SnoreSense is a TinyML project that detects snoring sounds in real time using the Seeed XIAO MG24 Sense development board.
It uses an onboard microphone, a TensorFlow Lite (TFLite) model, and runs completely offline — enabling low-power, privacy-safe sleep monitoring.
Snoring is not just noisy — it can be a symptom of sleep disorders like sleep apnea.
Conventional snore monitors are often expensive, power-hungry, or require cloud connectivity.
SnoreSense aims to solve this by:
- Detecting snores locally on a microcontroller
- Working offline to preserve privacy
- Using affordable open-source hardware and ML tools
- The XIAO MG24 Sense continuously listens through its onboard PDM microphone.
- Audio samples are processed in small frames and passed through a TinyML classifier trained to detect snore sounds.
- When a snore is detected, the system can:
- Light up an LED
- Log the event to memory
- Or send data via BLE (future update)
All this happens entirely on-device — no cloud required!
| Component | Description |
|---|---|
| 🟩 Seeed XIAO MG24 Sense | Main MCU + onboard PDM microphone |
| 🔵 USB-C Cable | For programming and power |
| 🔴 Optional: LED/Buzzer | To indicate snore detection |
| 🪫 Optional: LiPo Battery | For portable operation |