Flexible Antenna Sensor in Thumb Spica Splint for Noninvasive Monitoring of Fluctuating Blood Glucose Levels

Abstract

Antenna is used as a sensor in wearable technology because of its small size. Reduced flexible antenna size avoids mounting problems and crumbling effects on the human skin region during signal acquisition. Moreover, continuous diabetes monitoring needs an efficient antenna sensor design. In this article, we develop a continuous glucose monitoring device using a flexible antenna sensor planted inside the thumb spica splint glove. The flexible antenna sensor comprises polyimide substrate with <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\xi _{r}$ </tex-math></inline-formula> = 3.5, tan <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\delta $ </tex-math></inline-formula> = 0.0027, and a thickness of 0.1 mm. The proposed flexible antenna sensor senses dielectric radiations from glucose in the blood and improves the accuracy of glucose-level prediction. Traditional antenna sensor methods use skin layers and muscles as sources of permittivity, resulting in errors in diabetic value measurement. The acquired antenna sensor signal from the designed thumb spica splint glove is processed using the tunable <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${Q}$ </tex-math></inline-formula> -factor wavelet transform (TQWT) algorithm. It differentiates the dielectric properties of glucose in the blood, skin, and muscle through subband energy levels. The subband energy level of glucose from the proposed antenna sensor and the quadratic regression algorithm is used to predict the human glucose level. The predicted glucose levels using the proposed flexible antenna are validated through Accu-Chek Active Glucometer. From experimental results, the proposed method-based glucose prediction-level accuracy achieved about 96.8% compared to traditional methods.