Field of "Bots"

The agBOT Challenge searches for technologies that test the status quo.

Wiring the new farm: The Colorado School of Mines team drove all night to compete at an event in West Lafayette, Indiana. Typical agBOT Challenge teams included electrical engineering, programming and computer science majors, Image by Dan Miller

“Farmers don’t want to be morticians,” says Steve Gerrish, explaining that growers would rather be proactive in addressing field problems. “[Can] we fix something before it happens?” he asks. “Can we ID and kill weeds in one action; can we dig down into the soil [on the go] and find disease or nematodes?”

These are questions that inform the direction of the agBOT Challenge, an event begun by Gerrish in 2015 on his Gerrish Farms, outside Rockville, Indiana. “I don’t want someone coming back and telling me that I need to do something. Tell me what you’ve done, that you sprayed with a fungicide, that you found nematodes and used a nematicide, that you killed a specific weed,” Gerrish says adding, “I’ll be really happy with that result.”

The Challenge is a demonstrator of a future with rural high-speed broadband service, how robust broadband would allow robots or “bots” to analyze crop environments and initiate actions.

Without high-speed connections, it is difficult at best for farmers to flow into analytics software the billions of data points they are collecting, says Gerrish’s daughter, Rachel Gerrish. And, that thwarts their ability to identify and act on management issues in real or near-real time. “[In 2015], we wanted to showcase the need for [high-speed] broadband to enhance agriculture,” she says. Rachel is the agBOT Challenge’s senior executive producer.

AG DISRUPTER

“Farmers want to go in this direction,” Steve Gerrish says. “We’re trying in a positive way to disrupt agriculture, to change how we think about [farming].”

P[] D[x] M[x] OOP[F] ADUNIT[] T[]

The 2019 agBOT Challenge moved from the Gerrish farm to Purdue University’s 1,400-acre Agronomy Center for Research and Education (ACRE), in West Lafayette, Indiana. Teams of electrical engineers, computer sciences majors and programmers worried over their automated “bots”--checking connections, cameras and sensors--as they prepared to kill weeds, fertilize corn and collect soil samples for prizes totaling a reported $100,000.

There were two tasks at this year’s Challenge.

A “Weed and Feed” task had teams build automated vehicles capable of treating ragweed, foxtail and cocklebur with herbicides, and corn with nitrogen. Plants were identified while cameras and sensors produced imagery matched against onboard, digitized images of the plants. All this was done as the bots moved down rows of simulated no-till ground laced with small weeds and small sprigs of corn.

A second event, “Mining for Microbes and Microfauna,” had teams autonomously gather soil samples and prepare them for diagnostic evaluation while still in the field. The process involved drilling into the soil and several mechanical steps that winnowed the soil collected down to ready-for-the-lab sized samples.

“[The agBOT Challenge] offers a glimpse into the future of how unmanned vehicles and other technologies will impact agriculture and society at large,” says Ronald Turco, head of Purdue’s Department of Agronomy. “These tools are almost certainly going to be critical to help farmers produce the food that will be required to feed a growing population, and it shows what is possible with high-speed broadband.”

NEW PARTNERS

Gerrish says the Challenge encourages “uncommon collaborations,” where the work of ag research intersects with other non-ag disciplines such as imagery, aerial sensing, electrical engineering and computer science.

“One of our objectives has always been to show how cross-disciplinary research drives innovation,” explains Rachel Gerrish. As an example of uncommon collaborations, the Purdue University team included ag engineering, biological engineering, electrical engineering and computer science students.

Over the four years of the challenge, 50 teams from the U.S., Canada and other nations have competed. The agBOT Challenge has tasked teams to produce systems that autonomously reproduce planting, weeding, fertilizing and harvesting functions. Here are the winning teams in the 2019 agBOT Challenge:

> Weed & Feed (identify specific weeds/fertilize corn). The Purdue University team finished first. IUPUI (Indiana University Purdue University Indianapolis) finished second. University of Manitoba finished third. Other entrants included Colorado School of Mines and ROS Agriculture from Texas.

> Mining for Microbes and Microfauna (automated soil testing). Virginia Tech finished first.

[PF_0719]

Past Issues

and