Deep Learning Assisted Body Area Triboelectric Hydrogel Sensor Network for Infant Care

摘要

Infants are physically vulnerable and cannot express their feelings. Continuousmonitoring and measuring the biomechanical pressure to which aninfant body is exposed remains critical to avoid infant injury and illness. Here,a body area sensor network comprising edible triboelectric hydrogel sensorsfor all-around infant motion monitoring is reported. Each soft sensor holdsa collection of compelling features of high signal-to-noise ratio of 23.1 dB,high sensitivity of 0.28 V kPa~(?1), and fast response time of 50 ms. With theassistance of deep learning algorithms, the body area sensor network canrealize infant motion pattern identification and recognition with classificationaccuracy as high as 100%. Additionally, a customized user-friendly cellphoneapplication is developed to provide real-time warning and one-click guardianinteraction. This self-powered body area sensor network system providesa promising paradigm for reliable infant care in the era of the Internet ofThings.