NAU leading NSF grant that looks at the potential for drones in responding to forest fires

Three NAU researchers won a National Science Foundation grant to study the use of drones in responding to forest fires.

Fatemeh Afghah, an associate professor in the School of Informatics, Computing, and Cyber Systems (SICCS), is the lead principal investigator in the $1.2 million grant in collaboration with Georgia Tech, the Desert Research Institute and the National Center for Atmospheric Research. The co-PIs are Peter Fulé, a Regents’ professor in the School of Forestry, and Abolfazl Razi, an assistant professor in SICCS. 

The research, which builds on work Afghah and Razi have been doing for almost three years, is exploring ways to use Unmanned Autonomous Vehicles (UAVs) in tracking forest fires. Their goal is to give firefighters a better situational awareness about the fire environment; provide up-to-date information on where the fire is; and help fire responders form reliable predictions about the fire activity. 

“In the Southwest, especially Arizona, fire is a very big deal,” Afghah said. “When you get to the fire, it’s going to have a huge impact in informing the strategies of first responders to control fire, helping the survivors and coming up with evacuation plans in time.”

Drones are autonomous and less expensive than helicopters; when equipped with cameras, operators on the ground can observe the fire, predict its direction, look for and even communicate with people who are at risk and help inform decisions about evacuation routes. Drones can collect real-time video streaming, track and detect object and track survivors.

All of this happens without putting fire officials in the line of fire. The UAVs will have thermal cameras as well, so they can get pictures at night or in areas of heavy smoke. It’s increasingly important as the number and severity of fires in the Southwest increases due to climate variability and the accumulation of fuel in the forests.

“It can potentially give us the fastest solutions,” Afghah said. “We are designing technologies that will give the drones the better view of the fire. We’re looking for ways to create safe and fast evacuation plans, and we’re also working on low-computation fire modeling techniques.”

Fulé, who is a fire ecologist focused on climate warming and increased fire severity, said he is always looking for better ways to understand and manage wildland fire to gain the most benefits from environmentally helpful fires and limit dangerous fires.

“Drone technology, like aviation and satellite technology, is another step forward in providing timely, detailed information to managers, helping them make safer and more efficient decisions that benefit communities and our forests and watersheds,” he said.

Razi’s role is to develop machine learning tools to model fire expansion based on the imagery the drones collect. His team will exploit key topography and geographical features using advanced image processing methods, then develop integrated modeling that incorporates these features along with environmental factors such as wind, temperature, humidity, terrain and vegetation.

“This data-driven model will help characterize fire expansion as a dynamic system, generalizable to similar scenarios,” he said. “The results of this research can be used to predict active and future fire behaviors and help the ground force and firefighters to better use their fire management resources. For instance, they could focus on areas that fire is expected to grow faster and affect the surrounding residential area.”

In addition to the academic collaborators, the team is working with the Flagstaff Fire Department and the Kaibab and Coconino national forest administrators. In cooperation with fire managers under approved conditions, the researchers are able to put drones into both prescribed and naturally occurring fires and work with the firefighters on the ground to help them use the equipment, interpret the data received and get feedback on its user-friendliness.

The drones right now are not being looked at to actively fight fires; they’re not large enough to carry water tanks or other equipment. Neither are they looking at managing fires with drones in this study. However, they can take in emergency equipment to people if needed, such as bringing in a cellphone, water or other supplies.

“The field tests will progress from a highly controlled environment to a realistic wildland fire environment,” the grant application reads. “We expect that the outcomes of our project and experimental validation will enable a transformational and practical framework for managing wildfires that typically occur in rugged and remote terrains where land access is slow, safety zones are scarce, and the air access is limited by flight hazards while specialized firefighting aircraft have high cost and limited availability.”