Sniffing the Metaverse
summary
An electronic nose is built with a ~ $30 metal‑oxide gas sensor and a microcontroller running tinyML inference in ~1 ms, transforms raw VOC traces via a DSP into features, and emits only high‑level insights into an Azure IoT graph and time‑series store. The pipeline uses Digital Twins/ontologies and Spatial Anchors for location and querying. Code and hardware are on GitHub, and a real project by a 13‑year‑old detects fungal pneumonia, showing the idea works beyond the coffee demo.
AI-written summary · about
chapters (11)
- 0:00Intro
- 1:04METAVERSE?
- 4:38CONCRETE APPLICATIONS
- 6:00TECHNOLOGIES THAT ENABLE METAVERSE
- 8:04TECHNOLOGIES ENABLING THE METAVERSE
- 10:22MODELING THE ENVIRONMENT
- 14:13FEEDING THE METAVERSE
- 20:01BUT HOW CAN A MACHINE SMELL ANYWAYS?
- 21:20AI MODEL FOR SMELLS?
- 25:36SPATIAL ANCHORS
- 31:29IN A NUTSHELL
related talks
39mFrom a Weekend Hack, to a Future-Proof IoT Device
19mAI, Technology & Health: What's Smell Got to Do With It?
39mFrom a Weekend Hack, to a Future-Proof IoT Device
34mInferencing Big Data with AI and Machine Learning in Metaverse
59mML and the IoT: Living on the Edge
31mJennifer Marsman, Principal Developer Evangelist, Microsoft
47mReal-Time AI with Open Source
01h 30mIt's in the Air Tonight. Sensor Data in RAG