When IoT means Internet of Trees | Avnet Silica

Display portlet menu

When IoT means Internet of Trees | Avnet Silica

Display portlet menu

When IoT means 'Internet of Trees'

Different types of Dryad sensors

An hour can make all the difference. In an hour you can weed the garden, mow the lawn, plant a tree. In that same hour, as gardeners who make compost may know to their cost, a smouldering pile of biomass can easily burst into flames. When this happens in a forest, it can start a fire that ravages thousands of square kilometers and costs hundreds of millions of dollars to control.

Being able to discover forest fires early is an increasingly urgent issue. According to Cal Fire, California’s firefighting agency, nine of California’s 20 largest fires to date have occurred since 2020. California’s Dixie fire, which began in mid-July 2021, grew to consume almost 4000km2 of forest by the end of September. According to the New York Times, authorities spent more than $610 million over those three months to bring it under control. There have been similarly extreme fires over the past five years in places as far afield as Australia, Greece, and Portugal. And it’s a growing issue: the map below shows just how much of the planet’s surface is now extremely vulnerable to wildfires.

 

Wildfire vulnerability map
A map of global vulnerability to wildfires (click to enlarge)

The ability to find and suppress forest fires before they take hold is therefore of immense value, but it’s a hard problem to solve given how large forests can be, the lack of infrastructure within them, and the challenge of sensing the earliest stages of a fire before it bursts into flame. A Berlin-based start-up has decided it is up for the challenge and is addressing it by developing advanced sensing nodes and highly distributed networking technology. Although Dryad’s initial focus is to detect fires, it expects that its system will also be able to monitor other aspects of forestry, such as soil moisture, tree growth, sap flow, and even illegal logging. The same technology should also enable air and water quality monitoring over large areas.

Listen to Dryad on the We Talk IoT Podcast


Wildfires across the globe are worsening. They are growing larger, becoming more intense and frequent, and even creating their own weather – from Australia to North America, from the Arctic to Africa. According to a United Nations report, Climate change and land-use change are projected to make wildfires more frequent and intense, with a global increase of extreme fires of up to 14 percent by 2030, 30 percent by the end of 2050, and 50 percent by the end of the century.

For this episode – which will air for Earth Day – we have invited Carsten Brinkschulte, CEO of Dryad. We will discuss how sensors and artificial intelligence can predict wildfires and protect ecosystems.

A network of networks

Dryad’s Silvanet Wildfire Sensor on a tree bark
Dryad’s Silvanet Wildfire Sensor is designed to operate for up to 15 years from onboard solar cells (click to enlarge)

Dryad’s Silvanet is made up of five major components: three pieces of hardware, a proprietary RF networking technology that extends the reach of the LoRaWAN protocol, and a cloud analytics and alerting platform. The whole system has been designed to overcome the overarching constraints of the lack of power and connectivity in forests.

The Silvanet Wildfire Sensor

The sensing node for Silvanet is a solar-powered device that uses a sensor to measure temperature, humidity, and air pressure, as well as the concentrations of gases including hydrogen, carbon monoxide, carbon dioxide and volatile organic compounds at sensitivities of parts per million. A low-power microcontroller from STMicroelectronics, which integrates support for the LoRaWAN protocol, runs algorithms based on machine-learning (ML) strategies to process sensor data to reduce the energy burden of communicating it in raw form over the LoRaWAN RF network. The Wildfire Sensor is designed to run for up to 15 years without maintenance or a battery, which avoids introducing potentially flammable materials such as lithium into the forest.

 

The Silvanet Mesh Gateway on a tree
The Silvanet Mesh Gateway connects multiple LoRa star networks into a daisy-chain with greater reach (click to enlarge)

The Silvanet Mesh Gateway

The Silvanet Mesh Gateway uses the LoRa concentrator chipset from Semtech to implement the low-power, low data-rate LoRaWAN protocol, which is used to communicate with the Silvanet Wildfire Sensors and with third-party sensors that properly implement the standard. LoRaWANs usually have a star architecture, in which each end-node device, such as the Wildfire Sensor, has a direct relationship with the LoRaWAN gateway and through it to the internet. Although this approach is simple and widely implemented, it limits the reach of LoRaWAN networks to a circle whose radius is defined by the RF range from node to gateway. This, in turn, limits how far a LoRaWAN network can reach into a forest and hence its utility for early fire detection over large areas.

