Robots Working in Farming and Agriculture
“High-tech farming” is no oxymoron. A contemporary agricultural operation is more likely to resemble Silicon Valley than American Gothic, what with apps that control irrigation, GPS systems that steer tractors and RFID-chipped ear tags that monitor livestock.
And robotics is an increasingly key part of that technological stable.
Robots pick apples, gather strawberries, harvest lettuce and strip away weeds. Drones gather aerial images that help farmers quickly assess crop health. And robotic greenhouses are sprouting up thousands of miles away from traditional farmland regions, growing vegetables in the backyards of high-consumption urban markets.
It all comes at a time when growers face a costly, long-term labor shortage and — with the global population expected to rise from 7.7 billion to 9.7 billion in just over 30 years — food demand is poised to rise significantly.
Here’s how robots can help mitigate some of those challenges.
On its face, crop harvesting seems ripe for automation. It’s physically taxing and highly repetitive — the kind of labor that's often most effectively targeted in the robot revolution. (See factories, manufacturing, mining, logistics processing.) But that’s not necessarily the case.
Picking crops also requires manual dexterity and a delicate touch. Many fruits bruise easily in the heat, and leafy vegetables are easily torn. And most robots just aren’t advanced enough to handle that level of precision. Remember, it wasn’t all that long ago that roboticists finally got a cutting-edge bot to catch a ball — simple enough for humans, far less so for robots.
But agtech companies in the private sector and robotics departments in academia continue their efforts to clear that hurdle.
Location: Groton, Mass.
How it’s using farming and agricultural robots: All those charming decorative grasses, flowers and shrubbery accents at your local gardening center are big business. The commercial greenhouse market is poised to become a $38 billion industry within the next four years, and growers are increasingly using robots to help fill the high labor demand.
Founded by former employees of Roomba inventors iRobot, Harvest Automation made its first product with this fast-growing market in mind. The behavior-based HV-100 robot handles the important, but highly repetitive and strenuous, work of spacing container crops and plants. (Greenhouse plants need space between them so they grow thick and bushy and resilient, but too much space means square footage isn't being optimized.) The HV-100 is built to keep running even in the scorching temperatures and less-than-pristine environments of nurseries that grow ornamental plants and specialty fruits and veggies.
Location: Plant City, Fla.
How it’s using farming and agricultural robots: Gary Wishnatzki is vocal about the labor pinch he says growers have faced over the last few years. The owner of berry supplier Wish Farms told the New Yorker in April that he relies on workers hired through expensive temporary visa programs. Among reasons for America's labor shortage, the article identifies suppressed immigration and a diminished appetite among low-skilled domestic workers to do the backbreaking labor of strawberry harvesting. Wishnatzki's push to automate isn’t about eliminating farm jobs, he argues, but meeting the demands of consumers who’ve come to expect fresh strawberries even in the dead of winter.
He’s making that push as co-founder of Harvest CROO, a startup that developed an advanced strawberry-harvesting robot called Berry 5. It uses a variety of robotic components — rather than a single arm — to grab the leaf, pick the berry and pack it. Computer vision helps Berry 5 decipher ripe berries from non-ripe ones before plucking. And it's fast compared to human laborers, purportedly able to pick a plant in eight seconds and shift to the next in one-and-a-half.
Developed thanks to millions in investment dollars from others in the berry industry, Berry 5 is currently working Florida fields on a trial run. The company reportedly hopes to commercialize it before the end of 2019.
Location: Cambridge, U.K.
How it’s using farming and agricultural robots: Lettuce-harvesting has remained stubbornly robot-resistant thanks to the plant’s fragile nature and close proximity to the ground. But researchers at Cambridge University made a breakthrough with their so-called “Vegebot,” another computer vision-powered prototype.
Here’s how it works: One camera scans the lettuce and gives a thumbs up or down for harvesting. A second camera (positioned near a blade) then guides the pick without crushing the plant. Meanwhile, a machine-learning algorithm “teaches” the robot to avoid unripe or diseased lettuce.
Vegebot hardly works with the speed or skill of human hands, but a series of test runs have purportedly established proof-of-concept success, which also augurs well for other above-ground fruits, vegetables and grains.
“Every field is different, every lettuce is different. But if we can make a robotic harvester work with iceberg lettuce, we could also make it work with many other crops,” Simon Birrell, of Cambridge’s Department of Engineering, said in a press release.
