Intelligent Edge Computing Framework for ECG Signal Analysis in IoMT Healthcare Systems

Abstract

Wearable electrocardiogram (ECG) sensors enable intelligent edge computer convergence and provide a paradigm shift in real-time, privacy alluring, and energy efficient cardiac health monitoring systems.Nonetheless, the current Internet of Medical Things (IoMT) designs are limited to cloud-related dependency, low latency, and unsupported hardware testing.The currently discussed paper introduces a completely built edge-native IoMT architecture of edge-based real-time ECG anomaly detection optimized to run on inexpensive embedded hardware.The suggested solution encompasses a lightweight convolutional neural network (CNN) to perform on-device arrhythmia focused classification on the Raspberry Pi 4, a Raspberry Pi 4-based edge computing and inference platform, and a lean data transfer plan to lower the bandwidth usage and improve privacy.The framework combines sensitiveness, edge perception, and energy-conscience functioning unlike the past simulation based methodologies.This was experimented in real-life research using a large variety of open ECG.The findings indicate great changes in accuracy, latency, and power consumption.They also demonstrate the absence of the need in cloud infrastructure and that they will have the clinical insights available when they require them.This work enables the following generation of secure, sustainable and safe IoMT cardiac monitoring systems through linking the latest deep learning with devices that can be deployed.