A Low-Power Arrhythmia Detection SoC Based on Lightweight Convolutional Neural Network

Abstract

Cardiovascular diseases remain the leading global health threat, and arrhythmia is a critical early indicator of cardiac abnormalities. The demand for continuous cardiac monitoring has driven the development of portable, low-power arrhythmia detection systems. This paper presents a System-on-Chip (SoC) solution for arrhythmia detection based on a 1D convolutional neural network. The proposed method adopts a 1D Convolutional Neural Network (1D CNN) architecture, which enables effective feature extraction for arrhythmia classification while maintaining low computational complexity suitable for deployment on low-power edge devices. Furthermore, the chip has completed synthesis and backend design using TSMC 65nm process technology. Through algorithm-hardware co-optimization, this low-power arrhythmia detection system is of great significance for real-time and long-term cardiac health monitoring in wearable devices.