STORY: Students Developing Autonomous Weeding Robot for Sustainable Farming

May 21, 2024

A team of Olin students is developing a robotics system that uses artificial intelligence (AI) to improve farming practices by limiting reliance on harmful herbicides in weed management - and they’ll be submitting their proposed system to the Farm Robotics Challenge this summer.

The project began last fall in the laboratory of Kenechukwu Mbanisi, assistant professor of robotics engineering, with conversations with several local farms to learn more about challenges they have that could be helped with both robotics and AI.

“We spent about a month interviewing local farmers, and one of the things that emanated from those conversations was the difficulty of weeding, especially on small farms,” says Mbanisi. “The conventional approach to weeding is to use chemicals, which can be detrimental to crops and has other negative environmental impacts. We decided to focus on building a robotic weeding system that minimally disturbs the soil and safely uproots weeds around crop plants without the use of chemicals.”

To that end, the students are creating a small weeding robot that can drive through fields, using computer vision and AI technology to autonomously identify weeds from crops and a mechanical arm to pluck them out of the ground. Students are organized into subteams that work on individual subsystems, such as the mechanical weeding arm or the weed identification algorithm; each subteam also has an agricultural liaison who coordinates with the farmers to ensure that the end design suits the users.

“Collaboration within our team and with the farmers has been something we’ve been trying to make sure we’re really focused on and have found to be quite helpful,” says Brooke Moss ’25, who began working with Mbanisi from the very beginning as part of an independent research project for her self-designed major in robotics for sustainability. “We want to make sure that there’s crossover between the subteams so that we’re not making assumptions on what other teams needed, as well as stay connected to the farmers so we’re in tune with their needs.”

“A lot of people had no prior experience with robotics before joining this project, but we all just hit the ground running,” says Maya Adelman ’27, who is a co-lead on the mechanical weeder subteam. “I worked in Kene’s robotics lab last semester, and I joined this project because I wanted to learn even more about robotics and sustainability.”

This summer, the team will submit a demonstration of and report about their robot's capabilities for the Farm Robotics Challenge, a competition aimed at revolutionizing the future of farming through robotics technology. If they are among the winners, they could be invited to present their prototype at the conference this fall.