Sensors, mesh networks and AI can be the future of controlling wildfires
As has become more and more common, wildfires have recently burned large tracts of land in several parts of the world, including the Western US, Brazil, Greece, Italy, Russia and Turkey.
According to the US National Interagency Fire Center, by the end of July 2021, large wildfires burned close to 3 million acres in the US, destroying forests, wildlife and homes. The fires severely lower air quality in the immediate location of the blaze and the smoke can drift as far east as Chicago and even New York. Wildfires have a double whammy of being a large source of CO2 emissions, on top of depleting the ability of forests to absorb carbon dioxide.
Although preventing all wildfires is neither possible nor necessarily desirable, the earlier a fire is detected, the more likely it can be contained. But early detection in deeply forested regions is challenging, relying on aircraft and satellites to monitor areas, and using weather reports as well as seasonal fire lookouts to concentrate on particular areas. It’s often like finding a needle in a haystack until things are already out of control.
It’s therefore encouraging that new connected technologies are being set up with advanced early-warning systems, in the hopes of limiting the severity of fires. A growing number of innovative companies that develop systems using the Internet of things (IoT) see the potential in using wireless sensors to give early warnings of wildfire activity. Two leading companies in this area are Dryad and LADSensors.
Dryad’s solar-powered, LoRaWAN-based sensors hang off trees, creating a mesh network infrastructure. As the sensors are powered by solar energy, they need no batteries to charge or replace. The sensors talk to base stations, which in turn talk to each other, passing on the message until it reaches a gateway with either a cellular or hard-wired Internet connection.
This mesh network architecture allows the system to cover vast areas, like thousands of square kilometres, without each base station needing its own Internet connection. The sensors use a Bosch BME688 gas sensor chip to detect the gas composition of the air, looking for hydrogen, carbon dioxide and carbon monoxide. Dryad’s sensor then uses machine learning and edge processing to detect the combination of gases typical for a wildfire.
In July, the US Department of Homeland Security successfully tested four prototype technologies for early wildfire detection in California. The test was the second phase of a wildfire sensor technology programme being carried out by its Science and Technology Directorate (S&T), called Wildland Urban Interface. This programme forms part of the agency’s Smart Cities Internet of Things Innovation (SCITI) Labs initiative, bringing together government and private sector companies to find technologies that meet the operational needs of first-responders, and ensure the nation’s critical infrastructure holds up and remains secure.
S&T worked with four industry partners — Ai4 Technologies in California, N5 Sensors in Maryland, Valor Fire Safety in New Hampshire and Breeze Technologies in Hamburg, Germany — to refine and enhance their unique wildfire sensors technologies and platforms. It also connected the companies with the eventual users of their products, including the Federal Emergency Management Agency as well as state and local fire services, to gain valuable feedback as the products are readied for the commercial marketplace. For more on this programme see here.
To be clear, these systems are still in the development and trial stages, but it’s exciting to learn about the practical potential of smart, connected sensors. The only disappointment is that such a promising technology is only being pursued by early-stage companies at present — if a few big tech players really got behind it, it could be so much more effective. The Internet of things, like many other technologies, is still finding its place in the world, going through a phase of hype before breakthrough applications appear. The trees, it seems, will be among them.