Fire Foresight's journey at the XPRIZE Wildfire Challenge

Ten Minutes, One Thousand Square Kilometers, One Shot

Fire Foresight's journey at the XPRIZE Wildfire Challenge

How a global team built around a Tasmanian core is working to catch a wildfire in its first minutes and act before it can spread.

Most bad fires start where no one is looking. A dry ridge. A vacant lot. A stretch of back country with no road and no house for miles. By the time someone sees the smoke and reaches for a phone, the fire is already running. Those first few minutes decide everything that comes after.

That problem is why we entered the XPRIZE Wildfire Challenge.

What the challenge is

XPRIZE Wildfire is an 11 million dollar competition with one goal: stop destructive wildfires by getting ahead of them instead of chasing them. The Gordon and Betty Moore Foundation, PG&E and Lockheed Martin are among its backers. XPRIZE Wildfire runs in two tracks: Space-Based Wildfire Detection and Intelligence, and Autonomous Wildfire Response.

We entered the harder one, the Autonomous Wildfire Response Track. The task is simple to say and brutally hard to do. Inside a thousand square kilometers, find a real fire, tell it apart from the decoys, reach it, and knock it down in under ten minutes, with no person at the controls. 

Ten minutes over that much country is a challenging engineering feat. A lab demo will not suffice. The system must detect the fire, decide what to do, and act on its own, outdoors, in real conditions, fast enough to matter.

Our role in it

Fire Foresight made the cut as one of four finalists in that track, we're currently in Alaska, getting ready for our test.

As global team built around a one core system that finds a fire early, sends a drone to pinpoint exactly where it is, and can move in to hit it while it is still small, our system is built to work inside the same airspace rules everyone else follows.

https://youtu.be/f3_KTJrPy2s

We have taken it out of the lab and into the field, against staged fires into live conditions.

Why we build a system of systems

Most fire technology leans on one tool. A camera. A satellite. A sensor in the grass. Each one is useful, and each one has a blind spot. A camera cannot see past a ridge. A satellite passes overhead and misses the first crucial minutes. A ground sensor only knows its own patch of ground.

So we stopped hunting for one perfect tool. We approach this challenge like a jigsaw puzzle. Weather is a piece. Lightning is a piece. So are the satellites, the smoke sensors, the cameras and the drones. On their own they sit scattered across the table, each showing a corner of the picture. Fire Foresight is the part that fits them together, so you can finally see the whole picture and act on it.

Most fire teams already hold some of those pieces. A camera network along the ridgelines. A few drones in a shed. We do not ask anyone to throw that out and start again. We layer over the top of what they already run. When an agency operates cameras, we turn those feeds into a single confirmed alert with a location on it. Where it has drones, we add the autonomy that lets them go and check a fire on their own. Every piece someone already paid for becomes part of a bigger picture.

The system works the way your own senses do. You do not trust your eyes alone. If you smell smoke but see nothing, you go and look before you act. Fire Foresight lines up weather, lightning, satellite, smoke and camera signals by place and time, and raises the alarm only when independent signals point to the same fire. One signal can be wrong. Several that agree can be trusted. An alert that cries wolf ruins trust with your team so its imperative to get it right. Most detection tools stop at the alert. The harder step, acting on what the system sees, is the one we are building toward in the XPRIZE challenge.

https://youtu.be/wCz8HaT-VcE

In the real world, where we are already deployed, our technology runs at the center of Integrated Fire Camera Networks, where we flag fires in their first minutes, sometimes as much as fifteen minutes before the first call for help. That detection system protects communities, forestry, utilities, mining sites and carbon projects across several countries. Detecting fires early is the part we do every day. Stopping them on our own, in their first minutes, is the harder problem the XPRIZE challenge pushes us to build next.

The path forward

Reaching the finals tells us we are on the right track, and it is only the start. From here we have to prove the whole system in harder conditions, make every layer faster and more precise, and get it into the hands of the people who carry the risk when fire season comes.

The XPRIZE Wildfire Challenge is one of the hardest tests of this idea anywhere in the world, and we are building to meet it.

If you protect land, infrastructure or a community from fire, please reach out to us. There is a good chance you already own a few of the pieces, and we will help you make them work as one.