Unleashing Intelligence Over NTN to manage critcal Assets Integrity
summary
Claims an edge camera (IBEX) that runs cloud-grade CV with 500× power efficiency, continuous inference under 100 ms from a 10×10 cm solar panel, relays sensor metadata over LoRaWAN/NTN, and supports 1-year field operations by sending ultra-lightweight metadata and monthly model updates of a few kilobytes. Demonstrated image delivery by chopping a VGA image into ~200×200‑byte packets and achieving full image transfer via LoRa GEO NTN in under three minutes; MVP prototyping with first devices slated for Q1 next year.
AI-written summary · about
chapters (14)
- 0:00Deploying cloud-grade AI anywhere
- 1:09Cloud AI limitations for critical assets
- 2:37Montaigne’s scalable AI infrastructure
- 3:20Ibex smart camera efficiency
- 4:24Montaigne middleware for AI lifecycle
- 5:05Open-source Age for AI deployment
- 6:09Real-world fire detection use case
- 7:14Digital watchman’s connectivity
- 8:17Connectivity requirements optimization
- 9:19Testing with Dutch telecom satellite
- 10:27Antennas for satellite communication
- 11:54White fire detection system
- 13:40Why sky suffices for Montaigne
- 14:30Q&A
related talks
59mML and the IoT: Living on the Edge
30mAI On-board of Satellites, Towards Autonomous Scientific Instruments
16mEdge Computing and next-gen of IoT Sensors
19mHow to Deploy Artificial Intelligence in Industrial Application
28mMachine Learning on the Edge
26mBuilding a Cloud Native Edge Platform for Accelerating AI Inference
20mComputer Vision at the Edge with OpenVINO
29mEdge Intelligence Revolution