Abstract:
Dehydration is a common health challenge among athletes. Many attempts
have been made to detect dehydration using the conductivity of perspiration
as a biomarker. However, due to the poor reliability and portability of those
devices, the requirement for an alternative detection method is required. Thus,
a real-time human sweat conductivity-based portable non-invasive device
combined with an interdigitated electrode sensor (IDE and DS18B20 sensor)
and an Arduino Uno platform was developed. Sweat impedance was used to
predict hydration status. An Android application was designed to provide
alerts via Bluetooth connection to alarm the user regarding the hydration
status. The device was calibrated with a 10 μS KCl solution and cross-checked
with the conductivity of known electrolytes. The sensor was tested for human
sweat before and after consuming commercial electrolyte beverages in a
controlled exercise environment. Results showed a noticeable 2.42 (± 0.26)
times average conductivity increase compared to the initial conductivity with
a 10.61% percentage deviation for each person before reaching a predefined
dehydration level without consuming any electrolyte. After electrolyte
consumption, this value was reduced to 2.03 (± 0.44) with a higher percentage
deviation of 21.72%. Initial average conductivity and the increase in
percentage sweat conductivity were used to programme the device to provide
a warning signal to the user regarding potential dehydration.
