The Atlanta Journal-Constitution
High tech is looking to transform farming
Labor shortage pushes the need for driverless agriculture equipment.
Of all the out-of-the-box products a Silicon Valley tech startup could offer, Bear Flag Robotics may be delivering the most unexpected: plowed fields.
The company is developing autonomous tractors, a goal equipment companies such as Case IH, John Deere and Kubota are chasing as well. But the business model of Bear Flag, based in Sunnyvale, California, has a twist — it does not build the tractors. Instead, it adapts the sensors and actuators needed for driverless plowing to existing tractors produced by major manufacturers.
That step is not as sci-fi as it might seem. From equipment automation to data collection and analysis, the digital evolution of agriculture is already a fact of life on farms across the United States.
Auto-steer systems, which use GPS receivers to keep rows straight and avoid gaps or overlap, are available for equipment ranging from tractors to harvest combines to sprayers with 100-footwide booms. Precision seeders and fertilizer systems can be satellite-guided to accuracy of an inch or less.
The difference: For the most part, those operations still depend on an operator at the controls.
“Autonomous operation will be
a service in agriculture before it’s a product,” said Igino Cafiero, Bear Flag’s chief executive, during a break from his work in a test field of cilantro about 60 miles southeast of the company’s headquarters. The company’s niche is providing secondary tillage, deploying its equipment after a harvest is complete to prepare the fields for the next planting.
The need for driverless farming equipment is intensifying, Cafiero said, because of a crushing labor shortage, which drives up wages and worker mobility. Tractors equipped with Bear Flag technology are able to work fields around the clock, with- out a driver, using sensors similar to those in autono- mous road vehicles under development: lidar, radar and digital video. The sensory devices provide more than what Cafiero calls situational awareness, vital for safe operation where workers and livestock may be nearby, also collecting data on the land to improve efficiency. While Bear Flag pursues expanding capabilities to tasks like planting and spraying that have long demanded human supervision, it also plans to expand to the labor-intensive harvest duties of crops, including tree nuts and row crops. The drive to increase productivity is urgent in all phases of agriculture. Feeding a world population expected to reach 9.7 billion by 2050 faces dire challenges, according to the summary of a U.N. report released in August. The effects of climate change — extreme weather, soil loss, migration pressures — will strain land and water resources, potentially disrupting food supplies.
Yet growing crops has historically been an uncertain enterprise, a livelihood that increasingly depends on forecasts of weather conditions, commodity prices and complex factors like maturity index and projected yield. Agriculture is seen as an industry ideally suited to large-scale data collection and analysis, and technol- ogy companies more closely associated with databases and computer hardware are seeing opportunities.
IBM, for example, made its move into the cloud — not the virtual data repository, but the puffy ones in the sky — in 2016 with the purchase of The Weather Co., bringing super computer prowess to what once depended on the centuries of record-keeping by trusted prediction tools like The Old Farmer’s Almanac.
Jeff Keiser, a manager for agribusiness solutions at The Weather Co., knows more than agriculture analytics. He also farms corn and soybeans in Indiana, where he has encountered many of the same conditions as the wide range of food producers who can make use of IBM’s Watson Decision Platform for Agriculture. “With the cold spring followed by high temperatures and a lot of rain, it’s been a very challenging year,” Keiser said. “I got some planting done in April, but it wasn’t finished until June.”
The decision platform, which will mark its first anniversary in October, is designed to ingest data from satellite imagery and from sensors on farm equipment that monitor, among other things, seed counts, nutri- ent levels and fertilizer flow, said Cameron Clayton, chief executive of The Weather Co. A sensor-equipped farm of 1,000 acres requires vast analysis and storage capabilities on the scale of what IBM can provide.
With more than 2 million acres of farmland around the world covered, the platform provides hyperlocal six- month weather predictions based on satellite and atmo- spheric data.
The system makes extensive use of IBM’s experience in artificial intelligence to build management models for corn, soybeans, wheat, barley and other crops. Each model takes IBM six months to a year to assemble and accounts for issues that include pest control and fertilizer requirements. It’s not a one-size-fits-all solution: The models are tailored to the specific crops, geared to produce longer potatoes for french fries or barley for malting in beer production.
The benefits of automation scale down to some smaller growers as well. Penny Gritt Goff, the third-generation operating manager of Gritt’s Midway Greenhouse in Red House, West Virginia, takes advantage of computerized monitoring to keep tabs on temperature, humidity, nutrient levels and other conditions for 3 acres of hydroponic greenhouses where lettuce grows in flowing water and tomatoes are raised in a bed of coconut husks.
The computer system can send alarms when it gets too hot inside the greenhouses, but it also takes action on its own, spreading a shade cloth covering to cut down on sun exposure (or retain heat in the winter), opening and closing vents, and regulating irrigation. “The automated controls narrow the chances of failure,” Goff said, and lessen the need for some aspects of the operation’s human monitoring. “We could add more advanced equipment, but at this size it’s not economically feasible.”