While a scalable commercial lettuce harvester remains elusive, efforts to produce one have ramped up. In 2017, for example, John Deere acquired CV agtech trailblazer Blue River Technology, which has made notable strides in lettuce-focused robotics.
If you've ever tended a personal garden, you're well aware that weed control is both important and difficult. Commercial agriculturists know it, too, but on a massive scale. Even when crop rotation is possible, many large outfits rely at least somewhat on the use of herbicides. But given the facts that plants can become resistant to weed killers and consumers are increasingly averse to chemically treated food, it's hardly a perfect solution. That's why weed-management robots — including ones that incorporate advanced AI to help distinguish between crops and weeds — are an attractive option.
Location: Ramonville-Saint-Agne, France
How it’s using farming and agricultural robots: One of the Old World’s preeminent wine producers, Château Mouton-Rothschild is very New World when it comes to vineyard maintenance. The famed estate is among several vintners that have partnered with Naio Technologies to enlist Ted, the company’s vine-tailored robotic weed killer.
Electric and shaped like an inverted U, the long-running Ted simultaneously rolls over and around a vine row, using RTK satellite navigation to stay on course. (A drone maps out the initial plot of land that Ted surveys.) Industry-standard blades and finger weeders run along the base, pulling unwanted weeds from the vines and consequently decreasing the need for herbicides. Ted's robotic cousins include weeding robot Oz and vegetable robot Dino.
Location: Dartmouth, N.S.
How it’s using agricultural robots: This scrappy Nova Scotia startup played David to some of North America’s most prestigious robotics-program Goliaths at last year’s Weed and Feed ag-bot competition. The company’s top prize-winning entry, a weed-yanking autonomous robot dubbed R2Weed2 (yep, you read that correctly), employs artificial intelligence to differentiate between weeds and crops so the former are stripped and the latter left to grow. As it runs, R2 also gathers data that help farmers with soil analysis and environmental monitoring. A commercial version will reportedly be available later this year.
Robotic Greenhouses/Robot Farming
Instead of bringing robots to the field, one of the next great advances in farming automation will bring the field to robots.
Below are two well-financed coastal startups —one West, one East — that are helping, um, plant the seed for a robotic-greenhouse future. Not everything is growable in this way, but for certain crops the improvements are striking. Both of these companies promise a dramatic decrease in the amount of water used (between 90 and 95 percent less) for an equivalent crop yield, and both boast controlled indoor environments that eliminate the need for pesticides. Here’s how they do it.
Location: San Carlos, Calif.
How it’s using agricultural robots: Inside what the company describes as “the world’s first autonomous farm,” an 8,000-square-foot space that more closely resembles a research lab than a farm field, two cloud-connected robots oversee the growth of leafy greens (romaine and butterhead lettuce, bok choy, kale, arugula) and herbs (basil, cilantro, chives, sorrel, parsley), all grown inside heavy hydroponic pods. Using computer vision and sensors as its “eyes,” one robot does the heavy lifting, transporting the pods across the facility; the second analyzes and picks the individual plants.
It all happens beneath high-efficiency LED lights and under the watchful eyes of some dozen on-site robot and plant scientists. The Y Combinator-backed company — founded by two Willow Garage alumni — claims the small space can yield as much as a one-acre traditional farm.
Iron Ox began production last October in its greenhouse-style facility and started selling at its first partner store, a market in its hometown San Carlos, a few months later. The long-term goal is to establish additional farms near other high-demand areas, which would cut out the long-slog transportation costs that exist when growing is isolated to just a few regions in the country.
How it’s using agricultural robots: Unlike the large, low-profile vats in which crops are nurtured at Iron Ox, Bowery Farming stacks up layers of trays, each filled with greens, in the traditional vertical-farming format inside its Kearney, N.J., growing space.
But it's also using robotics, artificial intelligence and LED to grow leafy greens and herbs with the same end-goal in mind: addressing the problems posed by labor scarcity, population booms and centralized farming.
Here, a proprietary operating system and complex array of sensors collect data and maintain an ultra-precise balance of water, temperature, nutrients and humidity. At the same time, a team of vertical farmers help to harvest and watch over the crops.
