Paper Based Smart Sweat Sensor for Health Monitoring

The paper presents a novel approach of noninvasive health monitoring using a paper-based smart sweat sensor. Human sweat is mainly water, containing a varied range of biomarkers such as electrolytes (sodium, potassium, chloride, ammonium etc.), metabolites, trace elements, neuropeptides, and cytokines. By leveraging image processing and machine learning (ML) techniques, the proposed system analyzes sweat samples by comparing the intensity of yellow color precipitates on filter paper strips due to a chemical reaction between sweat and a chemical reagent to determine electrolyte concentration in the sweat. We developed a model to predict health conditions based on input biomarkers such as pH, glucose, sodium, and potassium. In this study, apart from testing potassium concentration, the pH and glucose levels are synthetically assumed. The work was carried out to provide a simple, cost-effective, and reliable alternative non-invasive method for determining electrolyte (potassium) concentration in sweat for continuous health monitoring and predicting health condition. Following precise chemical analysis on collected sweat samples, accurate concentration of potassium (2.5mmol/ltr) is predicted through image processing. The proposed smart sweat sensor is a significant advancement towards the development of a non-invasive, simple and accurate wearable sensor in the future.