Feb. 4, 2025

NDSU professor’s electromagnetic and machine learning research working to tackle agricultural challenges

SHARE

Ben Braaten isn’t afraid of tackling a big challenge and there are plenty of those in the current agricultural landscape.

Braaten’s experience suits him well for confronting the problems facing producers. He is the chair of the NDSU electrical and computer engineering department. His background is in applied electromagnetics, which studies how electric and magnetic fields interact with materials. He is aiming to solve the complex problems in the electrical and computer engineering realm – and applying that discipline to solve the difficult challenges facing agriculture. 

Those challenges include changing climate conditions, producing more food for a growing world population and environmental sustainability. 

“Agriculture is one area that is very complex from an electromagnetics point-of-view,” Braaten said. “Engineering design problems in this space include various types of soils throughout North Dakota, different plants at different developmental stages, farming equipment consisting of many different materials that result in numerous possibilities for electromagnetic scattering, all while being spread across a vast landscape for farming. I’m drawn to these problems because they are particularly difficult to solve.”

Braaten’s work with electromagnetic sensors and machine learning is part of NDSU’s Food, Energy and Water Security initiative. Machine learning is a form of artificial intelligence that enables computers and machines to imitate the way humans learn and to perform tasks autonomously. Braaten is applying machine learning to the data that has been collected from sensors he is developing. Electromagnetic principles can help improve seed germination and dormancy release in agriculture.

“The fusion of machine learning and electromagnetic sensors is our initiative to address problems and shortcomings in North Dakota agriculture with the latest technology,” Braaten said. “The complex environment of agriculture includes not just electromagnetic principles; the needs for sensing environmental variables. Collecting a vast amount of data is required, and utilizing this data in a useful way is important and getting it into the hands of growers in North Dakota is paramount.”

Two examples of this are the research that impacts sugar beets and honeybees. In 2023, the sugar beet industry in North Dakota and eastern Montana had an estimated economic impact of more than $6 billion. However, during the winter months, under certain weather conditions, sugar-beet storage can result in the loss of tens of millions of dollars in lost revenue.

“Incorporating disciplines outside of traditional agriculture, like Dr. Braaten’s work in electromagnetic sensing and machine learning, is crucial for advancing precision agriculture in North Dakota,” said Frank Casey, North Dakota Agricultural Experiment Station associate director. “By integrating cutting-edge technologies, data collection methods, and artificial intelligence, we are positioning the state as a leader in agricultural innovation. These advancements not only enhance efficiency and sustainability but also ensure North Dakota remains at the forefront of national and global ag tech developments.” 

But there are indicators in a sugar beet pile that can show if a portion of it is beginning to spoil – these include temperature fluctuations, changes in the shape of the pile, gasses emitted from a particular region of the pile and changes in electrical properties of the pile. Using machine learning, NDSU is pairing with Minn-Dak Farmers Cooperative in Wahpeton, North Dakota and the USDA Fargo site to build technology that would indicate to growers what is causing the spoilage. 

“Our team is working closely with personnel from those two agencies to develop technology that can measure these indicators and use machine learning to inform producers of this spoilage, allowing for actions to be taken to minimize further loss,” Braaten said.

Another area of research is focusing on the populations of bee colonies. North Dakota is the top honey-producing state in the nation, having produced 38.3 million pounds of honey in 2023, according to the USDA.

Bees are also responsible for pollinating more than 100 crops in the U.S. But bee colonies are facing obstacles to maintaining their numbers, including the habitat loss, invasive pests, nutrition and pesticides. Braaten’s team is working with the USDA to assess the environments near honey bee colonies and transporting.

“The team is using ultra-wideband radar to explore how electromagnetic fields can measure the amount of honey production in a hive,” Braaten said. “Discrete sensors measuring temperature, humidity and air pressure are being developed to determine environmental stress around bee colonies during transportation. We are also developing machine learning-based software to interpret this data and present it in a manner that is useful for North Dakota’s honey producers.”

The research comes at a time when the world population is growing, meaning food demands on existing natural resources is increasing. In addition to honey production, North Dakota ranks first in the nation for crop production of 11 different grain varieties. Accurately and efficiently applying products like fertilizers, herbicides and insecticides can reduce costs for growers. In addition, there are environmental benefits by reducing water usage and chemical runoff, and consumers may benefit with lower grocery prices, which all ties into the FEWS initiative’s goals.

“Precision agriculture gives growers tools to tailor their application rates down to the square foot resolution. Improving growers’ working resolution from the size of a field down to a square foot is a win-win,” Braaten said. “Today’s farming equipment is very capable and is equipped with a large array of sensors and is programmable; in many cases, the equipment is already capable of adjusting outputs real-time or on-the-fly. What is lacking is sensor input and situational awareness.”

The further development and use of sensors will be part of North Dakota’s agriculture landscape for years to come.

“Machine Learning is a powerful theory for collecting, organizing, interpreting and presenting data in a tool for growers, whereas the theory of electromagnetics lends itself to the physical principles of sensing technologies and wireless capabilities in the field,” Braaten said. “Many of the challenges in North Dakota agriculture and elsewhere can be addressed with machine learning and electromagnetics.”

This research work was supported by the United States Department of Agriculture under agreement 58-3060-3-022.

Submit Your News Story
Help us report what’s happening around campus, or your student news.
SUBMIT