An Autonomous Drone for Avalanche Search and Rescue: Integrating ArduPilot and Tracking-Beacon Detection

摘要

Avalanches claim numerous lives annually, underscoring the critical need for rapid and reliable rescue solutions. This paper presents an innovative autonomous drone system explicitly designed to enhance avalanche search and rescue (SAR) operations in challenging mountainous terrains. Utilising a hexacopter equipped with the CubePilot flight controller and integrated with a Raspberry Pi onboard computer, our system uniquely combines ArduPilot’s precise navigation capabilities and GNU Radio’s sophisticated signal processing techniques to detect avalanche transceiver signals swiftly and accurately. Rigorous testing through simulation and extensive field trials conducted in real alpine conditions demonstrate the drone’s robust performance. In our field trials at a controlled avalanche debris site, the drone detected a buried standard avalanche transceiver and autonomously home in its location in only several minutes. This is a noteworthy improvement compared to conventional search times – manual rescuer sweeps can take 20–30 minutes or more to locate a victim, even when using transceivers. Furthermore, the cost-effective and open-source nature of our drone solution promotes wide accessibility and adaptability for various SAR scenarios. By bridging advanced aerial robotics with practical rescue operations, this study delivers a vital technological advancement, significantly boosting the effectiveness and safety of mountain rescue missions. Code is available on GitLab: https://gitlab.com/albatros-uav/albatros.