Key Takeaways
- Researchers at East Carolina University’s Center for IoT Engineering and Innovation have built a low‑cost sensor network (Piton) that feeds data to an artificial‑intelligence agent capable of delivering real‑time farm management advice.
- The AI‑driven system has already reduced a Hyde County farmer’s daily water‑level checks from a 20‑minute manual task to instant text alerts, saving time and reducing flood risk.
- Beyond water monitoring, the platform is being adapted to track salinity intrusion, grain‑bin storage conditions, greenhouse temperature and humidity, and propane levels—addressing multiple pain points identified by local growers.
- By demonstrating how IoT and AI can make farming more efficient and less labor‑intensive, the project aims to attract younger generations back to family farms, bridging the perceived divide between technology and agriculture.
- Funding remains a challenge; the team relies on modest grants and personal time, emphasizing that their goal is to create an affordable, scalable service for small‑ and medium‑sized farms that lack access to high‑budget tech solutions.
Introduction and Project Overview
East Carolina University researchers are turning data collected from Hyde County canals and fields into actionable, real‑time information for farmers through artificial intelligence. Dr. Ciprian Popoviciu, assistant professor in the Department of Technology Systems and director of the Center for IoT Engineering and Innovation, explained that the initiative began as a effort to help farmers “better understand what’s happening on their farms.” The center’s flagship tool, Piton—Platform for IoT Open Networks—deploys inexpensive sensors to gather environmental data, which is then processed by an AI agent that communicates directly with growers via a simple mobile app.
Development of the Piton Platform and AI Agent
When the project launched, Popoviciu and his team performed the data analysis themselves, creating a bottleneck that delayed information delivery to farmers. Graduate student Colby Sawyer, a Camden County native whose grandfather was a retired farmer, joined the effort while researching how large language models (LLMs) could be applied to agricultural data. Sawyer recalled, “I started working with a large language model, a popular AI program that is used for ChatGPT and Google’s Gemini. LLMs know how to read and write.” By teaching the LLM to interpret sensor outputs and to generate farmer‑friendly recommendations, the team transformed raw numbers into concise advice such as, “Your water level is getting pretty high. You may want to consider something.”
Water‑Level Monitoring in Hyde County Canals
The first real‑world test involved a Hyde County farmer who manually checked canal water levels with a measuring stick once or twice daily. Popoviciu noted, “What the pump is actually doing is preventing portions of the county from flooding… It’s a super important job, but this is not what he’s paid to do.” Sawyer and Popoviciu installed sensors costing about $150 each in the canals. The sensors streamed data to the AI agent, which then sent text alerts to the farmer’s phone. The farmer could reply with queries like, “What is the water level? What does it look like? Do you think it’s going to be a problem today or is it a problem for tomorrow?” The AI, drawing on historical patterns and the farmer’s own responses, would reply with guidance such as, “Actually, you’re OK. It’s less than 2 feet and you’ve considered that pretty normal. You don’t have to do anything today,” or, “Maybe you should consider it tomorrow.” This exchange eliminated the need for a 20‑minute field trip each day.
Time Savings and Operational Benefits
Sawyer emphasized the tangible impact: “This exchange has saved the farmer a 20‑minute trip to measure the water levels, he said.” By automating routine monitoring, the farmer can allocate more time to core farming activities—planting, harvesting, and market decisions—while still maintaining optimal water management. The system also reduces the risk of accidental flooding or over‑draining, which can damage crops and soil structure.
Expanding to Salinity Intrusion Monitoring
During a subsequent visit to Hyde County, the farmer highlighted another concern: saltwater intrusion compromising field productivity. He asked how he could monitor the intrusion and assess its effect on crops. Sawyer explained, “We can check the salinity of the water, the ditches, and we’re checking the salinity in the soil. So we can see if maybe that water is leaching into the soil, what effect that’s having, what temperature does to that, what weather does to that, and we can deliver that to the farmer as another portion of that AI.” By integrating salinity sensors into the Piton network, the AI can now alert growers when rising salt levels threaten soil health, enabling timely mitigation such as flushing with fresh water or adjusting irrigation schedules.
