Anatomically Informed Multiresonant RF Sensor for Personalized Fingertip-Based Glucose Monitoring

Non-invasive glucose monitoring stands as a key pursuit in digital health, aiming to reduce the burden of frequent blood sampling for diabetic individuals. Radio frequency (RF) sensors emerge as promising solutions due to their non-contact nature and sensitivity to dielectric changes. However, their accuracy remains constrained by inconsistent tissue interaction resulting from subject-to-subject variability in skin thickness. In this study, we developed a stepped-line interdigital capacitor (SLIDC)-based RF sensor for anatomically-informed, multi-resonant glucose monitoring at the fingertip. The SLIDC is incorporated with a substrate-integrated waveguide (SIW) to ensure compactness for fingertip application, and a Defected Ground Structure (DGS) for enhanced operational bandwidth. The sensor was validated throughin-vitroandin-vivoexperiments. Forin-vitrotesting, 3D-structured PET-based fingertip-mimicking microfluidic layers with varying thicknesses (0.3–0.6 mm) were used, whilein-vivotesting was conducted using human volunteers. Verification results showed that each resonance frequency was associated with a specific thickness, demonstrating the ability of the sensor to identify individualized optimal sensing frequencies. Strong mean correlations were seen with glucose concentration, recording |R| up to 0.98, and |R| up to 0.9 forin-vitroandin-vivotests respectively. These findings confirm the system’s potential for tissue-thickness-informed, subject-specific glucose monitoring, supporting the future development of fingertip-deployable RF biosensors for digital health applications.