Chest-Laminated and Sweat-Permeable E-Skin for Speech and Motion Artifact-Insensitive Cough Detection

Assessment of the cough severity is essential when dealing with respiratory diseases such as chronic obstructive pulmonary disease and COVID-19. Although a few wearable devices have been reported for cough detection, they mostly rely on microphones, accelerometers, or throat-fixed flexible sensors, which suffer from key issues including privacy disclosure and speech/motion artifacts. This study presents a chest-laminated electronic skin (e-skin) for reliable cough detection. Mixed dumbbell-like networks and through-holes are engineered on hard-to-stretch composite films for high stretching force sensitivity and sweat permeation, respectively. The e-skin can effectively reduce speech-signal and motion artifacts owing to firm adhesion and conformal contact with the chest even on sweaty skin. Experimental results show that the specificity for cough identification is as high as 99.75% through machine learning of automated acoustic analysis, even in the presence of hard-to-distinguish daily activities such as throat clearing. The developed chest-laminated e-skin is a simple, comfortable, yet reliable method to detect cough for the primary diagnosis of respiratory diseases by extracting subtle acoustic information from cough.