AI-driven digital twins in agricultural research hold the promise for better crops

12-06-2024

Purdue CS Professor and Associate Head, Bedrich Benes and a computer simulated sorghum field.

 

Researchers are developing digital corn twins to simulate “what-if” scenarios for future crops.

Researchers are using AI to change agricultural research, and the results may reach your plate. Bedrich Benes, professor and associate head of computer science at Purdue University, is working with a team of researchers, including James Schnable from the University of Nebraska and Baskar Ganapathysubramanian from Iowa State University on a three-year project sponsored by $2M from the National Science Foundation (NSF) to develop computer models of crops that will allow for the exploration of possible scenarios and the discovery of genes that control plant shape and function, which can lead to better crops that will be more efficient in the current climate and adapt to possibly changing climate.

___________________________________________________________________________________

“This project develops digital twins of corn fields, simulating their function, shape, and interaction with light. The digital twin will allow for fast simulation of possible scenarios and automatic optimization of fields," says Professor Bedrich Benes.

___________________________________________________________________________________

This project brings together researchers from different disciplines. Benes leads this effort with 25 years of experience modeling trees and digital vegetation. James Schnable is a Professor of Horticulture and Nebraska Corn Checkoff Presidential Chair of Agronomy & Horticulture. Baskar Ganapathysubramanian is Anderlik Professor of Engineering, Faculty Fellow of the Plant Science Institute at Iowa State University, and the director of the NSF/USDA funded AI Institute for Resilient Agriculture.

The digital twin of Sorghum (left) and the same physical plant in the phenotyping facility (middle) simulates growth and interaction with light (Bosheng Li, Mathieu Gaillard, Ian Ostermann, Bedrich Benes from Purdue University, and James Schnable from University of Nebraska).

 Plant Ideotypes and Crop Optimization

Despite the unprecedented success of corn production efficiency in the past decades, from 20.5 bushels per acre in 1930 to 183 in 2024, according to the USDA National Agricultural Statistics Service, additional improvements are needed to meet the nation's needs. This project promises to use advanced computer simulations and shape modeling, with the help of AI, to explore possible future crops.

While field experimentations require planting and measuring the yields after the harvest, computer simulations will shorten this cycle by simulating certain features of real fields in silico (i.e., by simulation software).  Moreover, the virtual experiments will allow simulations of plants that don’t yet exist and under hypothetical environmental conditions.

A plant ideotype is a theoretical plant featuring optimal characteristics for an environment. This NSF-sponsored project aims to find plant ideotypes in a changing environment. Researchers will create carefully calibrated computer plant models reflecting the existing conditions and then use computer simulations to determine possible shapes and behavior in a changing environment. Eventually, the project will identify the genes that could lead to better crops.

The Research

The project will focus on two aspects of maize. First, the researchers will experiment with plant shape and light absorption to increase yield (specifically light use efficiency). Experimenting with hypothetical scenarios can lead to plants that do not exist in nature, for example, with longer and broader leaves, different leaf curvature, or different reflectivity, but could be better suited for increasing yield. Second, the plant distributions will be optimized to find a better plant seeding distance, different planting regimes, and plant orientation.

The environment is the most important aspect affecting plant growth. This project will use measured current conditions, such as temperature, carbon dioxide, light and water, and integrate the conditions into the simulation. The researchers will be able to modify the environment in the computer and simulate new plants.

These simulations are computationally demanding, and the team will use high-performance computing to evaluate plant illumination on a photonic level, their interaction with light, and their self-shadowing. The digital twins will grow and adapt to their environment. Simultaneously, the project will optimize plant arrangements to maximize their efficiency.

Simulated illumination of a field composed of digital twins of plant provides the amount of light reaching each leaf (displayed as the intensity of gray on the left). The GPU-supported method uses data from real sky illumination (Bosheng Li and Bedrich Benes, Purdue University).

This computer-generated simulation shows an initial result in which the sorghum field is illuminated from a clear sky, and the amount of light on each leaf is calculated. The highly optimized calculation takes under 30 seconds on a standard desktop computer.

“We are developing a suite of algorithms that will provide reliable 3D reconstruction and creation of digital twins of real plants that will be quickly put into the simulation and used for virtual experimentation,” says Bedrich Benes.  “This will eventually lead to plant ideotypes, a plant and the field best suited for the given environment.”

However, the project does not stop with the shape and function of maize. Schnable and Ganapathysubramanian will study the genes that control the biological and behavioral traits of digital ideotypes, research which may ultimately lead to the production of better crops.

 

About the Department of Computer Science at Purdue University

Founded in 1962, the Department of Computer Science was created to be an innovative base of knowledge in the emerging field of computing as the first degree-awarding program in the United States. The department continues to advance the computer science industry through research. US News & Reports ranks Purdue CS #8 in computer engineering and #19 and #18 overall in graduate and undergraduate computer science. Additionally the program is ranked 6th in cybersecurity, 8th in software engineering, 13th in systems, 15th in programming languages and data analytics, and 18th in theory. Graduates of the program are able to solve complex and challenging problems in many fields. Our consistent success in an ever-changing landscape is reflected in the record undergraduate enrollment, increased faculty hiring, innovative research projects, and the creation of new academic programs. The increasing centrality of computer science in academic disciplines and society, and new research activities—centered around foundations and applications of artificial intelligence and machine learning, such as natural language processing, human computer interaction, vision, and robotics, as well as systems and security—are the future focus of the department. cs.purdue.edu

About Purdue University

Purdue University is a public research institution demonstrating excellence at scale. Ranked among top 10 public universities and with two colleges in the top four in the United States, Purdue discovers and disseminates knowledge with a quality and at a scale second to none. More than 105,000 students study at Purdue across modalities and locations, including nearly 50,000 in person on the West Lafayette campus. Committed to affordability and accessibility, Purdue’s main campus has frozen tuition 13 years in a row. See how Purdue never stops in the persistent pursuit of the next giant leap — including its first comprehensive urban campus in Indianapolis, the Mitch Daniels School of Business, Purdue Computes and the One Health initiative — at https://www.purdue.edu/president/strategic-initiatives.

 

Writer: Molly Walker, walke598@purdue.edu

Source:  Bedrich Benes, professor and associate head of the Department of Computer Science at Purdue University