Microwave-Based Non-Invasive Blood Glucose Sensors: Key Design Parameters and Case-Informed Evaluation

Microwave-based non-invasive blood glucose monitoring (NIBGM) has emerged as a promising alternative to traditional finger-prick and enzymatic methods, offering painless, continuous, and real-time glucose tracking. This paper presents a comprehensive technical review and analytical framework for the design, modeling, and evaluation of microwave NIBGM sensors. Critical sensor architectures such as split-ring resonators (SRRs), complementary SRRs (CSRRs), dielectric resonator antennas (DRAs), and microfluidic electric-LC (ELC) structures are reviewed in terms of sensitivity, operating frequency, coupling efficiency, and material integration. We develop analytical and empirical models to simulate dielectric behavior, resonant frequency shifts, and permittivity changes across physiological glucose concentrations. Safety compliance is addressed using specific absorption rate (SAR) metrics aligned with IEEE C95.1 and ICNIRP standards. To enhance prediction accuracy, we apply machine learning models artificial neural networks (ANN), support vector regression (SVR), and long short-term memory (LSTM) to sensor outputs. These models are evaluated using mean absolute relative difference (MARD), coefficient of determination (<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$R^{2}$ </tex-math></inline-formula>), and Clarke Error Grid zones, benchmarked against commercial glucometer data from phantom and preliminary in vivo measurements. The validation strategy includes cross-validation and repeatability assessments, with plans for future deployment on standardized clinical datasets. Structured visual workflows and comparative performance graphs are presented to aid interpretability. Overall, the proposed framework synthesizes electromagnetic theory, safety constraints, and data-driven intelligence into a cohesive architecture for next-generation wearable glucose monitors that are accurate, safe, adaptive, and clinically aligned.