Arrhythmia Detection using Distributed Fuzzy Logic on ECG Signals

Accurate and scalable arrhythmia detection is essential for modern cardiac health monitoring systems, especially in the context of remote and cloud-integrated diagnostics. This paper proposes a modular, cloud-compatible arrhythmia detection system that combines feature-based fuzzy inference with distributed computing using Microsoft Azure. ECG data from the MIT-BIH Arrhythmia Database is used for training and evaluation. Key temporal features Mean RR Interval, Standard Deviation of RR Intervals, and QRS Complex Duration are extracted and processed through a fuzzy inference system to classify heartbeats into five categories: Normal, Supraventricular Arrhythmia, Ventricular Arrhythmia, Fusion Beat, and Unknown. To ensure scalability and high throughput, a simulated distributed processing framework was implemented using Python multithreading and Azure Blob Storage. The system achieved a classification accuracy of 92.1%, with precision, recall, and F1-score values above 90%. Additionally, a severity estimation module using fuzzy rules provides an interpretable risk score ranging from 0 to 10, offering nuanced insights beyond categorical classification. This architecture demonstrates potential for real-time deployment in wearable healthcare monitoring systems and large-scale telemedicine infrastructures.