Killer Robots: Weeds Won’t Know What Hit Em’

Palmer amaranth found in the Bootheel of Missouri, 2016.
( Sonja Begemann )

While weeds occasionally dodge herbicides—can they avoid robots? University of Illinois researchers are using a USDA grant to find out.

Resistant weeds cost up to $6 billion dollars each year in yield loss and researchers say that number could jump to $100 billion if chemicals continue to rapidly lose their efficacy. Research into robotic control has started to address this growing, costly issue.

“Big equipment can’t reach between plants after they’ve grown to a certain point,” said CSL's Girish Chowdhary, assistant professor in the Department of Agricultural and Biological Engineering, with appointments in Electrical and Computer Engineering, Computer Science and Aerospace Engineering, in a recent press release. “The robots could autonomously and continuously go through and take care of the weeds underneath the canopy.”

The scientists’ goal is to create autonomous, collaborative robots that weed fields with mechanical implements. They hope this research will give farmers access to a team of reliable, proficient and cost-effective robots to help manage weeds.

This project is bringing together scientists from a variety of disciplines including experts in machine learning, weed scientists and environmental ecosystems specialists. One reason is because weeds can look a lot like the crops they’re growing with—especially during seedling stages.

Machine learning helps address this issue. Experts apply principles that allow robots to acclimate to real-word uncertainties and differentiate between weeds and crops. The team has also developed the robot in a way that allows it to traverse variable, and sometimes rough, field conditions.

“This project could revolutionize integrated weed management, giving farmers a novel, highly effective tool for physical weed control while reducing reliance on, and improving stewardship of, herbicides,” said Adam Davis, head of the University of Illinois Department of Crop Sciences. He provides agronomic and weed management experience to the team building the weed-killing robots.