← Back to Home

ECG Monitoring System

Real-Time BPM and Atrial Fibrillation Detection

Details

Project Type: Biomedical Instrumentation Design
Course: ECE 50600 - Biomedical Instrumentation Design
Status: Completed
Focus Area: Cardiac Arrhythmia Detection & Real-Time Monitoring
Key Achievement: Real-time ECG monitoring with automated AF detection
Collaborators: Development Team: Al Shafayet Haque Silvy, Ahmad Shadman, Fairuz Haq, Advisor: Dr. Bin Chen

Technologies & Hardware

ECG Sensor: AD8232 Heart Rate Monitor
Microcontroller: Arduino Nano
Signal Processing: MATLAB with Arduino interface
Deep Learning: VGG16 CNN with Transfer Learning
Electrodes: Disposable Ag/AgCl gel electrodes
Dataset: MIT-BIH Arrhythmia Database (PhysioNet)

Problem Statement & Motivation

The modern medical field, especially for the treatment of heart disease, relies heavily on electrocardiogram (ECG) equipment to detect abnormalities and illnesses. An electrocardiogram records the electrical signals in the heart through a painless examination for detecting heart problems and monitoring heart health. Atrial fibrillation (AFib or AF) is defined as an irregular and often very rapid heart rhythm that can lead to blood clots in the heart, increasing the risk of stroke, heart failure, and other heart-related complications.

Critical Healthcare Challenges:

  • Detection Speed: Quick and efficient ECG measurement is crucial for diagnosing life-threatening conditions like atrial fibrillation
  • Accessibility: Traditional ECG monitoring requires hospital visits and expensive equipment, limiting access for routine monitoring
  • Early Detection: Atrial fibrillation can be asymptomatic, making continuous monitoring essential for early intervention
  • Cost Barrier: Professional ECG monitoring is expensive and not feasible for continuous home-based health management
  • Real-Time Analysis: Immediate BPM calculation and arrhythmia detection enable faster medical decision-making
  • Data Interpretation: ECG signals require expert analysis, creating a bottleneck in diagnosis

Understanding ECG Components:

Each ECG cycle comprises five waves (P, Q, R, S, and T) corresponding to different periods of the heart's activity:

  • P Wave: Represents depolarization of the atria
  • QRS Complex: Represents depolarization of the ventricles (R-peak is highest amplitude)
  • T Wave: Represents repolarization of the ventricles

This project addresses these challenges by developing a cost-effective, portable ECG monitoring system using the AD8232 sensor and Arduino Nano microcontroller, coupled with MATLAB signal processing and deep learning validation for automated cardiac condition assessment.

Our Solution: Real-Time ECG Monitoring System

This project delivers a complete ECG monitoring system that empowers patients to monitor their heart rate and assess cardiac conditions conveniently at home or on the go. The system combines hardware signal acquisition with sophisticated software processing and machine learning validation.

System Architecture

1. Hardware Implementation

ECG Sensor Module: AD8232 single-lead heart rate monitor for signal conditioning

Microcontroller: Arduino Nano for analog-to-digital conversion and serial communication

Electrode Configuration: Three disposable Ag/AgCl gel electrodes placed on:

• Right Arm (RA) - Red electrode

• Left Arm (LA) - Yellow electrode

• Right Leg (RL) - Green electrode (ground reference)

Breadboard Circuit: Stable connections for AD8232 to Arduino Nano with proper grounding

Power Supply: USB-powered through Arduino Nano (5V logic)

2. Signal Acquisition & Arduino Processing

Arduino IDE Implementation: Custom code for reading analog input from AD8232 output pin

Sampling Rate: Continuous analog readings at high frequency for accurate ECG capture

Serial Communication: 9600 baud rate for transmitting ECG data to computer

Lead-Off Detection: LO+ and LO- pins monitor electrode contact quality

Real-Time Visualization: Arduino IDE Serial Plotter displays live ECG waveform

Limitations Addressed: Serial Plotter only shows latest 500 values, necessitating MATLAB integration for comprehensive analysis

3. MATLAB Interface & Signal Processing

Hardware Support Package: MATLAB Support Package for Arduino Hardware enables direct communication

Real-Time Data Acquisition: readVoltage() function captures analog values from A0 pin

Signal Normalization: Data normalized between -1 and +1 using min-max normalization:

Formula: X_normalized = 2*(X - X_min)/(X_max - X_min) - 1

Continuous Plotting: 1000 iterations (approximately 60 seconds) for comprehensive heart rate analysis

Dynamic Visualization: Real-time plot updates with detected R-peaks highlighted in red

4. R-Peak Detection Algorithm

Peak Finding: MATLAB's findpeaks() function identifies local maxima in normalized signal

Threshold Setting: MinPeakHeight = 0.5 filters noise and isolates R-peaks

Peak Storage: Arrays maintain all detected peak values and time indices

Visual Marking: Red asterisks (*) overlay detected R-peaks on ECG plot

Adaptive Detection: Minimum 3-sample window ensures reliable peak identification

5. R-R Interval Calculation & Calibration

Oscilloscope Verification: Parallel oscilloscope connection validates MATLAB timing

Latency Compensation: Accounts for signal transmission delay through hardware chain

Scaling Factor Derivation:

• Oscilloscope measurement: R-R interval = 690ms (actual time)

• MATLAB plot: R-R interval = 32 units (x-axis difference)

• Scaling factor = 690ms / 32 = 21.5625 ms per unit

R-R Interval Array: Calculates absolute differences between consecutive peak indices

Mean R-R Interval: Average of all R-R intervals provides stable heart rate estimate

6. BPM Calculation & AF Detection

BPM Formula: BPM = 60,000 / (Mean R-R interval × Scaling Factor)

Classification Logic:

• Normal: 60-100 BPM with regular rhythm

• Atrial Fibrillation: 101-200 BPM indicating irregular heart rhythm

• Error: Values outside ranges prompt re-measurement

Real-Time Feedback: Immediate display of BPM and cardiac condition assessment

Mean Absolute Difference Approach: Robust to signal noise from 3-electrode placement

System Advantages:

  • Cost-Effective: Uses consumer-grade hardware (Arduino Nano ~$5, AD8232 ~$10)
  • Portable: Compact design enables home monitoring and emergency use
  • Real-Time Analysis: Immediate BPM and AF detection without manual calculation
  • Accurate Calibration: Oscilloscope validation ensures clinical-grade timing precision
  • User-Friendly: Visual ECG display with clearly marked R-peaks and automated diagnosis
  • Extensible: MATLAB environment allows additional signal processing features

Algorithm Validation with Real-World Datasets

To verify the robustness and accuracy of our R-peak detection and BPM calculation algorithm, we tested it on established medical datasets from PhysioNet's PhysioBank ATM database.

Dataset Testing Methodology:

Dataset 1: Normal Sinus Rhythm (NSR)

Source: MIT-BIH Normal Sinus Rhythm Database

Signal Characteristics: Regular heartbeat pattern with consistent R-R intervals

Processing: CSV data imported into MATLAB, processed with identical peak detection algorithm

Results: BPM = 89, Classification: Normal

Validation: Algorithm correctly identified regular rhythm and appropriate BPM range

Dataset 2: Atrial Fibrillation (AF)

Source: MIT-BIH Arrhythmia Database

Signal Characteristics: Irregular rhythm with variable R-R intervals, elevated heart rate

Processing: Applied same findpeaks() threshold and BPM calculation formula

Results: BPM = 150, Classification: Atrial Fibrillation

Validation: Successfully detected AF condition based on elevated BPM and irregular intervals

Key Validation Findings:

  • Algorithm Robustness: Consistent performance across real-time hardware data and stored CSV datasets
  • Accurate Classification: 100% correct classification on tested NSR and AF samples
  • Threshold Effectiveness: MinPeakHeight = 0.5 reliably isolates R-peaks across different signal qualities
  • Clinical Alignment: BPM values match expected ranges for normal and arrhythmic conditions
  • Generalizability: Algorithm works on diverse ECG morphologies from different patients

Deep Learning Validation: VGG16 Transfer Learning

To provide additional validation and enable automated arrhythmia classification, we implemented a Convolutional Neural Network using transfer learning with VGG16 architecture.

