Firefighting has always been a difficult and dangerous activity. Even with advances in equipment, materials, and training, teams battling blazes frequently are left to guess about wind conditions, fuel sources, toxic materials and other factors.
In the U.S., the National Fire Protection Association (NFPA) reports that fires resulted in 3,500 deaths, over 15,000 injuries and $21.9 billion in property loss in 2020. What's more, wildfires are growing larger and hotter, and urban fires represent enormous risks to lives and property.
However, the battle lines are shifting. Low-tech trucks, ladders, and hoses, while not going away, are taking a back seat to digital technologies like artificial intelligence (AI), machine learning, satellite imagery, drones, digital thermography and others. These tools are helping firefighters identify the dynamics and characteristics of a fire—and respond more effectively.
Says Wai Cheong Tam, a mechanical engineer in the Fire Fighting Technology Group at the National Institute of Standards and Technology (NIST), "Digital technology enhances traditional experience-based firefighting with data-driven firefighting. It can help save property and protect lives."
Out of the Flames
Digital technology increasingly appears on the front lines of firefighting. For example, the Office of Wildland Fire, part of the U.S. Department of the Interior, has adopted a desktop software application called FlamMap to predict the spread and intensity of wildfires. It relies on remote sensing technology that incorporates lidar, drones, automated weather stations, and satellite imagery to understand factors such as topography, vegetation, weather, soil, and past fires.
"This type of technology also helps us calculate the immediate consequences of a fire, from tree mortality to hazardous smoke and damaged soil," says Erin McDuff, a public affairs specialist for the Office of Wildland Fire.
At the University of California, Riverside, Richard A. Minnich, a professor in the Department of Earth and Planetary Sciences, builds heat and fuel maps from satellite and aerial imagery. His goal is to gain a better understanding of human impact on land. However, the benefits of his research do not stop there. "These maps provide valuable insights into how much fuel exists and how fires will burn," he explains. "With this information, it's possible to know how to manage land and manage responses (to wildfires) more effectively."
Digital technology is also having an impact on a smaller scale. Drones increasingly are used to survey fires and provide situational awareness and thermal assessments, aid in rescues, and reach difficult places that could put humans in danger. Hundreds of fire departments now rely on drone technology, which can incorporate special lenses and thermal imaging that sees through smoke to gauge the heat of a fire. Some have also begun to use instrumented robotics.
The Internet of Things is also fueling advances in firefighting. The Office of Wildland Fire has introduced high-speed broadband with satellite and edge connectivity to improve firefighter coordination and safety. "The installation of remote tracking systems enables wildland fire managers and dispatchers to view the locations of vehicles and personnel in real time, improving situational awareness and rapid decision-making," McDuff explains.
The U.S. Forest Service and others are turning to virtual reality systems to train firefighters. Meanwhile, researchers continue to explore the use of augmented reality, wearables, and embedded sensors in clothing that could display current conditions and identify hazardous situations.
AI Ignites Advances
One of the hottest areas of research and development surrounds the use of AI and machine learning (ML). Tam says more advanced analytics and AI systems can provide actionable information to firefighters. This can help them avoid dangerous scenarios and extreme fire events, including so-called "flashovers" that lead to a rapid acceleration of fire, often exceeding 800°C, within a few seconds. "This is the second most common cause of death and injuries among firefighters," he says.
Traditional numerical approaches cannot model risks and conditions effectively. The factors surrounding these events are simply too complex, and the models can lag actual events by 5 minutes or more. However, using a more advanced AI algorithm developed through synthetic machine data training, NISTs model can now predict a flashover in real time with about an 89% level of accuracy. "The model uses limited temperature signals from heat detectors embedded in the building structure before they fail," Tam says.
The framework, called P-Flash, could eventually be built into mobile networks used by firefighters but also could lead to sensor technologies that would enable far more sophisticated smoke detectors, Tam explains. In fact, NIST is currently working with the fire protection industry and companies in the building automation space to steer the concept into the real world. NIST is now testing the technology with existing fire protections systems, such as heat detectors.
NIST and others also are exploring broader smart firefighting technologies, including wearables, embedded sensors in clothing, and new types of augmented reality glasses. Says Tam: "There are many uses for digital technology in firefighting."
Samuel Greengard is an author and journalist based in West Linn, OR, USA.