BioWake is a high-performance ESP32-S3 firmware designed for real-time physiological data logging and autonomous sleep-cycle alarm triggering using Edge AI.

The system uses a 100Hz dual-core RTOS architecture to decouple high-frequency sensor sampling from Edge Impulse ML inference. By detecting specific sleep stages (REM/Awake) locally on the device, BioWake can autonomously trigger a wake-up sequence on a paired iPhone via Bluetooth Low Energy (BLE).

• Controller: ESP32-S3-WROOM-1
• Flash: 16MB QIO 80MHz
• PSRAM: 8MB OPI (Used for 112KB asynchronous circular buffer)
• Sensors:
  • MAX30102: PPG (Red/IR) @ 400Hz (Averaged to 100Hz)
  • MPU6050: 6-Axis IMU (Accel/Gyro) @ 100Hz
  • GSR: 12-bit Galvanic Skin Resistance

• Core 1: Dedicated to strict 100Hz sensor polling and 12.5Hz AI downsampling.
• Core 0: Handles SD Card I/O, BLE Keyboard HID service, and Edge Impulse Inference (TFLite Micro).
• Asynchronous Logging: A 4000-sample PSRAM buffer prevents data loss during SD card write latency spikes.
• Closed Loop: Syncs with NTP atomic time on boot, then transitions to BLE HID mode to act as an iPhone media controller.

1. Clock Sync: Connects to WiFi at boot to sync UTC-4 time, then disables WiFi to prevent BLE interference.
2. Window Check: Monitors if the current time is within the user-defined WAKE_WINDOW (e.g., 6:30 AM - 7:00 AM).
3. AI Verification: If the Edge Impulse model predicts REM or Awake_Light with >70% confidence, the ESP32 sends a KEY_MEDIA_PLAY_PAUSE command to the iPhone.
4. Action: Apple Music/Spotify resumes instantly to wake the user during their lightest sleep stage.

1. PlatformIO: Open project in VS Code with PlatformIO extension.
2. Pairing: Boot the device and pair your iPhone to "BioWake" in Bluetooth Settings.
3. Monitor: Use pio device monitor --baud 115200 to view the real-time HUD.