The company — which has received investment capital from the likes of celebrity chef Tom Colicchio and Uber CEO Dara Khosrowshahi — sells its greens and herbs at Whole Foods, Foragers and Westside Market. In the New York metro area, they're also available through Peapod, Jet.com, Amazon and in two of Colicchio's high-end restaurants.
Aerial Imagery Drones and Seed-Planting Drones
Aerial imagery can save farmers a lot of time by giving them a bird’s eye view of crops; that way, they can quickly get a sense of vegetation’s health, insect issues, irrigation layouts and weed growth. It even allows them to precisely determine how much pesticide the crops require.
Farmers can use a variety of subscription services to access these valuable flyover images (including thermal, infrared and NDVI) of their fields, but fewer companies have taken the full plunge into unmanned aerial vehicles (UAVs). That’s most likely because of FAA restrictions on autonomous drones, which require that pilots be immediately ready to take control of a drone. Small unmanned aircrafts must also be kept within a would-be pilot’s line of sight when airborne.
But they're out there, if not yet in large numbers. Here are some companies that prove there's something’s in the air when it comes to agricultural imaging, seed planting and cloud seeding.
Location: Marlborough, Mass.
How it’s using agricultural robots: One of several Boston-area companies making notable agtech breakthroughs, American Robotics is the team behind Scout, an aerial imaging drone that fits the so-called “drone in a box” model.
Between flights, Scout lives inside a weatherproofed box, where it self-charges and processes (via edge computing) all the data it collects. When it takes flight to examine fields, the box top opens and the fully autonomous drone lifts off, using artificial intelligence to plot and conduct the run. During missions, which can be scheduled or launched on demand, Scout gathers crop stress data that farmers can use throughout a crop’s life cycle. It’s the kind of aerial surveillance that remains a key facet in the growing field of so-called precision agriculture.
Location: Las Vegas
How it’s using agricultural robots: Any UAV pilot who wants to fly a drone that weighs more than 55 pounds at takeoff (including cargo) needs to petition for a special exemption through the F.A.A. That means non-exempt farmers who want to drone-scatter seeds over their acreage are limited in how many pounds of seed they can spread in a single run. Still, a number of manufacturers have developed drones marketed specifically for that purpose.
UAV Systems International sells two drones that spread seed and fertilizer, one with a payload capacity of about four pounds and another with a roughly 11-pound payload. Both have a two-mile flight range and a 20-minute limit, according to the company. UAV also markets crop sprayer drones and surveillance drones that inspect crop health.
Location: Tel Aviv
How it’s using agricultural robots: Given the FAA’s restrictions on autonomous drones, perhaps it’s no surprise that some of the leading purveyors of agriculture-focused UAVs are headquartered outside the U.S. and cater to an international clientele. That includes Israel-based Taranis, which has brought its high-resolution scans to farmers in Europe, South America and North America. Along with more traditional aircraft, the company operates drones that use computer vision and data science to monitor crop stress and self-improve identification capabilities.
Location: Reno, Nev.
How it’s using agricultural robots: Cloud seeding is all about making it rain — literally. The concept dates back to at least 1946, when Dr. Bernard Vonnegut discovered that, under certain conditions, introducing silver iodide particles to clouds could spur the creation of ice crystals, potentially precipitating precipitation. (Fun fact: Bernard was brother to literary legend Kurt Vonnegut.) But decades later, the process still has critics who note that variables abound and that it's tough to measure how much rain or snowfall is actually the product of a cloud seed. (The weather-modification practice also has military applications and, it should be noted, is internationally-banned for use in war ). But as droughts worsen, government agencies keep trying. In July, Indonesia’s disaster mitigation board used cloud seeding in an attempt to save drought-stricken crops.
Today, clouds are seeded using airplanes, ground generators or gun- or rocket-propelled canisters, but scientists are also exploring whether drones can handle the task. In 2017, the Desert Research Institute — part of Nevada’s higher education system — in conjunction with Drone America notably launched an unmanned cloud-seeding drone during an hourlong autonomous fight and beyond researchers’ naked line of sight.
"Reaching this milestone," said Mike Richards, president and CEO of Drone America, "allows us to now focus on higher altitude, longer distance flights; as well as the extreme challenges of flying our advanced unmanned fixed-wing aircraft in the harsh, icy weather that come with optimal winter cloud-seeding conditions."
Images via Shutterstock, social media and company websites.