IoT Sniffs for Smoke to Prevent Wildfires

By Claire Swedberg

image_pdfimage_print

 

  • Dryad’s fire detection system now leverages HF RFID chips to enable shippers to set the LoRa frequency on the device according to where the device is used.
  • The sensors are now in place across North America, Europe, Asia and South America to detect smoke before a fire burns out of control.

With another fire season underway in the Northern Hemisphere, the forests are more vulnerable than in years past.

According to the National Interagency Fire Center (NIFC) the wildland fire potential for the U.S. shows fire activity predictions above normal for this summer in portions of the Southwest, the Great Lakes and areas of the Northwest and California in the late summer. And fires are getting bigger and more complex. That means detecting them before they burn out of control is paramount.

Providing one technology-based solution is IoT startup Dryad Networks, expanding its wireless wildfire detection solution that uses LoRaWAN sensors and gateways, within proprietary mesh networks, to offer a single point of connection between sensor data and the cloud.

Beginning this spring, the system also leverages a passive HF RFID chip, built into each sensor or gateway, to enable the devices to be wirelessly set to the appropriate LoRa frequency based on where they will be used.

A second application for the RFID chip aims at installers who can tap their phone to the device as they attach it to a tree or other infrastructure, to create a GPS location for each device as it is being installed, explained Carsten Brinkschulte, Dryad Networks co-founder and CEO.

IoT Network in Fire Prone Areas

Many fires ignite at the point where human activity meets wild-lands—in fact 85 percent of fires start at the intersection between humanity and nature. In some cases, technology can help identify when a fire is starting so that it doesn’t spread out of control.

To help combat this, Dryad builds networks that often cover points of human activity, such as roadsides, that then extend into the forest. Unlike standard LoRaWAN based systems, the network from Dryad uses a proprietary mesh network that can carry a signal across hundreds of sensors via gateways, even in remote areas.

Dryad launched in 2020 to provide IoT-based wildfire detection, with small tree-mounted sensors and gateways and cloud-based Software as a Service (SaaS) to manage sensor data. Machine learning on each device helps the system identify whether conditions being detected require an alert.

Last year the company sold 20,000 sensors for about 50 deployments in southern Europe, North America, Thailand, Indonesia and Chile.

Tree Mounted Sensors

To deploy the technology, installers attach the 19 by 4.2 centimeter sensors to trees with a wooden nail, drilling a small hole in the tree trunk or branch. The tree will seal the hole with sap and the solar-powered sensors can then operate on that tree for about ten years.

Behind its IP67-rated, plastic membrane, the device has a gas sensor that picks up hydrogen and carbon monoxide and then employs its built-in machine learning in the MCU to identify a fire.

“By using ML signal processing to identify gas patterns, Dryad can scale a deployment into hundreds or thousands of units  that only send data when necessary, thereby ensuring users never overload their network,” said Brinkschulte.

Gateway devices are deployed to receive LoRa transmissions from the sensors, attached to trees with a bracket every two miles or so to create a mesh network. Like the sensors, the gateways are battery free, and are powered with a solar panel and super capacitor for energy storage. That’s key, said Brinkschulte, “because we didn’t want to bring potential fire starters into the forest.”

The proprietary mesh network enables one installation to leverage a single cellular connection to a server, thereby reducing the need for multiple backhaul devices. More traditional LoRaWAN solutions require that each gateway has its own connection to the Internet.

RFID Enables Automated Frequency Setting

The latest version of the devices includes 13.56 MHz HF RFID tags for wirelessly setting the frequency for each product as well as for provisioning during installation.

Use of RFID streamlines manufacturing for Dryad, by enabling a single product to be set to operate in any part of the world. LoRa devices must operate at different frequencies in different parts of the world: 868 MHz in Europe, 915 MHz in the U.S. and 923 MHz in Asia.

Until now, Dryad produced three different versions of its products to meet those regional requirements. “That was very inefficient,” Brinkschulte pointed out, and could be a challenge because it required anticipating the volume of orders from each part of the world.

Single Device

Now the company builds a single device that can be set to operate at any of the three different frequency bands. In that way the company can streamline and optimize its manufacturing and warehousing according to one SKU.

The sensors and gateways are fully packaged and sealed in the cardboard boxes prior to shipping. When they are prepared for shipping, warehouse workers use a mobile device such as smart phone or tablet to interrogate the RFID chip in the device.

They can then assign the frequency for that device, based on where it is going (such as Asia) and write that frequency to the RFID chip. When the device is received by a customer, as they power it on, it automatically captures the setting written to the tag and updates to the proper LoRa frequency.

Provisioning Sensors as they are Installed

The RFID tag provides an easier way for installers to provision the devices in the field as well. Traditionally they have used a QR code on the sensor, by taking a picture of the code and using the GPS location in their smartphone to link the location to the device.

With new devices, however, the installer taps their phone near the sensor or gateway and automatically links it to the geolocation. By using RFID instead of QR codes, the company reports the deployment is faster with fewer errors.

Over the Air Updates

Dryad’s cloud application, running on a cloud-based server, can not only manage sensor transmissions and send alerts, but also keep the system up to date. The system is capable of receiving updates over the air via LoRa so that Dryad can update the software in the sensors.

“It’s challenging to accomplish firmware updates over LoRaWAN,” due to the small data sets LoRaWAN typically can manage, said Brinkschulte, attributing the feature to years of engineering by Dryad.

“We need that capability because the machine learning model in the sensor has to be updated.  So the firmware update over the air is an essential feature of the product,” he said.

The sensors sell for less than $100 each. Customers are typically power line companies, forest services, railroad firms and national parks.

Read More: