It’s a hard truth: although wildfires in North America have increased in recent years, the situation will only get worse. Climate change continues to worsen the problem, putting people, property, nature and animals at risk.
Drones have been a staple of the fight for more than a decade now, and the Fire Apparatus Manufacturers’ Association projects its fleet will grow to 30,000 by next year. Although effective, these systems certainly have their limitations.
“Currently, the military-grade drones used by those fighting wildfires are high-altitude aircraft that fly well above the trees,” notes Andrew Jong, a doctoral student at Carnegie Mellon University. “They can’t fly low because they can’t see through the smoke. We want to fly just above the trees or even under the canopy.
Researchers at CMU’s Robotics Institute are building drones that can navigate through smoke, providing firefighters with a map and escape routes, while drawing attention to dangerous areas.
Retired firefighter Josh Wilkins, who is collaborating with researchers on the project, notes somberly: “I’ve helped bury too many people who died for lack of actionable information.” He adds that firefighters often receive 12-hour-old information when they begin their eight- to 12-hour shift. At this point, the information is outdated and poses a very real risk to first responders.
The project builds on research conducted as part of CMU’s DARPA Subterranean Challenge, which focused on underground navigation for mine rescues. Forestry drones use pathfinding to determine optimal trajectories with obstructed views. The goal is to efficiently build a 3D “digital twin” of the impacted area as quickly as possible.
Wilkins notes, however, that his former colleagues may be hesitant to adopt new technologies. “There is always resistance, especially among firefighters,” explains the second generation firefighter. “But once we show them the safety factors built into these systems and the good data we can collect with drones, I think we will convince them.”
techcrunch
