Poop sensors, drones, and robots: What automation looks like at the farm of the future

A visit to Mount Vernon in Virginia revealed some new possibilities in making agriculture smarter.

At first, the horse that greeted press attendees at an October 20 event at Mount Vernon looked decidedly analog. But as a Virginia Tech professor turned Mikey around, an unusual bit of technology came into view: a small box attached to a wrap around his tail.

The use case involved poop.

“We’re in the middle of the data revolution,” said Robin White, an assistant professor at Virginia Tech’s Department of Animal and Poultry Sciences and its Center for Advanced Innovation in Agriculture, before explaining that this pod included sensors to detect the telltale motions of the horse going to the bathroom.

“It seems like it might be an invasion of Mikey’s privacy,” she joked. But manure management matters at farms, where more data could help to reduce their fertilizer applications and limit runoffs into nearby waterways.

White next pointed to a less obvious sensor attached to the horse’s bridle that recorded Mikey’s activity and heart rate.

“We’ll know when he’s resting and eating and when he’s perhaps playing in the field with his friends,” he said. “We can improve management of their health.”

Low-stress handling could pay bigger proceeds in cows and other “food animals,” as White put it, but horses make for easier beta testers: “They’re so used to having devices attached to them.”

The smart-farming demo then moved outdoors to a field near a recreation of the 16-sided barn George Washington designed to allow more efficient threshing of wheat by horses treading on sheaves to separate the heads of grain.

Virginia Tech decided to stage these exhibits at Mount Vernon not just because of the photogenic scenery, but because of its history of agricultural innovation under Washington, who was an early adopter of such practices as crop rotation and composting. Washington, however, could conduct those experiments without having to pay the wages of enslaved workers, a fact made plain in today’s exhibits at Mount Vernon.

A few feet from a row of kale in that field, a DJI Phantom 4 Pro quadcopter drone sat next to a four-wheeled Clearpath Husky “field research robot” with a Lidar sensor mounted on top.

Hasan Seyyedhasani, an assistant professor of Automation and Connected Technologies in Virginia Tech’s School of Plant and Environmental Sciences, explained how two drones could be better than one in an agricultural context.

While farmers have been using drones to check their crops for years now, his vision is to have robots working collaboratively. For instance, aerial hardware like the DJI drone could use their color or thermal cameras and highly-accurate real-time kinematic positioning GPS to provide intelligence for the Husky robot on the ground, helping it to gather more up-close data and perform tasks like pruning or harvesting.

“We fuse these two sensing systems together to collect comprehensive information,” he said. “Hopefully in the next five years we can have a real automated farm.”

But there’s still plenty of room for improvement in just the data-gathering part of drone agriculture. Another Virginia Tech researcher noted over email that he’s seen promising results from using a DJI drone equipped with 4K cameras to inspect grapes at the university’s research winery in Winchester, Virginia.

“Even in my small vineyards, I was able to find some information that could have been missed if I did not fly my drone,” he said. But he complained that the software he’d purchased kept yielding misaligned images and needed better documentation.

Advancing to a state of agricultural automation in which drones lead other drones will require tackling many other problems. Battery life may be the biggest among them; Seyyedhasani said the Husky only lasts two to three hours on a charge.

Bandwidth can be an even bigger obstacle in many agricultural settings. A study released in October by the Benton Institute for Broadband and Society, a Wilmette, Illinois, nonprofit, cited 2019 U.S. Department of Agriculture data showing 22% of farms were limited to a digital subscriber line (DSL), a slow phone-based technology; 18% relied on smartphones; 3% still limped along on dial-up access; and 18% had no internet connectivity at all.

But the promise of using drones and robots to amplify a farmer’s senses—in effect, returning a hands-on level of care to industrial-scale farming—remains real. As White put it: “We have the opportunity to leave those animals in an environment as natural as possible and still understand their day.”