AI-Tuned Smart Potentiometric based Sensor for Chloride Concentration Assessment in Sweat

This paper presents a wearable potentiometric sensor system for sweat chloride measurement, targeting cystic fibrosis screening. The platform integrates Ag/AgCl electrodes, an instrumentation amplifier, and analog signal conditioning to enable low-noise acquisition of electrochemical signals. A low-power microcontroller performs on-device signal preprocessing and implements a TinyML-based regression model for sensor calibration and chloride concentration estimation. Environmental validation is achieved through an integrated humidity sensor to ensure reliable sweat detection. Experimental results demonstrate accurate chloride quantification within the clinically relevant diagnostic range, while maintaining low power consumption and autonomous operation, making the proposed system suitable for point-of-care and wearable sensing applications. This study validates the feasibility of deploying "smart," self-contained sensors that can autonomously map raw potentiometric transients to clinical values.