Dryad’s key innovation is to extend the LoRaWAN with proprietary technology that links the gateways to each other to form a mesh of star networks, which together can reach much more deeply into a forest. The Silvanet Mesh Gateways are also solar-powered and designed to be mounted on trees at about 3m above the ground.

Carsten Brinkschulte, CEO of Dryad, said: “The mesh is the magic of what we're doing. Standard LoRaWAN has a star network architecture so you can buy dozens of LoRa gateways, but they all have one problem. They all need to be connected directly to the internet. We place a mesh gateway, based on the Semtech solution, every three kilometres and that extends the coverage of the LoRaWAN network deep into the forest. Then we have a proprietary protocol to connect from mesh gateway to mesh gateway.”

 

The Silvanet Border Gateway

Silvanet Border Gateway with Dryad logo
The Silvanet Border Gateway brokers communications between the mesh of LoRa networks and the internet (click to enlarge)

Silvanet WildFire Sensors and Silvanet Mesh Gateways together can create an extensive ‘Internet of Trees’ among themselves, but they need a Silvanet Border Gateway to reach the wider internet.

The Silvanet Border Gateway is designed to be placed at the edge of forest, perhaps in a visitor’s car park or nearby building, to link the mesh network to the wider internet and then on to the Silvanet Cloud Platform. Each Border Gateway can act as a LoRa gateway, using the Semtech chips. It can also implement 4G connectivity using the LTE-M standard and fall back to 2G and GPRS protocols if necessary. For even more remote deployments, the Silvanet Border Gateway can connect to the internet using the SWARM satellite network.

The Border Gateway can be powered by solar cells, if necessary, although there is also support for wired power and Ethernet connectivity if that is available at the border site.

The Silvanet Cloud Platform

The sensor nodes, mesh networks and gateways are designed to bring pre-processed data from throughout the forest to the Silvanet Cloud Platform, which provides a complete solution for wildfire detection and monitoring the health and growth of the forest. The Silvanet application groups sensor devices and gateways either by geographical area, or by characteristics. Details related to each sensor, especially their fire status, are shown on a map.

Wildfire vulnerability map
The network architecture of Silvanet (click to enlarge)

 

The engineering challenge

Dryad has an experienced team of founders with a wide variety of relevant skills but implementing Silvanet is still a challenge. Along with the hardware, it includes firmware development for energy-constrained data analytics, support for over-the-air updates, IoT security, implementing the LoRaWAN (using Semtech parts) and ensuring standards compliance, creating the proprietary links between Mesh Gateways, integrating multiple forms of RF connectivity in the Border Gateways, and the cloud platform and analytics. All this must be achieved in equipment working at low power, with solar panels that are often in shade, in the difficult RF environment of a forest, and with the ruggedness to ensure long operating lifetimes. And, especially for the Wildfire Sensors, it must be done at the lowest possible cost to enable wide deployment.

Brinkschulte said: “If you want to put something else on the list, RF communications in the forest is challenging because every tree is an obstruction full of water that absorbs radio waves.”

There are other challenges with the RF implementation. Both the LoRaWAN and the 4G radios in the Border Gateways must be on all the time because they must be ready to accept data from the Wildfire nodes or inbound messages from the cloud. If 4G LTE-M connectivity is unavailable, falling back to 2G protocols also boosts power consumption.

“A key challenge for us is the ultra-low power design of the sensor hardware,” said Brinkschulte. “We need to work in an environment where we have very little energy input, yet we have to do something complex. We need to scan for gas compositions and run ML software, it needs to work reliably, and we need to detect fires very quickly. These are contradicting goals, and so one challenge is to select the right ultra-low power components. 

“At the same time, price is key because we will need to put hundreds of thousands if not millions of Wildfire Sensors in a forest for the system to work, and the price of each component at that scale determines the system cost. We can’t choose a component that performs extremely well but pushes the Bill of Material up by €5 and kills the product.”

 

Partnering for market insights

Dryad has worked with electronics component distributors Avnet Silica to explore its options for components that can meet these tough challenges.

For example, the initial design uses an STM 32 WL microcontroller because it offers ultra-low power operation and integrates LoRa radio, while having the performance to analyse sensor data with enough sophistication to detect wildfires early, without creating false positives.

 

“We really appreciate the competent-level discussions we're having with Avnet Silica. They are bringing a lot of ideas and components which may optimise the performance of the system.”
Carsten Brinkschulte, CEO of Dryad

 

“We have to do that in the sensor because we do not have enough bandwidth to transmit all of the sensor data to the analytics in the cloud,” said Brinkschulte. “We really appreciate the competent-level discussions we're having with Avnet Silica. They are bringing a lot of ideas and components which may optimise the performance of the system.”

