New AI Wildfire Sensors Deployed in Hawaii

Seven months after the deadliest wildfire in U.S. history devastated the island of Maui, a new system of artificial intelligence-powered sensors is being installed in Hawaii to provide early detection and warning when fires break out, officials announced Friday.

The smart sensors, developed by the Department of Homeland Security and U.S. Fire Administration in partnership with tech firm N5 Sensors, combine thermal imaging, gas detection and AI algorithms to sniff out wildfires in their earliest stages. Able to detect minute changes in gases, chemicals, and particulates in the air, the cloud-connected devices promise to give firefighters precious extra minutes or even hours to respond to budding fires before they rage out of control.

“These sensors aren’t just looking, they’re also smelling the environment at the same time,” said Maui Fire Chief Brad Ventura. “That’s a very important part of this project, because at night, when the thermal imaging capability is working and the gas detection is working, we’ll be able to earlier detect fires.”

Hawaii will be the first state to receive 80 of the next-generation “Beta” sensors as part of a nationwide rollout to deploy 200 of the devices in high-risk wildfire areas across the country this year. An initial batch of 20 sensors will be switched on in Maui on April 8, concentrated in the wildland areas bordering the densely populated communities of Kihei, Wailea and Lahaina.

A local news segment from Hawaii News Now.

The wireless sensors, small enough to attach to existing utility poles and streetlights, are solar powered and transmit real-time data around the clock via cellular networks. When the AI detects concerning readings, it sends text message and email alerts to local fire officials.

Over time, the machine learning algorithms will get smarter at distinguishing between benign events like backyard barbecues and dangerous flare-ups, officials said. The system will also learn the unique characteristics of Hawaii’s environment.

“You have volcanic ash. You have other conditions,” said Homeland Security’s Dmitri Kusnezov. “Smart sensors like this will learn what does the air feel like here. And what does an anomaly look like?”

While not a panacea, officials say the sensors represent a powerful new tool in the fight against increasingly ferocious wildfires fueled by climate change. On Maui, where 115 perished in an inferno that incinerated the town of Lahaina last August, the system aims to avoid a repeat of the catastrophe.

“As firefighters and public servants, we often vow never again,” said U.S. Fire Administrator Lori Moore-Merrell. “The impact of climate change-driven drought conditions, the warmer conditions, extremely unusual weather, storms, high winds, all of these conditions are very likely to occur here again.”

A local news segment from KHON2.

The AI sensors were developed over the past five years, building on earlier work to create high-tech flood detectors. After testing 200 “Alpha” wildfire sensors in 11 states, the new Beta version boasts improved solar panels, wind detection, and the ability to operate in areas with spottier cell coverage, officials said.

Following the initial deployment in Maui, the second wave of sensors will be installed on Oahu, Kauai and Hawaii Island in the coming months. Homeland Security is providing the devices to the state at no cost for now, but expanding coverage may require investments in more infrastructure, according to Kusnezov.

While not failsafe, Ventura, the Maui fire chief, said he believes the sensors will make a difference this wildfire season. When every second counts, firefighters will take any technological edge they can get.

“We’re looking at our wildland urban interface, which is the area outside of town, where a fire would be starting and then working its way towards town,” Ventura said. If the sensors work as intended, emergency responders should have a jump on the flames long before 911 calls start coming in. “That’s something that these sensors will be picking up prior to maybe somebody seeing it.”