Collaborative Research: VINES: Track 2: AgSlicing: Predictable RAN and Spectrum Slicing for Precision Agriculture

Precision agriculture, which is essential to food security, critically depends on last-acre connectivity for real-time data exchange, edge-and cloud-based analytics, and Artificial Intelligence (AI) model updates. Yet today's wireless solutions cannot support the heterogeneous communication requirements of precision agriculture in a predictable manner, and they cannot enable affordable, predictable spectrum access required for pervasive deployment of quality wireless services. To address the challenges, this project will develop the AgSlicing system that enables predictable Radio Access Network (RAN) slicing and spectrum slicing as well as their integration with precision agriculture. Helping unleash the full potential of precision agriculture, AgSlicing unlocks significant economic benefits in the U.S. and globally, and it helps fight hunger, reduce fuel use and pollution, lower agriculture water use, and reduce chemical applications. This project is led by academic and industry leaders across the wireless and agriculture ecosystems. Through the project team's contributions to the NextG Alliance, O-RAN Alliance, and OnGo Alliance, its leadership at the O-RAN Alliance Open Testing and Integration Center ARA-OTIC, and its partners across the wireless and agriculture industries, this project will contribute to national as well as international policies and programs related to 6G, rural broadband, and agriculture. Tackling critical challenges in far-edge wireless, AgSlicing is also expected to have a transformative impact on other industries such as utilities, public safety, and defense. AgSlicing builds on algorithmic and systems breakthroughs that are already at Technology Readiness Level (TRL) 3-4 and that enable field-deployable solutions for predictable interference control, predictable control of per-packet communication reliability and timeliness, predictable communication capacity allocation, scalable massive Multiple Input Multiple Output (MIMO), accurate channel modeling, and reinforcement learning. The development, demonstration, and evaluation of AgSlicing will leverage the 272 square miles of farmland covered by the NSF Agriculture and Rural Communities (ARA) wireless living lab, and the project will also conduct commercial pilots in rural America (e.g., rural Montana). Through predictable RAN and spectrum slicing, AgSlicing supports the most demanding use cases of agriculture (e.g., sprayer teleoperation and monitoring) that require predictable low latency and high reliability, and it enables affordable, predictable access to spectrum. Its standards compliance, software-based implementation, and O-RAN support further facilitate AgSlicing's capability of tapping into growing radio performance while leveraging the cost-benefits of O-RAN. Besides the TRL-8 AgSlicing system, this project will generate first-of-its-kind real-world measurement data and models of agriculture wireless channels and communication systems. These data and models will serve as a foundation for studying agriculture wireless systems and are of significance by themselves. Curated AgSlicing experiments will also facilitate the reproducibility of AgSlicing solutions and inspire community participation in agriculture wireless research and innovation. This project will create exciting opportunities for broadening participation in computing and networking, and it will help enrich undergraduate and graduate research and education as well as K-12 outreach at Iowa State University and University of Virginia. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.