SafeSwarm: Decentralized Safe RL for the Swarm of Drones Landing in Dense Crowds

2025-01-13

AI Summary

Ensuring safe, high-precision autonomous landing of multi-rotor UAV swarms in dense, dynamic environments remains challenging due to heterogeneous dynamics, mobile landing platforms, and coupled static and inter-agent collision risks—especially without centralized coordination. Method: This paper proposes a decentralized Safe Barrier Net, integrating safety-critical reinforcement learning, control barrier functions, and distributed policy learning to enable individualized dynamical modeling, cooperative tracking of moving landing platforms, and simultaneous avoidance of static obstacles and inter-UAV collisions—entirely without a central controller. Contribution/Results: Evaluated on Vicon motion-capture systems with Crazyflie 2.1 nano-quadcopters, the approach achieves a mean landing accuracy of 2.25 cm and completes each landing task in 17 seconds, with zero collisions throughout. It represents the first closed-loop experimental validation of formal safety constraints embedded within distributed deep reinforcement learning on real-world micro-UAV swarms, significantly enhancing robustness and safety for swarm landing in complex, dynamic scenarios.

Technology Category

• Intelligent Robots: Multi-Robot Systems
• Natural Language Processing: Safety and Robustness
• Multiagent Systems: Adversarial Agents

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Decentralized Web and Fediverse systems
• Security and Privacy: Large-scale security measurements

Abstract

This paper introduces a safe swarm of drones capable of performing landings in crowded environments robustly by relying on Reinforcement Learning techniques combined with Safe Learning. The developed system allows us to teach the swarm of drones with different dynamics to land on moving landing pads in an environment while avoiding collisions with obstacles and between agents. The safe barrier net algorithm was developed and evaluated using a swarm of Crazyflie 2.1 micro quadrotors, which were tested indoors with the Vicon motion capture system to ensure precise localization and control. Experimental results show that our system achieves landing accuracy of 2.25 cm with a mean time of 17 s and collision-free landings, underscoring its effectiveness and robustness in real-world scenarios. This work offers a promising foundation for applications in environments where safety and precision are paramount.

Problem

• Drone Swarms
• Safe Landing
• Crowded Environment

Innovation

• SafeSwarm
• Reinforcement Learning
• Safe Learning