Curriculum Reinforcement Learning for Quadrotor Racing with Random Obstacles

TL;DR

This paper introduces a vision-based curriculum reinforcement learning framework that enables quadrotors to race through obstacle-rich environments with higher success rates and faster lap times, advancing autonomous drone racing capabilities.

Contribution

The paper presents a novel curriculum reinforcement learning approach combining multi-stage curriculum, domain randomization, and multi-scene updating for robust obstacle avoidance in drone racing.

Findings

Achieves faster lap times in obstacle-rich environments.

Demonstrates higher success rates compared to existing methods.

Validates effectiveness through hardware-in-the-loop and real-world experiments.

Abstract

Autonomous drone racing has attracted increasing interest as a research topic for exploring the limits of agile flight. However, existing studies primarily focus on obstacle-free racetracks, while the perception and dynamic challenges introduced by obstacles remain underexplored, often resulting in low success rates and limited robustness in real-world flight. To this end, we propose a novel vision-based curriculum reinforcement learning framework for training a robust controller capable of addressing unseen obstacles in drone racing. We combine multi-stage cu rriculum learning, domain randomization, and a multi-scene updating strategy to address the conflicting challenges of obstacle avoidance and gate traversal. Our end-to-end control policy is implemented as a single network, allowing high-speed flight of quadrotors in environments with variable obstacles. Both hardware-in-the-loop and real-world experiments demonstrate that our method achieves faster lap times and higher success rates than existing approaches, effectively advancing drone racing in obstacle-rich environments. The video and code are available at: https://github.com/SJTU-ViSYS-team/CRL-Drone-Racing.