A team from the School of Electrical and Electronic Engineering, here at the University of Sheffield, has been working on the development of vision-based, real-time algorithms for wildfire detection and localisation with swarms of uncrewed aerial vehicles.
Team AURA, led by Professor Lyudmila Mihaylova, participated in the semi-finals of the XPRIZE Wildfire competition, which was held between September 21 - 24 in Lancashire and demonstrated the technology developed here at Sheffield.
Professor Lyudmila Mihaylova said, “We are excited to be working in an area that can make a difference for firefighters and save human lives. We are part of a consortium formed of the University of Sheffield, the University of Bristol, and the Lancashire Fire and Rescue Service (LFRS).
The University of Sheffield has developed a computer vision system that is responsible for autonomously detecting wildfires in real time, and its overall integration with the UAV swarm. Together with the digital twin for UAV control developed by the University of Bristol, our AURA team delivers a competitive and efficient swarm and AI-driven system.
The developed distributed AI-enabled system of UAVs will significantly increase firefighters’ observation horizon and facilitate early detection of wildfires, all while preventing the incineration of large swathes of forest cover, reducing the amount of carbon and pollutants released into the atmosphere, and allowing remote inspection.
The team is grateful to the Innovate UK Future Flights II project “Protecting Environments with a Swarm of UAVs”, to the EPSRC, Grantham and NVIDIA for the financial support, in making this work possible.