Excel's Coffee Shop

[basserdan]

These giant robots can pick strawberries. What does that mean for humans?

By Laura Reiley
Friday, October 27, 2017 5:01am


Ken Gardner, director of autonomous development at Harvest CROO Robotics, checks on the company’s strawberry-picking robot at G?&?D Farms in Duette, Fl.

DUETTE — It stands in the dirt-floored barn. Its 16 arms, with fierce-looking spikes at the ends, emit pneumatic hisses as they punch downward. It is 30 feet wide, 12 feet long, 14 feet tall. It's like Optimus Prime, if the Transformer became a majestic steed with tractor-tire hooves.

The robots have arrived. And they'll be picking crops in Florida fields soon.

Robots can do things humans can't. They can pick all through the night. They can measure weight better. They can pack boxes more efficiently. They don't take sick days, they don't have visa problems.

Google "are robots taking our jobs?" and you get millions of theories: Robots will take over most jobs within 30 years; yes, but it's a good thing; yes, but they will create jobs, too; chill out, they won't take them all. Truckers, surgeons, accountants and journalists have all been theoretically replaced by prognosticators.

But harvesting specialty crops is different: Plants vary in shape and size and determining ripeness is complex — experts have said there are too many variables for robots. Until now.


Bob Pitzer, Chief Technical Officer and Chief Engineer at Harvest CROO Robotics, stands with the company's Strawberry picking robot at G & D Farms in Duette.

• • •

Gary Wishnatzki is the third-generation owner of Plant City's Wish Farms. He has 600 acres of strawberries, and he's the leading blueberry supplier in the state. You've also seen the Wish Farms label in the grocery stores on blackberries and raspberries. Their name is on the soundstage at the Florida Strawberry Festival and Wishnatzki has testified before the House Committee on Agriculture about the future of agriculture.

He and robotics designer Bob Pitzer started prototyping agricultural robots in 2013. They will have their alpha in the fields this year. Next year they hope to have a tweaked version, with full commercialization of the system in 2020.

They have raised $5 million to date from investors. Really, they got the money from their competition. More than 60 percent of the U.S. strawberry industry in both Florida and California are investors in this project.

"I went and personally visited all these competitors of my company," Wishnatzki says. "We've got future customers lined up. Now we just have to deliver the technology. It's like self-driving cars: We don't have to be perfect, we just have to be better than the humans."

The push toward field robotics, they say, is about a shrinking labor pool.

"There have been several years where we haven't been able to get our whole crop harvested, we've had to leave berries in the field," Wishnatzki says. "That's a pretty painful thing when you raise a crop. It's clear to anyone in the industry that we're on a very bad course right now and nearing crisis stage when you increase cost of labor by 50 percent."

The east Tampa office of Harvest CROO (that's Computerized Robotic Optimized Obtainer) looks like the very large man cave for a fraternity of BattleBots fans. And it kind of is. There are 10 employees and interns, all guys, mostly machinists and engineers. Two dogs wander around, mostly well-behaved. Robot action figures crowd desks.

Before we head into the lab to see the real robots, Pitzer pauses to consider what we can photograph, deciding it's all fine.

"A lot of our secret sauce is still buried," he says as the door swings open. "It's the software."

There's a part of a robot with a rotating wheel of soft plastic grabbers, grabbers that will hopefully pick a whole strawberry plant of ripe berries in 8 seconds. Grabbers that, if this all works out, could be retooled to pick tomatoes and other Florida specialty crops.

Pitzer served in the first Gulf War, running reactors on Navy nuclear submarines. After graduating from the University of Florida, he worked on robotics for Intel. He met Wishnatzki while consulting for a company building DVD disk machines. He got a call from the strawberry giant in 2007 about an automation project in Wishnatzki's warehouse.

"The thing that is driving all of it is labor," Pitzer says, insisting that this isn't about the Trump administration, tighter immigration policies or anything political. It has to do with the birthrate in Mexico, where the vast majority of the farm laborers hail from.

In the 1950s and '60s, the birthrate in Mexico was 6.7 kids and now it's 2.2, about what it is in the United States, according to data from the World Bank. It doesn't matter what laws you pass or walls you build, he says, there just aren't enough pickers for specialty crops. Wishnatzki and others predict that Mexico, a huge agricultural exporter, will soon not have enough labor to pick its own crops — it may become a net importer of labor.

