Electret-based flexible pressure sensor for respiratory diseases auxiliary diagnosis system using machine learning technique

Abstract

Respiratory diseases have been increasingly affecting people worldwide, posing a major public health challenge due to rising morbidity and mortality rates. Subtle abnormal respiratory sounds s may present early in the pulmonary or respiratory tract diseases, therefore urging an instant intervention in critical clinical conditions. However, the current monitoring and signal analysis technologies pose significant challenges in achieving real-time, convenient, and accurate respiratory disease monitoring. Here, we propose a novel automatic auxiliary diagnosis system that utilizes a flexible electret-based self-powered sensor (FESS), signal processing technology, and machine learning algorithms. The FESS is based on a strain-enhanced laminated electret with an edge-to-edge hollow hemisphere array structure, which enables the system to detect and prevent various respiratory diseases with high accuracy rates of 99.43%, sensitivity of 98.30%, and specificity of 99.02%. Our system holds immense potential in reducing medical burden and improving the overall health of individuals.