As greenhouse agriculture continues to expand to meet growing food demands, researchers at the University of Kentucky are developing new technologies to produce healthier crops more efficiently, while also reducing the time and labor required to grow them.
Biyun Xie, Ph.D., associate professor in the Stanley and Karen Pigman College of Engineering's Department of Electrical and Computer Engineering, is the principal investigator on a nearly $1.2 million U.S. National Science Foundation grant to develop an autonomous robotic system that could transform how greenhouse-grown tomatoes are monitored and evaluated.
Xie has partnered with co-principal investigator Qinglu Ying, Ph.D., in the Martin-Gatton College of Agriculture, Food and Environment's Department of Horticulture, on the project that combines robotics, artificial intelligence (AI), computer vision and wireless power technologies to create a mobile robotic platform capable of autonomously collecting detailed information about tomato plants in large-scale commercial greenhouses.
Xie's research aims to improve the speed and accuracy of plant phenotyping - the process of measuring a plant's physical characteristics - while helping growers make more informed management decisions.
"Greenhouse tomato phenotyping provides quantitative measurements of plant traits such as fruit number, size, maturity, health status and canopy structure, thereby improving productivity and resource-use efficiency in controlled-environment agriculture," Xie said.
The project includes four primary research objectives:
- Developing deep learning models that enable the robot to identify tomato plants, understand its greenhouse surroundings and select fruit for analysis.
- Creating new robotic motion-planning algorithms that position cameras for optimal image collection while avoiding obstacles and preventing damage to plants.
- Designing AI models capable of simultaneously measuring dozens of tomato fruit traits and automatically evaluating whether additional imaging is needed to improve data quality.
- Developing an efficient, reliable and cost-effective wireless charging system that allows the robotic platform to operate continuously in humid greenhouse environments.
Together, these advances will enable an autonomous phenotyping platform unlike conventional systems currently used in greenhouses.
Xie's platform uses a mobile robotic arm to navigate greenhouse rows and move directly into the plant canopy. From within the canopy, the robot captures high-quality images without disturbing the plants, allowing AI models to generate precise, noninvasive measurements of individual tomato plants.
The system will also incorporate innovative dynamic wireless charging technology, allowing the robot to recharge while it operates. This continuous power supply enables uninterrupted data collection across large commercial greenhouses - eliminating the need for frequent battery changes and maximizing efficiency.
Beyond improving greenhouse automation, the research addresses fundamental scientific questions across multiple engineering disciplines.
"The technologies developed in this project - including robotic active perception, computer vision models and dynamic wireless charging - have broad applicability to other agricultural operations," Xie said. "It can be applied to selective fruit harvesting and leaf pruning, as well as a wide range of robotic automation systems beyond agriculture."
The research also has important workforce implications. Autonomous phenotyping can provide growers with faster, more accurate information about plant health and fruit development, supporting increased crop productivity.
"Kentucky is home to a large and growing greenhouse industry - making it an ideal setting for the deployment of advanced agricultural technologies," Xie continued. "This project has the potential to improve worker well-being by reducing labor-intensive and repetitive tasks, broaden participation in computing and engineering through student training and outreach activities, and strengthen workforce development in agricultural technology."
Research reported in this publication was supported by the U.S. National Science Foundation under Award No. 2437812. The opinions, findings, and conclusions or recommendations expressed are those of the author(s) and do not necessarily reflect the views of the U.S. National Science Foundation.
