In a greenhouse in Belgium, a small robot moves through rows of strawberries growing on trays suspended above the ground, using machine vision to locate ripe, flawless berries, then reaching up with a 3D-printed hand to gently pluck each berry and place it in a basket for sale. If it feels that a berry isn’t ready for harvest, the robot estimates the date it will be ready for it to return and pick it.
It’s a test of a system that Octinion, the R&D company building the robot, believes could replace traditional strawberry farming and harvesting:
In California, where tough immigration policies plus broader economic conditions mean the number of immigrant farmworkers is decreasing (and native-born workers don’t want the job), strawberry growers are finding it harder to find workers to harvest fruit. In the U.K., Brexit is making farm labor less appealing for Eastern European workers who used to do the job. Most developed countries are facing similar challenges with agricultural labor shortages.
“Agricultural labor, at this point, is not sustainable, in the sense that it’s often people who come a long way–a few thousand kilometers–do that work, and after the season they go back, or people come over as immigrants and do that kind of job to get started, and afterwards move on to other, better jobs,” says Tom Coen, CEO of Octinion.
The new robot can pick one berry every five seconds; a human can do the job slightly faster, picking and packing a berry every three seconds. “We’re a bit slower, but we’re already economically profitable because the cost per berry is similar,” Coen says.
The company began designing the robot based on cost constraints, and other specific requirements for strawberries. The stem, for example, shouldn’t stay on the berry when it’s picked, because it could poke other sensitive fruit in the basket. When the berries go in the packaging, more red should show than green berry tops, to attract customers. The robot’s vision system can accomplish this.
The robot is designed to work with “tabletop” growing systems, where strawberries are grown in trays, rather than in fields, because the industry is moving in this direction. In Europe, greenhouse-grown strawberries are already standard. In California, which produces most of the strawberries eaten in the United States, major producers such as Driscolls are also beginning to move to tabletop growing systems, because the height is better both for robots and for people who would otherwise hunch over low plants growing in a field. (Driscolls has been testing another strawberry-picking robot, called Agrobot, which can harvest multiple berries at once but is rougher on the fruit; Octinion’s device automatically calculates whether it could possibly bruise a berry, and if so, doesn’t pick it).
In addition to making berries easier to pick, growing them in a tabletop system saves water because only a small amount of soil around the plant needs to get wet, and it makes it possible to grow more fruit in the same amount of space. Vertical farming systems–some of which are figuring out how to grow strawberries–can grow even more food in a small footprint, and the robot could be adapted to move up and down high walls filled with fruit.
As the world continues to become more urbanized, Coen believes that more vertical farming is inevitable, and the robot could help make it economical. “Something like 80% of the strawberry production in the U.S. comes from California, which means if you eat a strawberry in New York it has been on a truck for something like two days. So both ecologically and economically, that’s not really okay.”
The robot is currently nearing the end of development as the company tests it in collaboration with Belgium’s Research Center of Hoogstraten. Octinion will begin pilot testing with strawberry farmers in 2018, and the company plans to begin selling it in 2019. It also plans to begin adapting the robot to pick other produce like peppers, tomatoes, and cucumbers. “We’re focusing on everything that can be grown in a greenhouse,” Coen says.