Agentic AI for Rural Connectivity and Spectrum-Aware Networks to Power Smart Agriculture Ecosystems
Anil B. Desai, Bhaksara Rallabandi, Krishna Kumar Gattupalli, Monish Sai Medarametla
Abstract
The use of smart agriculture requires reliable, low-latency and energy-efficient connections needed to support real-time decision-making in farms. Nevertheless, conventional rural wireless networks usually have difficulty about scarcity of spectrum, unreliable backhaul and a low level of edge intelligence. In the current research, the author suggests agentic, spectrum-conscious architecture, able to combine IoT sensor, lightweight edge agents, cooperative spectrum management as well as multi-layer orchestration to facilitate the connectivity of smart farming ecosystems. Based on real-world inspired datasets of specimen smart farming logs and data sets from the IoT sensors, the proposed system was compared to a baseline network. There are major improvements in the results of experiments, end-to-end latency was cut in half, going down to 145 ms, delivering packets ratio rose to 97.8, and energy per message decreased by 41. Moreover, the use of spectrum was at 73.5% and SLA at 92% with more than 90% accuracy in the edge. The results support that agentic architectures that are spectrum-aware provide scalable, resilient, and energy-aware connectivity to provide data-driven agricultural automation and support decision-making in rural areas.