Dryad is already looking at potential optimizations of the hardware designs with even lower power and even lower costs, “and that's a constant discussion that we're having on which chipsets to use.

“We need to have it really low cost and ultra-low power because if you halve the power consumption of the device you can halve the size of the solar panel and energy storage, which are key price-driving factors of the sensor node.”

Dryad is using super-capacitors to store energy from the solar cells, rather than introducing potentially toxic and flammable rechargeable batteries into the forest.

“If you can reduce power consumption you can reduce the capacity of the super-capacitors, which would also drive down the price,” said Brinkschulte. “Super-capacitors are not yet commoditized, and so selecting the right components is definitely something where Avnet Silica has the competency to advise us on selecting which may perform better, have better availability, or a better price.”

Dryad’s use of the parts is also a little unconventional, in that super-capacitors are more widely used to provide enough energy to enable systems that have suddenly lost mains power to do an orderly shutdown.

“There's lots of specific things that we need to do to charge the super-capacitors from the solar cell, so we need charging circuits that actually do that and that is a bit more challenging because it's not the standard way of doing things,” said Brinkschulte.

Avnet Silica is also helping Dryad understand current supply constraints for various parts, caused by supply-chain issues ranging from the pandemic through to factory fires. It has been trying to ensure that Dryad chooses components that will be available in enough volume to allow it to rollout Wildfire sensors and Gateways in volume.

 

“When you move to your go-to-market strategy and try to roll out your network and the quantity goes up, you need to have the microcontrollers and other parts delivered on time,” said Ingo Seehagen, senior field application engineer at Avnet Silica.
Ingo Seehagen, senior field application engineer at Avnet Silica

 

“When you move to your go-to-market strategy and try to roll out your network and the quantity goes up, you need to have the microcontrollers and other parts delivered on time,” said Ingo Seehagen, senior field application engineer at Avnet Silica.

 

Dryad’s vision

According to Brinkschulte, wildfire sensing is just the first application of the Dryad networking technology.

“This is a generic IoT communications infrastructure for the forest,” he said. “There are tonnes of applications that forestry can benefit from once our infrastructure is in place. Soil moisture is one of them, tree growth, sap flow, anti-logging alerts using sound detection, and emergency communications for forest workers.”

Forestry equipment maker Stihl is one of Dryad’s backers.

“We create an IoT network infrastructure that digitises the forestry industry, and I guess that's of strategic interest to them,” said Brinkschulte.

The deal also gives Dryad access to the electronics manufacturing services, and all the tacit knowledge and expertise it has in making rugged electronics for use in challenging environments.

“That is an advantage for us because if the mothership uses them, and the baby uses it as well, that means they give us good attention.”

Dryad continue to rely on Avnet Silica as it ramps production to hundreds of units to populate 10 pilot projects around the world.

 

“What is impressive is the amount of attention that we are getting and the professional support that we are receiving, even though we're still in our very early stages. Avnet Silica’s support is an investment in our future and that's very impressive from our perspective.”
Carsten Brinkschulte, CEO of Dryad

 

Brinkschulte said: “What is impressive is the amount of attention that we are getting and the professional support that we are receiving, even though we're still in our very early stages. Avnet Silica’s support is an investment in our future and that's very impressive from our perspective.”

Kick start your IOT journey with Avnet’s IOT Connect

Follow Avnet Silica on LinkedIn

When IoT means Internet of Trees | Avnet Silica

Display portlet menu

Sign up for the Avnet Silica Newsletter!

Stay up-to-date with latest news on products, training opportunities and more!

When IoT means Internet of Trees | Avnet Silica

Display portlet menu
Related Articles
A cnc machine is milling a piece of metal
Explore options for choosing an optical rotary encoder for motion control and position sensing
July 24, 2023
Rotary position sensing is used in motor control and user interface panels. It comprises various encoding methods. Optical encoders can offer advantages. We look at the options and how to choose the right one for your application.
three electric motors on glass table
How are magnetic rotary encoders used in industrial automation?
July 24, 2023
The Hall effect is used in rotary encoders to monitor position, speed and direction in electric motors and rotary dials. The encoder designs vary, so it is helpful to understand the main features and how they affect performance.

When IoT means Internet of Trees | Avnet Silica

Display portlet menu
Related Events
Renesas Tech Day Germany
Date: May 14, 2024
Location: Germany, Stuttgart