As global food demand increases, Mississippi State University researchers are developing autonomous agricultural technologies aimed at improving efficiency and reducing labor demands in farming.
Led by the Department of Agricultural and Biological Engineering (ABE) and the Agricultural Autonomy Institute, the work focuses on robotics that not only perform field tasks but also collect data that can help farmers improve yields and overall crop performance.
The research is particularly important in Mississippi, where agriculture is the state’s largest industry and farms cover 10.3 million acres, according to the Mississippi Department of Agriculture and Commerce.
ABE assistant professor Hussein Gharakhani is one of the researchers leading the efforts. He is developing an autonomous cotton-harvesting robot designed to operate without a human driver and harvest cotton more precisely than current equipment.
Cotton is one of Mississippi’s most valuable crops, with an estimated value of $372 million in 2024, according to MDAC. However, cotton presents a challenge in harvesting. Bolls ripen at different times on the plant, and current systems are designed to harvest all of the bolls in a single pass. Gharakhani’s system aims to solve that challenge.
“That’s one of the main concepts as a land grant university. It’s one of our highest priorities,” Gharakhani said. “We make sure that we are solving high-priority issues related to agriculture in the state first, as well as the nation.”
ABE Department Head Alex Thomasson founded the Agricultural Autonomy Institute in July 2023 and became its director. As farms across the country struggle with labor shortages and rising production demands, Thomasson said he believes the development of autonomous equipment is essential for agriculture to keep pace with a growing global population.
“We want to be a leader nationally and worldwide,” Thomasson said. “In fact, as far as I can tell, we’re the only university institute that’s completely focused on autonomous systems for agriculture.”
Thomasson said this work is an opportunity for economic development in Mississippi.
“We want to do corporate engagement like we’re doing with John Deere and the Sugar Cane Growers Cooperative of Florida, to try to bring their business in the state,” Thomasson said. “That could mean, you know, that they’re just funding projects, but it could also mean that they are, you know, putting people here, like in a research capacity, or maybe they’re building a distribution facility or a manufacturing facility that’s part of our long-term economic development plan for the state.”
One major difference in much of the technology being developed is the move from large, single machines toward fleets of smaller, autonomous ones. Thomasson said that, historically, equipment had to be large to maximize efficiency for a human driver. By eliminating the need for a human operator, autonomous machines can be smaller, more precise, less harsh on the soil and more reliable.
“If you’ve got, let’s say, five machines going through the field and one of them breaks down, you still have 80 percent of your operation functioning,” Thomasson said. “If you have one large machine and that large machine breaks down, you have 0 percent of your operation moving forward.”
Gharakhani’s cotton harvester reflects this shift. Because current large harvesters only allow for one pass for harvesting, all of the bolls have to be harvested at the same time. Gharakhani said that by using machine vision and learning, the robot is able to identify individual cotton bolls and harvest them selectively.
Gharakhani said advances in artificial intelligence have made this level of precision increasingly possible.
“Artificial intelligence is improving drastically, and it has been applied on different things, starting from perception systems or machine-vision systems,” he said.
Researchers in the department are also developing systems to aid in field data collection. Moeen Ul Islam, a second-year Ph.D. student in ABE, is developing an autonomous ground vehicle designed to work in coordination with drones to collect high-quality agricultural data.
“The thing about the new era of agriculture is people are trying to get as much data as possible,” Ul Islam said. “…They’re not just making robots that can go on the field and make labor easy. But rather, they’re more focusing on getting quality data.”
This data can then be analyzed by specialists, identifying diseases and different problems with crops and helping growers improve their yields.
“So once you know a lot of information about your field, then you can effectively work to improve it,” Ul Islam said. “So what people normally do is, like, they go by their instinct. Sometimes their instincts are right, sometimes not.”
Ul Islam said that artificial intelligence has been key in training robots to work in unpredictable environments.
“I think nowadays it’s really hard to not use AI in like really advanced robots, because especially when in agriculture, everything is so unpredictable,” Ul Islam said. “So you need to have, like, some sort of training on the robot that you do prior to releasing it on the field for a mission so that it can handle unpredictable situations based on what it has learned from, like AI or machine learning models.”
Ul Islam plans to field test his robot this week, with a target completion date of April. Ul Islam said the data gathered by these systems, with the help of his ground robot, will be not only useful to local farmers, but also to researchers globally working to learn about agriculture and make it more efficient.
“So that’s, I think, like a goal for us, so that we can develop a system that can provide, like a really solid backbone to research done by people all over the world,” Ul Islam said.
Autonomous agricultural technology is already in the early stages of being applied. Thomasson said he believes that, by 2040, a significant portion of the machines doing agricultural work will be autonomous, and that MSU hopes to lead that charge.
