Dynamic Space Partition Algorithm with an Archimedean Spiral for Wildfire Detection Using a Swarm of UAVs

Due to climate change in recent years, wildfires have become one of the most harmful hazards to the environment and society. In firefighting operations, the early stages are crucial to controlling wildfires successfully. In this paper, we propose an improvement to an existing dynamic space partition...

Full description

Saved in:
Bibliographic Details
Published inConference proceedings (International Conference on Unmanned Aircraft Systems. Online) pp. 1057 - 1063
Main Authors Shi, Yinan, Tzoumas, Georgios, Hauert, Sabine
Format Conference Proceeding
LanguageEnglish
Published IEEE 14.05.2025
Subjects
Aerospace electronics
Area coverage
Autonomous aerial vehicles
Climate change
Hazards
Heuristic algorithms
Monitoring
Multi-UAV system
Particle swarm optimization
Partitioning algorithms
Prevention and mitigation
Robots
Swarm Intelligence
Wildfire
Wildfires
Online AccessGet full text

Cover

More Information
Summary:Due to climate change in recent years, wildfires have become one of the most harmful hazards to the environment and society. In firefighting operations, the early stages are crucial to controlling wildfires successfully. In this paper, we propose an improvement to an existing dynamic space partition (DSP) algorithm by adding an Archimedean spiral to enable wildfire detection in large areas on the scale of California. Compared to the baseline DSP controller, the improved algorithm provides more efficient area coverage with the same number of robots in the simulation. With a swarm of 30 robots, the DSP algorithm with an Archimedean spiral (DSP-A) can identify 87.81% static fires. With the same configuration, the baseline DSP algorithm covered 79.77% of total fires. Furthermore, the DSP-A controller is resilient when the number of robots decreases. When the number of robots in the swarm drops from 30 to 10, the DSP-A algorithm can still cover 70% of wildfires, while the performance of the baseline DSP controller is reduced to 44%.
ISSN:2575-7296
DOI:10.1109/ICUAS65942.2025.11007934