UC Davis developing faster, more accurate robotic cultivator

Led by Professor David Slaughter of the UC Davis Department of Biological and Agricultural Engineering, the project recently received a $2.7 million grant from the U.S. Department of Agriculture Specialty Crop Research Initiative. It addresses a problem that has vexed precision weed management for years: How do you devise a robotic cultivator that can quickly distinguish friend from foe?

“Machines can recognize a weed, and they can recognize a crop plant, but they have trouble distinguishing one pattern from another when they are co-mingled, as is often the case with weeds and young crops in the field, particularly when traveling at a typical tractor speed of three-feet-per-second or more,” Slaughter said.

Slaughter's team is designing a robotic cultivator that can remove weeds in commercial fields as carefully as gardeners pull weeds in their own backyard, without the time-consuming labor and cost. They're developing a “smart” cultivator with small knives that reach out to uproot weeds and retract to keep crops intact.

It will weed the beds of any row crop and will be especially useful in wide beds of densely seeded crops like spinach and baby lettuce, which can turn green almost overnight with weeds and leafy crops.

“Current vision-sensing mechanical cultivators can sometimes recognize weeds along the edges of wide beds, or seed lines, as we call them, but they get lost in the middle,” said Steve Fennimore, Cooperative Extension weed specialist with the UC Davis Department of Plant Sciences, a member of the new robotic cultivator team. “Workers often have to go back through and hand-weed them.”

How will the new cultivator distinguish friend from foe? Thanks to a safe, simple seed coating, the plants will signal the cultivator by emitting a faint, fluorescent glow that will appear when seedlings emerge and are most vulnerable, then vanish as plants grow and can out-compete weeds for sun, water and nutrients.

“It won't involve biotechnology or any genetic engineering,” Slaughter said. “The seeds will be coated with a safe, inert, fluorescent material.”

Seeds are commonly coated with various materials (like fertilizers, fungicides or herbicides) for different reasons. Some seeds are coated to protect them from birds, rodents and stress. Some seeds are coated with colorful, inert material to make them easier to plant and easier to pick up if they spill. Some seeds are coated as a way for companies to protect their brand identity — a process they call “track and trace.”

To develop the seed coating, Slaughter's team will work with researchers from the Seed Biotechnology Center at UC Davis and Aginnovation, a company that specializes in seed technology, located in Walnut Grove in Sacramento County. Aginnovation is a founding member of the Centor Group, a group of independent companies from around the world that work together to deliver leading-edge technology and services to the agricultural seed industry.

This new smart-cultivator technology could be the breakthrough needed to help crop plants communicate with existing machines like automated lettuce thinners, machines that drive through heavily seeded fields and remove all but the most viable plants.

The new cultivator should move more quickly through a field than current vision-sensor models because it won't take as long for the machine to distinguish the good guys from the bad guys. That's good news for vegetable growers like Alain Pincot, managing partner of Bonipak Farms in Santa Maria.

“As the cost of labor rises in California, mechanical cultivators become more important to both organic and conventional ag production,” Pincot said. “We've been fairly happy with our existing automatic weeders, but we would be interested in a new type of cultivator if it moved more quickly and could accommodate beds of various widths. A robotic weeder with a higher speed and good accuracy along the row would be a winning machine.”

That's no doubt why the USDA is investing $2.7 million in the new cultivator design, which will take shape over the next five years.

Several researchers from various disciplines in different states are teaming up on the project. In addition to Slaughter and Fennimore, they include Professors Ken Giles, Shrinivasa Upadhyaya and Stavros Vougioukas, all with the UC Davis Department of Biological and Agricultural Engineering; Richard Smith, Cooperative Extension farm advisor in Salinas County, who specializes in vegetable crop production and weed science; Laura Tourte, Cooperative Extension farm advisor in Santa Cruz, Monterey and San Benito counties, specializing in farm management; Mark Siemens, professor and Cooperative Extension specialist with the University of Arizona Department of Agricultural and Biosystems Engineering; Professor Manoj Karkee with the Washington State University Biological Systems Engineering Department; and Professor Qin Zhang, director of the Center for Precision and Automated Agricultural Systems at Washington State University.