Wearable Smart Sensor System for Monitoring and Intelligent Prediction of Sodium Ions in Human Perspiration

To date, many studies have been carried out for monitoring analytes in human perspiration with wearable sweat sensors, but few of them have done an in-depth investigation on the relationship between the acquired sensing data and human health status from a system scenario. In this article, we report a wearable smart sensor system (WS3), which can not only monitor the concentration of sodium ions in human perspiration but also predict the dehydration state of the human body. The proposed WS3 consists of an entire three-layer Internet of Things (IoT) structure. The perception layer includes a sweat sensor with a sensitivity of 60.3 ± 2 mV/decade, an analog front-end (AFE) with a signal amplification gain of 3, and a rechargeable 3.7-V Li-ion battery. The signal conditioning circuit can read the sodium ion concentration data obtained by the sweat sensor and wirelessly send it to the mobile phone APP in the application layer through Bluetooth. Besides, the mobile phone APP can exchange the data with the cloud server in the network layer through hypertext transfer protocol secure (HTTPS) requests, allowing the real-time post of concentration data and acquiring predicted dehydration states. Moreover, a lightweight deep-learning (DL) algorithm based on a Seq2Seq long short-term memory (LSTM) model with Luong attention is implemented in the cloud server, which achieves an overall accuracy of above 91% in the prediction of dehydration. The performance achieved by the WS3 combined with its high level of convenience and compactness makes it a promising wearable system for deployment in the IoT for daily human healthcare.