ML for Sensor Array
Budget: $25 – $50 USD
ICS-43434 microphone and BME688 (VOC/smoke) sensors have built in ML capabilities. I want a very primitive template designed whereby I could later send new model information to the devices via OTA. So more the architecture of the ML integrations within the hardware and software without the actual data modeling
What I'd like to do with the ML (at a high level):
• Use quantized models: Reduces size and speeds up inference.
• Limit audio logging: Only store short clips around detection events.
• Modular firmware: Separate audio, ML, and alert logic for maintainability.
• OTA capability: Enables remote updates of models and thresholds.
What I'd like to do with the ML (at a high level):
• Use quantized models: Reduces size and speeds up inference.
• Limit audio logging: Only store short clips around detection events.
• Modular firmware: Separate audio, ML, and alert logic for maintainability.
• OTA capability: Enables remote updates of models and thresholds.