KSU Research Advances Computer Circuitry to Measure Soil Content
Wednesday, February 1, 2023
filed under: Research and Development
Can an electronic circuit — not much larger than a postage stamp — help to speed up U.S. farmers’ goal to feed a hungry world?
It might, says Raj Khosla, if it’s taught to measure soil properties of a farm field so that in a matter of seconds, farmers can adjust water, nitrogen and other inputs to abundantly grow crops.
Think of it as farming in bits and bytes, in real time.
“The United States has constructed an agriculture innovation agenda that in the next 28 crop cycles — which gets us to the year 2050 — we want to grow 40% more food than what is currently grown,” says Khosla, a precision agriculture specialist, professor and head of the Kansas State University agronomy department in the College of Agriculture.
“But here is the caveat: We need to achieve this goal by using 50% less water and 50% less nitrogen applied to the crops, which are the two biggest drivers of crop production systems,” Khosla adds. “This means we have to account for everything that goes into the water and nitrogen budget. We can’t leave room for error.”
For the past 10 years, Khosla has been methodically reducing the chance of error in several projects to develop biodegradable sensors that measure soil moisture and nitrogen content.
“Ten years ago, I was hypothesizing that someday we would have soil moisture sensors that we could literally throw out in a field, then ‘ping‘ping’ them with a computer to get a measure of that soil’s moisture content,” Khosla says.
Khosla began conducting field experiments in 2012 in collaboration with private partners that provided sensors mounted on a post and connected to cables that measured soil moisture at five depths. In that setup, Khosla determined that to cover a 22-acre field, a farmer would need about 100 sensor nodes— each at a cost of about $3,000.
“They were expensive," he says. “Even today, they are cost prohibitive for a farmer. Installing those sensors is labor intensive and it’s a logistical nightmare to have 12-foot-tall posts sticking out of your 22-acre field at such a high density.”
Those early experiments, though, served a purpose: to fuel the research group’s motivation to provide a low-cost option that could still gather detailed information of the farmer’s entire field.
Now, Khosla is co-leading a collaborative team that is building on those early experiments. In 2018, Khosla and colleagues at the University of Colorado, Boulder and the University of California, Berkeley received a U.S. Department of Energy grant that funds high-risk/high-reward projects through the Advanced Research Project Agency.
The team of researchers includes materials scientists and computer and electrical engineers from the collaborative institutions. Other Kansas State agronomy researchers involved include Jeff Siegfried and Dipankar Mandal, both postdoctoral research fellows; Wub Yilma, doctoral student; and Ross Unruh, assistant scientist.
Together, the researchers aim to develop biodegradable sensors capable of measuring soil moisture content at high spatial densities. That would allow researchers to estimate soil moisture at every inch of a field and provide huge volumes of data that are crunched by computer algorithms to build an easily readable guide for the farmer.
“There is no spot in the field where there will be guesswork,” Khosla states.
Farmers already can apply water, nitrogen and other nutrients in very precise ways, using such current technologies as variable-rate irrigation that can be adjusted to provide different rates of water in a field, Khosla relates. But knowing the field’s needs, foot by foot, is limited to measurements provided by satellite images or unmanned aerial vehicles. Those are good ways to accommodate a field’s needs, but still not entirely precise, diagnostic or immediate.
“One idea is that as a pivot is applying water in a field, you can ping sensors that are lined up in the next 20, 50 or even 100 feet of the pivot arm,” Khosla notes. “That information is sent back to the computer to re-create the real-time soil-moisture data surface that the pivot is encountering while it is applying the water, and the farmer or artificial intelligence-based decision tools can change the rate of application if necessary. I think that’s going to be a big deal.”
If farmers were to deploy 100 sensors throughout a field, the sensor cost — at 50 cents to $1 each — would be $50 to $100. Setting them up would be as simple as walking the field and tossing or inserting them about. Because they would be biodegradable, they would never have to be collected.
The eco-friendly biodegradable chips, which Khosla thinks will start to deteriorate in about 200 days, are not yet available. Currently, the research team is using larger, more expensive circuits that are not biodegradable to make sure the huge volumes of data they are collecting can be processed by computers using algorithms developed by KSU researchers to translate data that enables farmers to make better decisions.
“We can only manage what we can measure,” Khosla affirms. “If we can’t precisely measure the resources that we’re trying to manage, then we won’t be able to help farmers. I think that’s particularly true for these two major inputs in crop production systems — water and nitrogen— that are environmentally so sensitive and important.”
This work, he says, is the first step toward an exciting agricultural era that involves more sensors and data-driven decisions.
“These types of technologies often are for organizations with a very high demand for information technology. They’re usually the first ones to get their hands on it,” Khosla says. “Well, this time it happens to be agriculture. It’s very exciting to be in that environment.” — Pat Melgares / K-State Research and Extension News Service