When AI Listens: How Firewave Outsmarted a Swarm of Cicadas

The power of AI lies in its adaptability. Our proprietary fire detection algorithm continuously learns from its environment - sharpening accuracy, reducing false positives, and responding in real time.

This month, we faced a real-world test. At an XPRIZE semi-final event in Davenport, California, our CTO, Eduard Greenberg, joined a live demonstration of Firewave integrated with Ember Flash’s autonomous wildfire suppression system.

For those unfamiliar, cicadas are insects known for their loud buzzing. While some species emerge only once every 13 to 17 years, most appear annually. The males vibrate their chests to produce a noise that can reach 100 dB - that's as loud as a nightclub, a power drill, or a lawnmower.

Well, it turns out cicadas can sound like a crackling fire and confuse our sensors. But not for long. After experiencing numerous false positives, Eduard recorded the cicada sounds and injected them into the proprietary Firewave AI algorithm. The next day, cicadas were no longer an issue, no false positives occurred. Firewave detected the fire within minutes, in real-life conditions, hearing it over a loud backdrop of cicadas.

That’s the secret sauce that makes Firewave unique. Our system learns the environment of each sensor and, with the power of machine learning, adapts to hear fire over ambient noise - whether it’s traffic, buzzing power lines, or a loud backdrop of cicadas.