Skip to content

UC3 — IoT / Edge

Lead: J. Chen  ·  WBS: 5.0  ·  DOI deliverable: Month 18

Story

Precision-agriculture sensor networks generate continuous streams of high-frequency telemetry — soil moisture, leaf temperature, weather, drone imagery. Most farms today either drop the data or batch-upload it; the result is that anomalies (irrigation failures, pest outbreaks, equipment faults) are detected hours or days late.

UC3 demonstrates MESA running an agentic triage loop at the edge: 5G/6G- connected sensors stream into a partner-farm iRODS zone, the Lakehouse catalog ingests metadata in near-real-time, and an agentic workflow flags anomalies for the farm operator.

What the prototype demonstrates

  1. Lakehouse ingest of high-frequency time-series data with time-travel for after-the-fact root-cause analysis.
  2. Federated authorization so the farm owner controls who sees what, even when data crosses institutional boundaries.
  3. Edge-aware MCP — MCP servers running on 5G/6G edge nodes so the triage loop tolerates partial network partitions.

Personnel

  • J. Chen — Faculty Senior Personnel, UC3 lead.
  • GRA 4 (ECE, 5G/6G Edge) — increased from 0.25 FTE to 0.50 FTE to provide dedicated research support on this use case.

Deliverables (Month 18)

  • Reproducible end-to-end workflow with DOI.
  • Open-source edge MCP runtime.
  • Field-test report from the partner farm.

Status

Draft — content matures through Phase 2.