Cyber-agricultural future arrives now

Drones flying over fields to assess crop status. Tractors steering autonomously, guided by soil maps to deliver tailored doses of fertilizer. Robots in the rows harvesting high-value fruits. What was recently science fiction may already be available from a nearby agricultural retailer.

The farm of the future is arriving, thanks in large part to Iowa State University researchers Soumik Sarkar and Asheesh “Danny” Singh.

“Iowa State has been at the forefront of creating a new discipline, cyber-agricultural systems (CAS), which brings together many areas of expertise to address big agricultural problems and create opportunities for smart, connected and response- agile farms, which could hardly have been imagined a few years ago,” said Singh, G.F. Sprague Chair in agronomy.

Cyber-agricultural systems draw from mathematics, engineering and computer science, building on cyber-physical systems that have revolutionized industries like manufacturing, aerospace engineering and transportation.

Sarkar, a professor of mechanical engineering, worked on cyber-physical systems in industry before coming to Iowa State University in 2014. He soon met Singh, who was experimenting with using machine learning and artificial intelligence to improve variety development for crop production. The two quickly realized the potential of adapting cyber-physical approaches for agriculture.

Recently, they reviewed the new field they helped launch for the journal Trends in Plant Sciences. With co-authors from around campus and the country, they describe how cyber-agricultural systems is complementary yet distinct from precision and smart ag. They outline three main components defining the cyber-agricultural systems framework.

• Data absorption from sensors and cameras of many kinds affixed to satellites, drones and robots that collect a multitude of data on plants, weeds, insects, diseases or animal behavior, along with information on landscape position, weather and environmental conditions.
• Modeling using the data to answer questions and make predictions.
• Decision-support applications based on the modeling results informing digitally based tools designed to answer specific questions, such as ‘What is this disease?’ and “What are recommendations for its management?’