May 2 2017
If you’ve moseyed past the greenhouse on the south side of Owens Science Hall (OSS) after dark, you may have wondered what the purple incandescence radiating from one end is all about. No, the Biology Department isn’t staging nightly raves in homage to Prince. Nor is it a nod to the university’s official color.
Engineering major Andrew Ryan adjusts the FarmBot as it hovers over a bed of lettuce. (credit: University of St. Thomas, Minnesota)
The source of the glow comes from a set of LED grow lights, the same kind used to grow plants on NASA’s International Space Station. Here, the lights are helping a large table of three types of romaine lettuce plants grow optimally. It is a greenhouse, after all.
The lettuce, however delicious and nutrient rich, is merely a conduit for the School of Engineering’s Precision Agriculture Research Team. The team’s focus is to provide students with rich learning opportunities and real-world scenarios relevant to working in the current industry. They are conducting research using a customized FarmBot. This modified robot is just one project aimed toward an overall goal of addressing the world crisis of food shortages and the development of new technologies to grow healthy crops using less man-made and natural resources.
Or more accurately, “We modified a FarmBot from an open-source template so the software met our requirements. The track it’s running on is movable, so it’s not strictly a FarmBot anymore,” said Andrew Hafferman, development systems administrator, Computer and Information Sciences Department (CISC), who oversees the integration of software and hardware of the bot’s components. “We’re calling it the ‘UST horticultural robot.’”
Currently in its developmental phase, “the immediate goal right now is to collect data for image processing and analysis of plant health,” which is being done by Dr. Chih Lai, a Graduate Programs in Software professor, Hafferman said.
A sophisticated multispectral camera that has been mounted face down on the bot takes images of the plants below. This camera has the capability to look at plants in a number of different spectral wavelengths. The collected image and metadata is sent to the Precision Agriculture Cluster where the information is held in a “data lake.”
The collected data is then analyzed to compute health indices (on the health of individual lettuce plants). These indices reveal important information on where and when the health of the plants is impacted due to environmental factors so the team (and eventually, farmers) can make decisions on how best to optimize the yield of the growing crop.
Ultimately the team plans to develop algorithms that are able to make recommendations to the farmer/grower on how best to: improve crop health; reduce/prevent infestation by insects/virus/fungus/weeds; determine areas where plants needs more or less water; determine areas where crops are not getting enough nutrients from the soil; and what, when and where to plant in the subsequent growing cycle.
In keeping with the team’s “precision” moniker, the UST horticultural robot requirements evolved from the previous work. Last year, the PA Research Team worked with a camera mounted on a mechanical stand that was not automated. Team members had to move the camera by hand and manually click the camera to take a picture. Now, the camera moves up and down, front to back and left to right using a programmable microcontroller housed in “the cluster” (a data center comprised of a cluster of machines in OSS). The bot has the technology to be connected to the internet and operated using any device that can connect to the internet. The team plans to develop and integrate this feature at a later stage.
The camera on the robot also may be interchanged with devices that can be used for watering, seeding, fertilizing and weeding.
“Where we’re heading is basically using automation to grow food,” said Dr. Christopher Greene, a professor in the School of Engineering and the computer engineering chair who has been leading this effort, in all its evolutionary phases, for several years. “It will water when it needs to, seed the soil, fertilize, and find and replace weeds. Ultimately, it will even harvest the crops.”
Dr. Cheol-Hong Min, a professor in the School of Engineering who works on the project, added that this type of automation is good for quality control and productivity but also addresses the shortage of farming help.
“The younger generations, generally speaking, don’t want to run farms today,” Min said. “There’s not enough farm workers and the world population is increasing. This is a perfect example of how technology can work with farmers to improve yielding, growth, productivity and much more.”
Peter Farley, a freshman computer engineering major, has been working, paid, roughly 10 hours per week on “whatever needs to be done,” he said. His duties have included helping to hang the LED lights and assisting Hafferman in developing the software used to automate the robot.
“I was surprised they gave me all the responsibility they have,” said Farley, who has been computer programming since elementary school when his brother brought home a book on the subject. “It’s been challenging but fun working with unfamiliar technology.”
Soon the team will introduce different seeds – corn and perhaps soybeans – to the lettuce crop to test the horticultural robot’s ability to differentiate between weeds and crop. Eventually, perhaps as early as this summer, the UST horticultural robot sub-system, with camera, will be attached to a rover (an unmanned autonomous ground vehicle) and tested outdoors on a farm as a precision agriculture system.
Scott Morgan, pro bono consultant to the School of Engineering and principal of Prescription Robotics, is the project’s sponsor, system architect and lead mechanical engineer/mechanic. Students, such as junior electrical engineering major Andrew Ryan, have been assisting him with the electro-mechanical detail design, fabrication and integration of the UST horticultural robot and rover.
“It’s an open-source project that we are redesigning, and it’s been challenging to figure out how to make things work,” Ryan said. “It’s not as simple as putting pieces together. I like that I’m getting actual experience. I’m designing and building the robot on a good project that can go places, which you don’t normally get to do as a student.”
Faculty and students from several departments at St. Thomas have contributed to the project, a joint venture between the University of St. Thomas School of Engineering, CISC and the Biology Department. Currently, 13 people are working on the horticultural robot project. Other faculty, staff and students not previously mentioned in the article include: Dr. Chong Xu, a fellow in the School of Engineering; Dr. Peter Gittins, Chemistry Department, College of Arts and Sciences; Catherine Grant, greenhouse manager; and St. Thomas students Jacob Komarek, Jiwoon Hur, Taro Vue and Karthik Kantipudi.