High-Speed Vision-Based Flight in Clutter with Safety-Shielded Reinforcement Learning

2026-02-09

AI Summary

This work addresses the challenges faced by quadrotor drones in high-speed autonomous obstacle avoidance, where traditional modular pipelines suffer from high latency and pure reinforcement learning approaches lack safety guarantees. The authors propose an end-to-end reinforcement learning framework integrated with a model-driven safety mechanism: during training, physics-informed priors shape the reward function, while at deployment, a real-time safety filter projects the policy’s output onto a provably safe action set. This approach uniquely unifies high-speed flight with rigorous safety constraints within an end-to-end learning paradigm. Evaluated in both dense indoor obstacle courses and real-world forest environments, the system achieves stable flight at 7.5 m/s, outperforming conventional planners and existing end-to-end methods in both performance and robustness, while demonstrating exceptional generalization and safety.

Technology Category

• Natural Language Processing: Safety and Robustness
• Search and Optimization: Learning to Search
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

Quadrotor unmanned aerial vehicles (UAVs) are increasingly deployed in complex missions that demand reliable autonomous navigation and robust obstacle avoidance. However, traditional modular pipelines often incur cumulative latency, whereas purely reinforcement learning (RL) approaches typically provide limited formal safety guarantees. To bridge this gap, we propose an end-to-end RL framework augmented with model-based safety mechanisms. We incorporate physical priors in both training and deployment. During training, we design a physics-informed reward structure that provides global navigational guidance. During deployment, we integrate a real-time safety filter that projects the policy outputs onto a provably safe set to enforce strict collision-avoidance constraints. This hybrid architecture reconciles high-speed flight with robust safety assurances. Benchmark evaluations demonstrate that our method outperforms both traditional planners and recent end-to-end obstacle avoidance approaches based on differentiable physics. Extensive experiments demonstrate strong generalization, enabling reliable high-speed navigation in dense clutter and challenging outdoor forest environments at velocities up to 7.5m/s.

Problem

• high-speed flight
• obstacle avoidance
• safety guarantees
• quadrotor UAV
• cluttered environments

Innovation

• Safety-Shielded Reinforcement Learning
• Physics-Informed Reward
• Real-Time Safety Filter
• End-to-End UAV Navigation
• High-Speed Obstacle Avoidance