Dataset Preparation:

Data Source & Processing

Database: MIT-BIH combined dataset including:

• MIT-BIH Arrhythmia Database

• MIT-BIH Normal Sinus Rhythm Database

• BIDMC Congestive Heart Failure Database

Data Structure: 162 × 65,536 matrix (162 patients, 65,536 samples each)

Sampling Rate: 128 Hz, 1-hour duration per recording

Classes Selected: ARR (Arrhythmia) and NSR (Normal Sinus Rhythm)

Subset Creation: 30 ARR + 30 NSR = 60 total recordings for focused binary classification

ECG to Image Conversion Pipeline

Segmentation: Each recording divided into 500-point segments

Total Segments: 60 recordings × 125 segments/recording = 7,500 segments

Continuous Wavelet Transform (CWT):

• Applies time-frequency analysis to capture dynamic frequency characteristics

• Reveals frequency components at different time points

• Ideal for signals with time-varying spectral content like ECG

Image Generation: CWT coefficients converted to 256-color RGB images (224×224 pixels)

Class Distribution: 3,750 ARR images + 3,750 NSR images = 7,500 total images

Dataset Split Strategy

Training Set: 5,250 images (70%)

Validation Set: 2,206 images (29%)

Test Set: 44 images (1%)

Balance: Equal representation of ARR and NSR in each split

VGG16 Architecture & Transfer Learning

Base Model: VGG16 pre-trained on ImageNet

Model Configuration:

• Input shape: (224, 224, 3) RGB images

• Include_top = False (remove original classifier)

• Weights = "imagenet" (transfer learning)

• Base_model.trainable = False (freeze convolutional layers)

Custom Classifier Layers:

1. Flatten layer: (7, 7, 512) → 25,088 features

2. Dense layer: 256 neurons with ReLU activation

3. Dropout layer: 0.5 dropout rate for regularization

4. Output layer: 1 neuron with sigmoid activation (binary classification)

Total Parameters: 14,714,688 (56.13 MB)

Training Configuration

Loss Function: Binary crossentropy

Optimizer: RMSprop

Metrics: Accuracy

Training Epochs: Tested at 150, 300, and 1000 epochs

Feature Extraction: Pre-computed features from frozen VGG16 base

CNN Model Performance:

~94%
Training Accuracy (150 epochs)
~93.7%
Validation Accuracy (150 epochs)
7,500
Training Images
100%
Test AF Classification

Training Analysis Across Epochs:

  • 150 Epochs: Achieved ~92% accuracy on both training and validation, demonstrating excellent convergence
  • 300 Epochs: Maintained ~93.75% accuracy with stable validation performance, no overfitting observed
  • 1000 Epochs: Consistent ~94% accuracy, validating model robustness and preventing degradation
  • Generalization: Minimal gap between training and validation accuracy indicates good generalization

External Data Validation:

  • Independent AF Sample: Tested on external atrial fibrillation ECG not in training set
  • Classification Result: Correctly classified as Arrhythmia (ARR)
  • Confirmation: CNN model validates our MATLAB algorithm's AF detection capabilities

Technical Innovations:

  • CWT Image Encoding: Converts temporal ECG signals into visual representations capturing time-frequency dynamics
  • Transfer Learning Efficiency: Leverages ImageNet pre-training for medical image classification
  • Frozen Feature Extraction: Trains only classifier layers, reducing computational requirements
  • Dual Validation: Combines rule-based BPM algorithm with deep learning classification

Comprehensive Results & Performance

Real-Time Hardware Testing:

77 BPM
Measured Heart Rate
Normal
Cardiac Condition
21.56 ms
Calibrated Scaling Factor
60 sec
Measurement Duration

Dataset Validation Results:

  • Dataset 1 (Normal): BPM = 89, Classification: Normal ✓
  • Dataset 2 (Arrhythmia): BPM = 150, Classification: Atrial Fibrillation ✓
  • Algorithm Accuracy: 100% correct classification on tested samples
  • Consistency: Identical peak detection performance on real-time and stored data

CNN Model Performance Summary:

  • Training Accuracy: ~94% across all epoch configurations
  • Validation Accuracy: ~93.75% with minimal overfitting
  • External Test: Successfully classified independent AF sample
  • Convergence Speed: Stable performance achieved within 150 epochs
  • Model Robustness: Consistent accuracy maintained through 1000 epochs

Key Performance Indicators:

  • R-Peak Detection: Reliable identification with 0.5 threshold across diverse signal qualities
  • BPM Accuracy: Validated against oscilloscope measurements (21.5625 ms/unit scaling)
  • AF Discrimination: Clear separation between normal (60-100 BPM) and AF (101-200 BPM)
  • Processing Speed: Real-time analysis with ~60 second measurement window
  • Signal Quality: Successfully handles 3-electrode configuration challenges

Impact & Clinical Applications

Healthcare Accessibility

  • Home Monitoring: Enables patients to perform routine ECG checks without hospital visits
  • Cost Reduction: Low-cost hardware (~$15 total) makes cardiac monitoring accessible to underserved populations
  • Early Detection: Regular monitoring facilitates early identification of atrial fibrillation before symptoms appear
  • Emergency Response: Real-time BPM calculation aids in critical decision-making during cardiac events
  • Chronic Condition Management: Patients with known AF can track heart rhythm patterns over time

Clinical Applications

  • Primary Care Screening: Quick ECG assessment during routine checkups
  • Post-Surgery Monitoring: Continuous cardiac surveillance after surgical procedures
  • Medication Adjustment: Real-time feedback on anti-arrhythmic medication effectiveness
  • Telemedicine Integration: Remote patient monitoring with data transmission to healthcare providers
  • Emergency Medical Services: Portable ECG for paramedics and first responders
  • Sports Medicine: Athlete heart rate monitoring during training and competition

Public Health Impact

  • Stroke Prevention: Early AF detection reduces stroke risk by enabling timely anticoagulation therapy
  • Reduced Healthcare Burden: Home monitoring decreases unnecessary hospital visits and emergency room utilization
  • Health Literacy: Empowers patients with real-time cardiac health data and education
  • Preventive Care: Shifts healthcare model from reactive treatment to proactive monitoring

Educational Value

  • Biomedical Engineering Training: Practical experience with medical device design and signal processing
  • Clinical Correlation: Bridges theoretical ECG knowledge with real-world implementation
  • Interdisciplinary Learning: Combines hardware, software, signal processing, and machine learning
  • Research Platform: Foundation for advanced cardiac monitoring research projects

Economic Benefits

  • Low Manufacturing Cost: Arduino Nano ($5) + AD8232 ($10) + electrodes ($5/pack) = $20 total
  • Reduced Medical Expenses: Decreases need for expensive Holter monitors and clinic visits
  • Scalability: Mass production potential for developing countries and rural healthcare
  • Insurance Savings: Preventive monitoring reduces costly emergency interventions

Technical Innovations

  • Oscilloscope Calibration: Novel approach to compensate for signal processing latency
  • Mean Absolute Difference: Robust BPM calculation method for 3-electrode configurations
  • CWT Image Encoding: Transforms temporal ECG signals into visual representations for CNN processing
  • Dual Validation Framework: Combines rule-based algorithm with deep learning for enhanced reliability
  • MATLAB-Arduino Integration: Seamless hardware-software interface for real-time analysis

Complete System Implementation

Hardware Components

  • AD8232 ECG Sensor: Single-lead heart rate monitor with integrated signal conditioning
  • Arduino Nano: ATmega328P microcontroller (16 MHz, 32KB Flash, 2KB SRAM)
  • Breadboard: Solderless prototyping platform for circuit assembly
  • Ag/AgCl Electrodes: Medical-grade disposable electrodes with adhesive backing
  • Electrode Cables: 3.5mm snap connectors for secure electrode attachment
  • USB Cable: Power supply and serial communication to computer

Software Stack

  • Arduino IDE: C/C++ programming environment for microcontroller code
  • MATLAB R2023a: Signal processing and data visualization platform
  • MATLAB Support Package for Arduino: Hardware interface library for direct Arduino control
  • Python 3.10: CNN model implementation environment
  • TensorFlow 2.x: Deep learning framework for VGG16 transfer learning
  • Keras API: High-level neural network interface