Grain‑Bin Monitoring and Fire Prevention
The team also deployed sensors inside grain bins to measure stored grain volume, removing the need for workers to climb to the top with a measuring device. This continuous data stream helps farmers plan future production or purchases with greater accuracy. In parallel, they explored temperature monitoring inside bins to prevent spontaneous combustion—a hazard that can ignite stored grain. Earlier this spring, they worked with a farmer to develop a process for measuring bin temperatures, aiming to trigger alerts before conditions become dangerous.
Collaboration with RIoT and Greenhouse Experiments
Popoviciu and Sawyer are engaging six additional Hyde County farmers about potential projects and are partnering with RIoT (Regional Internet of Things), an incubator that supports rural entrepreneurship by weaving technology and AI into farming and other businesses. RIoT operates a facility in Wilson where growers can trial non‑traditional crops suited to the local climate. The ECU team provides sensors and instrumentation to monitor temperature, humidity, and irrigation in greenhouses and hoop houses. When a cold snap earlier this year caused propane tanks to run dry overnight, Popoviciu recalled, “The growers didn’t find out until the morning.” The AI is now being trained to warn growers when temperatures and gas levels fall, with the long‑term aim of enabling automatic corrective actions.
Inspiring the Next Generation of Farmers
A motivating factor behind the work is the hope of retaining young people in agriculture. Popoviciu noted that the Hyde County farmer’s son, a technology student at NC State University, is uncertain about returning to the farm. “Maybe now that we have this IoT, we have these AI capabilities on the farm, and we’re doing smart agriculture, we can entice those students, entice those kids to come back and serve the family,” Popoviciu said. Sawyer added, “He can see it on his phone, he can interface with AI. He can see how the technology layer makes these things better, and it’s not that technology and farming are completely separate… We’re bridging that gap, bringing technology to the farms, hopefully inspiring the next generation of farmers to stick around.”
Funding, Sustainability, and Future Directions
While the project benefited from initial grant funding, Popoviciu and Sawyer have invested considerable personal time, acknowledging financial constraints. Sawyer remarked, “We’ve been working kind of from the grassroots up and so we’ve been keeping that in mind since day one and so far it’s worked out for us.” Popoviciu contrasted their approach with well‑funded initiatives in wealthier parts of North Carolina, stating, “There are other projects that are built for communities in wealthy areas of North Carolina… They got millions in grants. I don’t think they achieved as much as we did. But they have millions in grants. Our mission was we want this for the folks who don’t even know that this is possible.” The researchers envision transitioning the effort into a self‑sustaining service—potentially supported by county or state funding—but remain uncertain about the exact economic model. Their overarching goal is to make the technology “easy for all these small, medium farms to take advantage of this technological revenue,” ensuring that innovation serves those who need it most rather than only high‑budget operations.
Conclusion
Through the Piton platform and an AI agent built on large‑language‑model technology, ECU’s Center for IoT Engineering and Innovation is delivering low‑cost, real‑time farm management tools to Hyde County growers. From automating water‑level checks and monitoring salinity intrusion to safeguarding grain bins and greenhouses, the system addresses multiple, interlinked challenges identified by farmers. By saving time, reducing risk, and demonstrating the tangible benefits of technology on the farm, the project not only improves current operations but also seeks to inspire a new generation to view agriculture as a viable, tech‑forward career. As the team navigates funding and scalability questions, their grassroots, farmer‑driven approach offers a replicable model for bringing affordable AI‑enabled solutions to rural communities across the state.
https://www.rockymounttelegram.com/news/local/researchers-use-artificial-intelligence-to-help-farmers/article_b269d86c-9487-4ed7-ae06-66c40082ca71.html