Mexican labor is in short supply, it's too expensive to bring in seasonal laborers from Southeast Asia or South America, and domestic American workers don't want to work the fields. Why not entice workers with better pay? Fat chance. Consumers balk at $10 tomatoes, and American agriculture would lose ground to cheaper imports.

Traditional mechanized harvesters are destructive, meaning they swoop through a field and rip and cut, wrecking the plants. That won't work for strawberries — berries ripen at different times, so each plant may be visited 40 times over the season by pickers. That's a tall order for robots, but technology has evolved so rapidly that Wishnatzki's dream is now possible, says Kenneth Parker, executive director of the Florida Strawberry Growers Association.

"Think of all the industries where robotics has improved productivity and reduced costs," Parker says. "I think agriculture in specialty crops is late to the game — it's about the speed of the processors, the ability to process the data to say that fruit is the right size and right color and does not have any decay. A robot has to make a lot of decisions instantaneously."

He lauds Wishnatzki's foresight and says that while some growers wonder if his implementation time frame is realistic, everyone recognizes the need.

A robotic harvester will pick a plant in 8 seconds, with another 1.5 seconds to move on to the next plant. Pitzer estimates that each harvester will be able to pick 8 acres in a day, the equivalent of 30 human pickers.

"Our business model is to build machines as a service," Wishnatzki says. He plans to keep the machines and rent them to farms so farmers wouldn't have to buy their own robots.

Using processors developed for driverless cars and cameras from cellphone technology, each robot right now costs $20,000 to $25,000 to build. Pitzer is hoping to get the cost down to $7,000.

Now, the question is, when can other farms start to use it? And how long until they break even?


The strawberry picker component of Harvest CROO's automated strawberry picking and planting machine at the Harvest CROO robotics lab in Tampa. The picker sits on a rail system and can delicately pick berries without crushing the fruit.

• • •

The robot's spiky arms punch the holes in which each plant nestles, GPS employed to locate each plant down to the sub-half inch. The system better forecasts ripening and production, and robots can report back pests in a specific GPS location in the field to catch a problem early.

Robots will be able to pick 20 hours a day, through the night and the cooler parts of the day so the fruit is less likely to bruise, and because much less energy is required to cool already-cool fruit, growers save refrigeration costs.

Lucky Westwood is the operations manager at California Giant, which has 3,000 to 4,000 acres of strawberries mostly in Watsonville, Salinas and Santa Maria. Each year they employ close to 4,000 human harvesters. The labor shortage is something people in his industry have seen coming for a while: The Mexican economy is better, there has been a decade of increasing border security and there's more upward mobility of Mexican emigres.

"This is pure physical labor," he says. "When people have an opportunity to move up and take another job, they do."

California Giant was an early investor in Harvest CROO, but Westwood doesn't think robotic harvesting will necessarily replace humans.

"How this machine will actually finish the job of picking and packing is unknown. We've experienced other automation where it couldn't quite do what humans do. The question is whether you'll need people to work with and around the machines."

Machine harvesting is not new, says Barry Estabrook, author of Tomatoland, an investigation of the tomato industry in South Florida. Since the mid 1960s, canning tomatoes have been machine harvested.

"When the mechanical harvesting came to the California tomato industry there was a lot of controversy about displacing thousands of workers," he says.

Attorneys of California Rural Legal Assistance sued the University of California, the institution that had developed the technology, on behalf of 19 farm workers, alleging that publicly-funded mechanization research displaced farm workers and eliminated small farmers.

Even if strawberry robots hit it out of the park, Reggie Brown, executive vice president of the Florida Tomato Exchange in Maitland estimates a full decade before specialty crops like tomatoes will be harvested robotically. And he's hoping we can last that long.

"If we don't solve our immigration problem, it will be a moot point because we won't be able to get through the next decade. We have to solve our guest worker problem. Our labor pool is small and getting smaller every year."

So how many human workers will be displaced? Brown doesn't think many. Because the transition to robotic harvesting is going to be capital intensive, he thinks it will be a slow migration, with humans and robots working together.

http://www.tampabay.com/things-to-do/consumer/these-giant-robots-can-pick-strawberries-what-does-that-mean-for-humans/2342492