A Low-Power Wireless System for Predicting Early Signs of Sudden Cardiac Arrest Incorporating an Optimized CNN Model...

Area of Science:

• Biomedical Engineering
• Cardiovascular Health Technology
• Artificial Intelligence in Healthcare

Background:

• Sudden cardiac arrest (SCA) has a low survival rate, with inadequate early warning systems for at-risk patients.
• Continuous electrocardiogram (ECG) monitoring outside clinical settings is crucial for early detection of cardiac dysfunction.
• Existing long-term monitoring solutions face challenges with battery life and power consumption during data transmission.

Purpose of the Study:

• To develop an integrated system for continuous, out-of-hospital ECG monitoring.
• To enable early detection and prediction of sudden cardiac arrest through advanced signal processing.
• To address the limitations of power consumption and battery life in wearable cardiac monitoring devices.

Main Methods:

• Utilized a zeolite-based dry electrode for safe, on-skin ECG acquisition.
• Implemented a low-power readout circuit with a 10-bit ADC and nRF24L01 transceivers for wireless data transmission.
• Developed and deployed a subject-wise cross-validated, three-fold optimized convolutional neural network (CNN) on NVIDIA Jetson for irregular heartbeat detection.

Main Results:

• The wireless transmission system operated at 2.4 GHz ISM band with GFSK modulation, achieving a 250 kbps data rate, low power consumption (11.2 mW transmit, 4.62 mW receive), and a bit error rate ≤0.1%.
• The CNN model achieved 89% accuracy in identifying irregular heartbeats within 5.31 seconds.
• Normal beat classification demonstrated high performance with an F1 score of 0.94 and an ROC score of 0.886.

Conclusions:

• The integrated system successfully combines ECG acquisition, low-power wireless transmission, and AI-driven analysis for effective irregular heartbeat detection.
• This technology offers a promising solution for continuous, out-of-hospital cardiac monitoring, potentially improving outcomes for patients at risk of SCA.
• The system's low power consumption and high accuracy pave the way for practical, long-term wearable cardiac health solutions.