Digital ag, but a lower price tag

FarmBeats from Microsoft hopes to put data-driven agriculture on the table for farmers balking at cost or hurting for bandwidth

Ranveer Chandra doesn’t have a background in agriculture; he has a PhD in computer science from Cornell University and is the chief scientist of digital cloud provider Microsoft Azure Global. Yes, Chandra grew up in India where he would spend his summer and winter vacations visiting his grandfather’s farm, but you won’t hear him wax nostalgic about his time there, and it was a far cry from the technology-integrated farm operations that his work hopes to promote.

“It was one of the poorest regions in India,” he said. There was no electricity, no toilets.”

Those months every year did, however, expose him to some of the most primitive forms of agriculture, and that was at least part of the impetus for him starting the FarmBeats project at Microsoft in 2015.

Why it matters: Cost has been an often-cited barrier against data-driven agriculture, but some Microsoft projects are hoping to lower that sticker shock.

The stated goal of FarmBeats is to provide solutions making data-driven agriculture (DDA) more accessible and more affordable for farmers around the globe.

“The world needs more food to feed a growing population,” Chandra said during his presentation at the Ag in Motion Discovery Plus virtual farm show in July. “Not just food, good food, but nutritious food, and we need to grow this food without harming the planet.”

Data-driven agriculture (the ability to capture large amounts of data from farms and use that data to create more efficient outcomes) is the most promising way to maximize that quality food production, according to the computer scientist.

The concepts of precision and data agriculture will be familiar to many producers, with the practices under its umbrella increasingly touted for higher yields, reduced costs and environmental sustainability.

Producers in recent years may have noted the rise of online or, increasingly, mobile farm management services that promise more efficient input use, data tracking and field-specific trends. Drone and satellite imaging are used to map out field problem spots. New spray technologies hope to improve existing variable-rate practices.

But while the benefits of data-driven agriculture are well known, Chandra notes that these technologies have yet to become widespread.

“The reason they haven’t taken off is the cost of available data-driven agriculture solutions,” he said.

And that’s where FarmBeats comes in.

One of the barriers to affordable data-driven agriculture is farm connectivity. Though it is improving, many rural areas still lack the bandwidth of urban environments. According to an early 2020 survey from the Keystone Agricultural Producers, just under two-thirds of rural households in Manitoba were somewhat or very dissatisfied with their internet and cellular service.

Most farm homes have at least some connectivity to the cloud, but relaying data from the field (often miles away) can’t be done with current Wi-Fi systems. To address this, Chandra and his team have turned to piggybacking on much older technology, called TV white spaces. The technology takes the Wi-Fi signal and overlays it in the “white spaces” between television channels at frequencies that will not interfere with broadcasts. This technique can boost the signal’s range by up to 1,200 per cent. In addition, Chandra notes that because television towers are in cities, and often don’t reach remote rural farms, “If you turn on a television on a farm, often most of the channels will be white noise.

“Even if 20 channels are available on a farm, we are talking of about half a gigabyte per second of streaming capacity,” he said.

The cost of gathering data is yet another challenge. Accurately determining a field’s moisture levels, for instance, would require a sensor every 10 metres, a costly proposition at $100 a pop. FarmBeats has developed artificial intelligence (AI) that combines different datasets (ground sensors, drone images, satellite imagery, etc.) and extrapolates and combines them to paint an accurate picture of what is going on in the fields.

However, while drones are becoming affordable resources here in North America, a farmer in India or sub-Saharan Africa may not spend $1,000 or more for a suitable drone. To address this, FarmBeats went low tech. While a drone might be out of the budget of a farmer in these regions, most have smartphones. Aerial data could be gathered by large helium-filled balloons with a mount for a cellphone and battery pack. The balloons were then attached to a tether and could either be left in one spot or moved around the farm to gather images.

Despite this kind of innovation, the struggle to make precision agriculture more affordable continues. In particular, Chandra pointed to a project called “Strobe,” designed to use a cellphone Wi-Fi chip as a sensor. As opposed to a sensor costing $1,000 or more, the user would bury a relatively inexpensive Wi-Fi receiver in the field. The system measures moisture and soil electrical conductivity (EC) by mapping signal strength and propagation time of signals received by the system’s different antennae. Those differences correspond to how wet or saline the soil is.

Microsoft is also transitioning FarmBeats into a product, “Azure FarmBeats,” that has been available for public review at since November 2019.

About the author



Stories from our other publications