NeoWear: An IoT-connected e-textile wearable for neonatal medical monitoring

Abstract

According to WHO, 15 millions babies are born preterm each year globally. Preterm infant (born before 37 weeks of gestation) are at a significantly higher risk of medical and surgical morbidity in comparison to babies born at term (around 37 weeks). Innovative solutions are warranted to meet the increased requirements of Neonatal Intensive Care Unit (NICU) with rising number of preterm babies. Various kinds of vital signs such as heart rate (HR), respiration rate (RR) or blood oxygen level (SpO2) are monitored in NICU. Considering the fact the lungs develop in the last few weeks of gestation, preterm babies in NICU demands sophisticated technology to monitor respiration and events related to the respiration. Current technologies rely on the indirect measurements from thoracic impedance or other invasive techniques for RR monitoring. This poses discomfort and risk of infections to babies. Also, the delivery of parental and clinical care is largely impacted by a large number of monitoring cables placed on the babies. To address this requirements, we have developed an Internet-of-Things (IoT) based smart textile chest belt called “NeoWear” to monitor RR and detect apnea events in babies. The NeoWear is a wearable system consisting of a sensor belt, a wearable embedded system, and an edge computing device. The sensor belt comprised of a pressure sensors made of smart-textile and an Inertial Measurement Unit (IMU) to monitor movements. These sensors are connected to a micro-controller equipped with wireless communication capabilities called as a wireless embedded system (WES). The WES wirelessly connects with an edge computing device (ECD) using an MQTT-based IoT networking architecture. ECD is capable to offer signal processing and computing services to detect RR and apnea events. Simulation experiments using a high-fidelity, programmable NICU baby mannequin and five healthy adults were conducted to test the efficacy of the NeoWear System. Our findings shows an average error of 0.89 BrPM in respiration rate measurement and ∼97 percent accuracy in apnea detection on baby mannequin. Our experiments also demonstrated that how the movements of the baby mannequin affected the respiration signal during apnea episodes and when breathing rate increased to 40 breaths per minute. In addition, the changes on human respiration data showed a meaningful increase for slow, normal and fast breathing. We also computed computation and communication latencies and they were found to be ∼66 and 22 ms, respectively. Our preliminary results are promising showing the efficacy of NeoWear to measure respiration and related events for babies.