CARLA [17] or Metadrive [29] with sensor simulation, or nuPlan [26] and Waymax [19] for data-driven simulation.”