TheSleepGuard 💤

Smart Sleep Optimization Wristband

A wearable device designed to monitor and improve sleep quality through real-time heart rate analysis and sleep phase detection.

🎓 Project Context

• Period: August 2023 - December 2023
• Institution: ESME SUDRIA (France) in partnership with UDESC (Universidade do Estado de Santa Catarina, Brazil)
• Location: Joinville, Brazil - International academic exchange (6 months)
• Course: Digital Signal Processing / Embedded Systems
• Team: Tom Huyghe, Adèle Faillé
• Supervisor: Prof. Alexander Paterno

🎯 Objective

Develop a wristband using Digital Signal Processing (DSP) techniques to monitor and improve sleep quality by providing real-time feedback and personalized recommendations based on heart rate variability (HRV) analysis.

🔧 Hardware Components

• Microcontroller: Arduino-compatible board
• Heart Rate Sensor: MAX30105 (PPG - Photoplethysmography)
• Accelerometer: Motion detection for sleep analysis
• Display: OLED screen (128x64) for user interface
• Storage: SD card module for data logging
• Interface: Physical button for menu navigation
• Power: USB connection

⚙️ Key Features

1. Real-Time Monitoring

• Continuous heart rate measurement during sleep
• Data recording to SD card in CSV format (timestamp, BPM)
• Infrared threshold detection (>50,000) for finger presence

2. Sleep Phase Detection

Based on Heart Rate Variability (HRV) analysis:

Sleep Phase	BPM Range	Description
WAKE	> 73 BPM	Fully awake and alert
REM Sleep	58-73 BPM	Rapid eye movement, vivid dreaming
Light Sleep	45-58 BPM	Transitional stage, easy to wake
Deep Sleep	< 45 BPM	Restorative sleep (SWS)

3. Interactive User Interface

• Main Menu:
  • "Starting Sleep" → Begin recording session
  • "Starting Analysis" → Analyze recorded data
• Real-time display of sleep phase durations
• OLED screen with intuitive navigation

4. Signal Processing Pipeline

Data Collection → Preprocessing → Segmentation → Analysis → Visualization
       ↓              ↓              ↓            ↓            ↓
   Record AVG     FIR Filter    Sleep Phase   Button/App   MATLAB/Python
   Heart Rate     (0.1 Hz)      Classification  Display      Graphics

Data Preprocessing

• FFT Analysis: Fourier Transform to identify signal frequencies
• FIR Low-Pass Filter:
  • Order: 30
  • Cutoff frequency: 0.1 Hz
  • Designed in MATLAB, implemented in Arduino
  • Smooths heart rate variations
• Convolution: BeatAVG signal convolved with FIR filter

📊 Results

• Functional prototype with wristband form factor
• Accurate sleep cycle detection over 7-hour night sessions
• Sleep cycle visualization showing 3-5 cycles per night (90-minute cycles)
• Real-time analysis displayed on OLED screen
• Data export for advanced visualization in MATLAB/Python

Example Output

WAKE:  0:26:00
REM:   1:23:00
LIGHT: 3:45:00
DEEP:  1:52:00

💻 Tech Stack

Embedded Software

• Language: C++ for Arduino
• Libraries:
  • MAX30105.h - Heart rate sensor
  • U8g2lib.h - OLED display
  • SD.h - SD card storage
  • heartRate.h - Beat detection algorithm

Data Analysis

• MATLAB: Signal processing, filter design, visualization
• Python: Data analysis, FFT, convolution plots
• Libraries: NumPy, Pandas, Matplotlib, SciPy

Hardware

• Arduino IDE
• Breadboard prototyping
• Sensor integration (I2C, SPI protocols)

🧠 Skills Developed

• Embedded C++ programming for real-time systems
• Digital Signal Processing (FFT, FIR filters, convolution)
• Physiological signal analysis (HRV, PPG)
• Sensor integration and calibration
• User interface design for embedded systems
• Scientific data visualization
• International academic collaboration

📈 Methodology

1. Data Collection: Record average heart rate using MAX30105 sensor
2. Preprocessing: Apply FFT and FIR low-pass filter to smooth signal
3. Segmentation: Classify heart rate into sleep phases using threshold-based algorithm
4. Display & Reporting: Show results on OLED screen and export data
5. Modeling & Interpretation: Visualize sleep cycles in MATLAB with proper graphics

🚀 Future Improvements

• Integration of oximeter for SpO2 measurement
• Smart alarm feature based on real-time sleep phase analysis
• Personalized sleep recommendations based on historical data
• Wireless data transmission (Bluetooth/WiFi)
• Mobile app for enhanced data visualization

📁 Repository Structure

TheSleepGuard/
├── main.ino                    # Main Arduino sketch
├── MesureSleep.cpp/.h          # Heart rate measurement module
├── Analysing.cpp/.h            # Sleep phase analysis module
├── Python/
│   ├── FFT.py                  # Frequency analysis
│   ├── CONVO.py                # Convolution visualization
│   └── untitled0.py            # Data plotting
├── MATLAB/
│   ├── FFTMAT.m                # FFT and filter design
│   └── codeProf24oct.m         # Sleep cycle analysis
└── docs/
    └── presentation.pdf        # Project presentation

📝 References

• ELVEE Pulse sensor documentation
• Heart Rate Variability (HRV) research papers
• Sleep cycle studies (90-minute cycles)
• MAX30105 sensor datasheet

🤝 Acknowledgments

Special thanks to Prof. Alexander Paterno (UDESC) for supervision and guidance, and to ESME SUDRIA and UDESC for facilitating this international academic partnership.

---

Project Status: ✅ Completed (December 2023)
Academic Grade: Validated - M1 Embedded Systems Course

---

Developed during international academic exchange at UDESC, Joinville, Brazil