Researchers at the University of Minnesota Twin Cities have developed an advanced AI-powered drone swarm system designed to meticulously monitor and analyze wildfire smoke plumes. This technology, introduced in August 2025, aims to provide a more precise understanding of smoke behavior and its impact on air quality. The system consists of a central manager drone overseeing a team of four worker drones, each equipped with a 12-megapixel camera on a gimbal, 6000 mAh batteries, advanced flight controllers, and NVIDIA Jetson processors. These components enable real-time smoke detection and adaptive flight, allowing the drones to navigate directly into smoke plumes.
During field tests, the drone swarm captured high-resolution imagery, facilitating the creation of detailed 3D reconstructions of smoke. This data is crucial for analyzing smoke shape, direction, and flow, thereby improving fire behavior models and air quality forecasts. The research, supported by the National Science Foundation and conducted with assistance from the St. Anthony Falls Laboratory, builds upon previous drone systems developed by the team, which already incorporated computer vision for smoke plume tracking. The current iteration enhances this by employing coordinated multi-drone efforts and exploring the use of fixed-wing VTOL drones for extended surveillance.
This advancement is particularly timely given the increasing frequency and intensity of wildfires, largely attributed to climate change. Traditional methods of smoke plume tracking, such as satellite imaging, often lack the necessary resolution for precise data collection. The University of Minnesota's AI drone swarm addresses this by providing high-resolution data collection across large areas at a lower cost than satellite-based tools. The technology's potential extends beyond wildfire monitoring, with researchers exploring its application in tracking other airborne hazards like sandstorms and volcanic eruptions. The system's ability to characterize smoke particles using Digital Inline Holography further enhances predictive models for a variety of environmental events.
The development has garnered attention within the broader scientific community, aligning with global efforts to integrate advanced technologies into wildfire management. For instance, a large-scale exercise in Greece in July 2025 demonstrated the integration of drones into wildfire response protocols. The University of Minnesota's innovation offers a cost-effective and efficient method for real-time data collection, promising to bolster early hazard detection and inform emergency planning, ultimately contributing to public safety and environmental preservation.