Purdue Uses IoT to Reinvent Farming, Boost Output

Digital agriculture harnesses the Internet of Things and wireless to revolutionize farming and increase food production for exploding population.

The goal is as lofty as they come: to feed the

world. Agricultural output will need to double by 2050, when the global population will hit an estimated 9 billion people. High tech -- in the form of the Internet of Things (IoT), wireless and mobile technologies, and automated data collection and analysis -- is leading the charge.

Purdue University's College of Agriculture has developed an agriculture-centric IT infrastructure including advanced IoT sensors and devices that will allow researchers to study and improve plant growth and food production processes. "Every process from farm to table and in between has potential for improvement," stresses Pat Smoker, director of Purdue Agriculture IT, in West Lafayette, Indiana. "Technology will play a big role in helping us achieve that."

The digital agriculture initiative kicked off about three years ago. Recognizing the need for a partner in innovation, Purdue sought proposals from a number of its IT partners. The university decided that Aruba, a Hewlett Packard Enterprise (HPE) company, was most closely aligned with its goals.

"Their proposal nailed the challenges and opportunities we were facing, recognized our strengths and had a vision on how to use them to grow the agricultural technology economy," Smoker says. "They had infrastructure solutions that enabled the very things we wanted to focus on."

Another plus: HPE's wireless management started not with the core but with the end devices, what Smoker calls "the adaptive stuff that happens at the edges." He adds, "They had the hardened equipment we needed in a demanding environment -- a wet, dirty field with many varieties of animal and insect cohabitants."

Purdue established a partnership with HPE blending research, innovations and disruptive technologies such as IoT to revolutionize farming research with digital agriculture. In fall 2016, the university began installing an IoT network on its 1,408-acre research farm, the Agronomy Center for Research and Education (ACRE). The system is now being built out to production levels and should be complete by Q1 2018.

The system captures terabytes of data daily from sensors, cameras and human inputs. To gather, aggregate, process and transmit such volumes of data back to Purdue's HPE supercomputer, the university is leveraging a combination of wireless and edge computing technologies. ACRE researchers can enter data into a mobile device onsite and transmit it via the Aruba wireless network, first to the HPE Edgeline systems and then to an HPC data center for analysis.

Innovations include solar-powered mobile Wi-Fi hotspots, an adaptive weather tower that provides high-speed connectivity across the entire ACRE facility and the PhenoRover, a semi-automated mobile vehicle that roams throughout ACRE research plots capturing real-time data from plant-based sensors. Pilot projects include an Unmanned Aerial Systems (also called drones) program to advance the effectiveness of plant-growth data collection.

The digital process is a huge timesaver. "Previously, our faculty had to design data handling protocol," Smoker explains. "They had to figure out how to get data off the sensors and back to the lab, develop analytics and get someone to write the code. We need to take all that off their plate so they can focus on the science and don't need to think so much about the mechanism."

It also increases the speed and accuracy of research. For example, researchers using mobile devices in the field can transmit information about seed growth back to central labs to analyze the impact of water levels, fertilizer quantities and soil types. In turn, the labs can communicate their analysis back to the field to allow quick adjustments.

The potential is enormous. Digital agriculture could enable "precision farming," in which fertilizer levels or irrigation requirements are adjusted for sections or individual plants, and automated equipment can apply the ideal treatment for specific weeds, a boon for sustainability. "At some point, could the plants themselves report the presence of disease or insect infestation for the farmer to act on?" he asks.

Initially, the College of Agriculture will be the only one using the Big Data pipeline. But that's just the beginning. "This college is one of 11 at the university," Smoker says. "At some point, the next-generation infrastructure we are creating will become mainstream and eventually provided to the research community at large. Then we'll be off doing the next new thing at the edge."

By Eileen McCooey