Autonomous Drone Swarming for 3D Mapping of Atmospheric Particle Transport from Micrometer to Kilometer Scale

Understanding particle dispersion from wildfires is essential for modeling its impact on regional air quality and global climate. However, field characterization of the complete spectrum of particle behaviors—from individual particles at the microscale to kilometer-spanning plumes—presents significant technical challenges. Here, we introduce an autonomous drone swarm system to address this challenge. The system features a manager drone that coordinates multiple worker drones, each equipped with computer vision cameras for navigation and plume imaging, and holographic microscopy sensors for detailed analysis of particle properties such as concentration, morphology, and types. Specifically, the manager drone analyzes particle plume structure in real-time and uses flow-intelligent AI to guide the worker drones, optimizing their formation and navigation paths to reconstruct 3D smoke plume geometry and particle distribution. This system, developed using a simulated environment incorporating fluid dynamics and drone operation, was successfully deployed during the Cedar Creek prescribed burn experiments, demonstrating its efficacy in real-world applications. Our work provides a versatile tool not only for the fundamental study of atmospheric particle transport phenomena, such as dust, volcanic ash, and pollen dispersion, but also for practical applications such as forest fire response and other environmental monitoring tasks.