I don’t think sci-fi saw this coming. For so long, futuristic books and films have promised us robots like C-3PO that translate alien languages and assist us in hijinks. Or ones like Rosie that clean our houses. Or, on the other end of the spectrum, robots that level our houses and destroy humanity. Looking at you, Arnold.
The reality of modern robotics couldn’t be more different. These days, it’s more about developing robots that … shake the bejeezus out of cherry trees.
Researchers at Washington State University have developed algorithms that scan a tree for individual branches, then determine what bit of each branch to grasp and shake to extract the most cherries—up to nearly 90 percent of them. Sure, that’s not as dramatic as the machine-driven apocalypse. But at least it gives us an intriguing vision of a robotics-fueled agriculture industry.
If the idea of a tree-shaking robot seems a bit oddball to you, know that you’re the reason it may soon exist. Americans don’t just want a lot of produce; they want a lot of flawless produce, free of dings and bruises and discoloration. And you can’t have flawless cherries if you’re using a giant machine to shake a tree trunk willy-nilly, flinging fruit all over the place. (Though that method works fine with hardier fare like almonds.)
So this algorithm is a little more genteel in its approach—starting by pinpointing the branches. The researchers were working with a special variety of cherry tree that grows “upright fruiting offshoots” (yes, known in the industry as UFOs). A more traditional bushy tree may hide fruit in its core, but this kind has a trunk that comes out of the ground and bends almost horizontally, out of which branches grow upright. It looks a bit like a menorah, really.
This kind of cherry tree may be optimally arranged for picking by human hand, but it’s a challenge for an algorithm. “In a lot of different situations, the branches were hidden by berries or leaves,” says Manoj Karkee, an agricultural roboticist at Washington State University. “We used some mathematical modeling to estimate their location based on the cluster of fruit, as well as some visible sections of the branches.”
So, the algorithm’s got sight of the obscured branch. Next is figuring out where to grasp and shake. Rule number one, of course: Avoid any clusters of fruit. “What we’ve found is that if we start shaking a tree at about one third of the height, we’d be able to remove about 70 to 80 percent of cherries,” says Karkee. “Then for the remaining cherries, we go to about two thirds of the height and shake it again.” Do this with all six or seven upright branches and a robot might be able to extract almost 90 percent of the fruit. (So a small crew would still need to trail the robot to pluck the remaining cherries, at least until the researchers improved the machine’s efficiency.)
Might. The robot to do the actual work doesn’t exist yet. But what Karkee envisions is a machine with six, maybe eight arms that rolls through the orchard grabbing branches and giving them a good shake. Specifically, each shake buzzing at 18 hertz for five seconds, is what he and his colleagues found works best.
Actually developing a robot that works outdoors, though, is no small task. “You’re potentially dealing with moisture, with driving on rugged ground,” says Jon Binney, co-founder & CTO of Iron Ox, which has developed an automated indoor farming system. “All solvable mechanical problems, but non-trivial.” Problems that are important to solve soon, as agricultural employment continues to dwindle in the United States. Between 2002 and 2014, American farms lost nearly 150,000 laborers, or 20 percent of the workforce. People just aren’t flocking to the jobs.
So the future of agriculture will be increasingly robotic. One startup, for instance, has developed a robot that spots apples and picks them with a suction tube. And a machine called the LettuceBot rolls through fields eyeballing weeds and automatically spraying them. “What you’re going to see is people trying a lot of very specific things outdoors, most of which aren’t going to end up being cost effective or reliable enough, but some of which work,” says Binney.
Which is not to say farmers can’t engineer outdoor farming to introduce some of the order you get with indoor farming. For instance, farmers in California plant a special kind of lettuce that grows like a bulb, which makes it easier for a fancy machine to cut the base with a knife made of water. It’s easy to imagine, then, that humans will create all manner of new varieties of fruits and vegetables—maybe even a differently-branching cherry tree—to better get along with robot harvesters.
No matter how, if humanity wants to feed its ballooning population, it must lean heavily on the machines. So stay out of our way, Arnold.