Researchers from Sheffield Hallam University have received the Best Paper Award at the 15th International Conference on the Internet of Things (IoT 2025) in Vienna for work exploring how drones can support search and rescue more efficiently.
The paper, titled “Real Time Edge Intelligence in UAV for Search and Rescue: Onboard Energy Efficient Video Summarisation with Reduced Data Transmission”, was authored by Chidike Vincent, Laurence Hirsch, Najam Ul Hasan, Caren Crizben, Ezlin Fernandes, Adriana Crainic, Jonathan Zasada James and James Baldwin.
The research addresses a key challenge in drone search and rescue. UAVs can capture large amounts of video footage but transmitting that footage can be slow in areas with poor connectivity, while analysing every frame onboard can quickly drain battery power.
The team developed an approach that first creates a much smaller summary of the recorded video and then analyses only the most relevant frames for signs of a person.
In simple terms, rather than asking the drone to process every second of footage, the system identifies the most useful moments first and focuses its analysis there.
The results showed substantial efficiency gains when running on low-power hardware such as a Raspberry Pi. Processing time was reduced from around 103 minutes to 13 minutes, while the amount of data was reduced from approximately 906 MB to 33 MB. Energy consumption was also reduced by around 77%, while maintaining reliable detection.
These improvements could have important practical benefits for search and rescue operations. Reducing processing and transmission demands means drones could stay in the air for longer, share useful information more quickly and support rescue teams in making faster decisions when time is critical.
The work demonstrates how artificial intelligence and edge computing can be designed for real world constraints, where energy, connectivity and processing power may all be limited.
Receiving the Best Paper Award at IoT 2025 recognises both the technical contribution of the research and its potential impact in real world search and rescue applications.
The paper is published through ACM:
Vincent C., Hirsch L., Ul Hasan N., et al. (2025). Real-Time Edge Intelligence in UAV for Search-and-Rescue: Onboard Energy-Efficient Video Summarisation with Reduced Data Transmission. IoT 2025.
DOI: 10.1145/3770501.3770508