Data Processing Pipeline

Step 1: Signal Acquisition

1. Patient electrodes placed on both arms and right leg

2. AD8232 conditions raw bioelectric signals (filtering, amplification)

3. Arduino Nano samples analog output at A0 pin

4. Serial transmission at 9600 baud to computer

Step 2: MATLAB Processing

1. Arduino object creation and port connection

2. readVoltage(a, "A0") captures analog values in real-time

3. Min-max normalization scales data to [-1, 1] range

4. Real-time plotting with 1000-iteration loop (60 seconds)

Step 3: Peak Detection

1. findpeaks() identifies local maxima with 0.5 threshold

2. Peak values and indices stored in arrays

3. Red asterisks overlay detected R-peaks on plot

4. Visual confirmation of detection accuracy

Step 4: BPM Calculation

1. Calculate R-R intervals (time between consecutive peaks)

2. Apply oscilloscope-calibrated scaling factor (21.5625 ms/unit)

3. Compute mean R-R interval for stability

4. BPM = 60,000 / (mean_RR × scaling_factor)

Step 5: Classification

1. Compare calculated BPM against thresholds

2. Normal: 60-100 BPM

3. Atrial Fibrillation: 101-200 BPM

4. Display classification result to user

System Requirements

  • Computer: Windows/Mac/Linux with USB port
  • MATLAB: Version R2020a or later with Signal Processing Toolbox
  • Arduino IDE: Version 1.8.x or later
  • Python (Optional): For CNN validation (3.8+)
  • Memory: Minimum 4GB RAM for MATLAB processing
  • Storage: 100MB for software and data storage

Setup & Calibration

  • Electrode Preparation: Clean skin with alcohol wipe, ensure proper adhesion
  • Lead-Off Detection: Verify LO+ and LO- pins for electrode contact quality
  • Signal Stability: Allow 30 seconds for signal stabilization after electrode placement
  • Threshold Adjustment: Modify MinPeakHeight if needed based on signal amplitude
  • Scaling Factor Validation: Periodic oscilloscope verification ensures timing accuracy

Future Enhancements & Research Directions

Hardware Improvements

  • 12-Lead ECG: Expand to full 12-lead system for comprehensive cardiac assessment
  • Wireless Transmission: Integrate Bluetooth module for cable-free monitoring
  • Battery Operation: Add rechargeable battery for true portability
  • LCD Display: Onboard screen for standalone operation without computer
  • SD Card Storage: Local data logging for extended monitoring periods
  • Improved Electrodes: Investigate dry electrodes for long-term comfort

Software Enhancements

  • Multi-Arrhythmia Classification: Expand CNN to detect additional cardiac conditions (PVC, SVT, etc.)
  • Real-Time CNN Inference: Deploy lightweight model for on-device classification
  • Mobile Application: Develop smartphone app for data visualization and alerts
  • Cloud Integration: Implement secure data upload for remote physician access
  • Trend Analysis: Add long-term heart rate variability and pattern tracking
  • Alert System: Automatic notifications for detected arrhythmias

Clinical Validation

  • IRB-Approved Studies: Conduct formal clinical trials with patient cohorts
  • Cardiologist Validation: Expert review of automated classifications against gold standard
  • Sensitivity/Specificity Analysis: Quantify false positive and false negative rates
  • Long-Term Monitoring: Validate system performance over weeks/months of continuous use
  • Diverse Population Testing: Assess performance across age groups, comorbidities, and demographics

Machine Learning Advancements

  • Attention Mechanisms: Implement transformer-based models for temporal ECG analysis
  • Multi-Task Learning: Simultaneously predict multiple cardiac parameters (BPM, QT interval, ST segment)
  • Few-Shot Learning: Enable model to learn rare arrhythmias from limited examples
  • Explainable AI: Develop interpretability methods to show which ECG features drive classifications
  • Ensemble Methods: Combine multiple models for improved robustness

Regulatory & Commercialization

  • FDA Clearance: Pursue 510(k) approval for medical device classification
  • HIPAA Compliance: Implement encryption and secure data handling practices
  • Clinical Guidelines: Develop user manual and clinical interpretation guidelines
  • Manufacturing Scale-Up: Design injection-molded enclosure for production units
  • Quality Management: Establish ISO 13485 quality system for medical devices

Research Extensions

  • Heart Rate Variability: Implement advanced HRV analysis for autonomic function assessment
  • QT Interval Measurement: Add automated QT/QTc calculation for drug safety monitoring
  • ST-Segment Analysis: Detect ischemic changes for myocardial infarction screening
  • Sleep Study Integration: Combine with pulse oximetry for sleep apnea detection
  • Fitness Applications: Adapt for exercise stress testing and sports performance