New study from Illinois examines the use of robotic weed control to combat superweeds

Robots work by dragging hoes through the soil, disrupting the emergence of weed seeds

PUBLISHED ON October 31, 2024

URBANA, Ill. – Most corn and soybean fields in the United States are planted with herbicide-resistant plant varieties. However, the emergence of superweeds that have developed resistance to common herbicides threatens current weed control strategies. Agricultural robotics for mechanical weed control is an emerging technology that could potentially provide a solution. A new study from the University of Illinois Urbana-Champaign examines which farmers and fields are more likely to use weed control robots and what stage of resistance development is.

“The exclusive use of herbicides for weed control has led to the emergence of superweeds and we have no new modes of action in the pipeline. If chemical control methods fail, it could result in millions of dollars in crop losses per year,” said corresponding author Madhu Khanna, professor of agricultural and consumer economics in the College of Agricultural, Consumer and Environmental Sciences (ACES) and director of the College of Agricultural, Consumer and Environmental Sciences (ACES). Illinois Institute for Sustainability, Energy and the Environment.

Small, lightweight robots that work under the hood are highly efficient, require little labor, and are environmentally friendly. They pull hoes through the soil and disrupt the emergence of weed seeds. The robots – which are not yet commercially available for corn and soybeans – rely on artificial intelligence for automation and navigation.

The study focused on controlling waterhemp (Amaranthus tuberculatus) in corn cultivation. Waterhemp poses a persistent threat to Midwestern farmland, and the weed has already developed resistance to several herbicides.

The researchers examined the effect of two different types of weed control strategies farmers could use: short-sighted management, which takes one year at a time, and forward-looking management, which takes future consequences into account. They also considered weed seed density, level of weed resistance, and economic thresholds that would trigger the adoption of robotic weed control at the farm level.

“We found that both seed density and resistance levels are important for myopia treatment. For a forward-looking approach, seed density is not important because resistant seeds are likely to spread in the future. This perspective does take into account the level of resistance, but almost any level is enough to trigger adoption,” said co-author Shadi Atallah, an associate professor at ACE.

“Assuming a robot costs $20,000, farmers with a forward-looking management perspective are likely to intervene when 0.0001% of seeds are resistant, while someone with an annual management approach will wait until resistance levels are above 5%,” Atallah noticed.

“So if you're going forward, don't even bother looking at seed density, just resistance level. And no matter how low that is, you should go ahead and let the robots take over.”

The researchers also examined adoption rates and intensity over time. Their calculations showed that farmers with a short-sighted management perspective would not use robots at all in the first six years. These farmers would apply herbicides until they were no longer effective, then switch to 100 percent robotic control—six robots per hectare—in the seventh year, when they had exhausted chemical options.

In contrast, forward-thinking farmers would start adopting robots much earlier and need fewer of them. They would adopt them gradually, not exceeding four per acre. They would use robots to supplement herbicide treatment, ensuring that their effectiveness is not exhausted. In the seventh year, they would use robots on 75% of their land, while 25% would be treated with herbicides.

“We find that short-sighted management initially leads to higher profits because there is no investment in the robots. Forward-looking management initially appears to be worse off because it buys the robots. But it pays off after the sixth year when their profits become higher,” Atallah said.

“Farmers may take the short-sighted perspective when, for example, they rent their land and have to renew it every year, so they cannot really plan for the future. But even for those who work on an annual basis, there will come a point where it is necessary to introduce the robots because other control options have been exhausted,” he added.

The different strategies have implications beyond the agricultural level, as resistant seeds can spread to neighboring fields. A proactive approach can help reduce the number of resistant seeds and potentially help reverse resistance.

Atallah warned that resistance is not reversible in all weed species, but in waterhemp there is a trade-off if the seeds develop resistance; their reproduction rate decreases. Therefore, it is likely that resistant seeds will be overtaken by non-resistant seeds when selection pressure is reduced, he noted.

Researchers focused on maximizing profits at the farm level, but an upcoming study will consider two neighboring farms to understand the spillover effect of resistant seeds. They also plan to conduct a landscape-level analysis to assess impacts over larger areas, which will have further implications for policymakers.

Atallah presented the study results and results of a survey among farmers in one farmdoc daily Webinar.

Khanna is also a professor at the Center for Advanced Bioenergy and Bioproducts Innovation, the Carl R. Woese Institute for Genomic Biology, the Center for Digital Agriculture and the National Center for Supercomputing Applications at the U. of I.

The study “Herbicide-resistant weed control with robots: An ecological-economic weed model” is published in Agricultural economics [DOI: 10.1111/agec.12856]. Other authors include Chengzheng Yu, Saurajyoti Kar, Muthukumar Bagavathiannan and Girish Chowdhary.

This research was funded by AIFARMS, an AI institute at the U. of I., and supported by the USDA's National Institute for Food and Agriculture.

– ACES at the University of Illinois