Investigations on a CSRR-Based Bio-Sensor for Dielectric Characterization of Blood

This paper investigates the use of a single-layer microwave filter for the dielectric characterization of biological materials, with the aim of developing a non-invasive, low-cost sensor for glucose monitoring. The proposed biosensor exploits the frequency response of a passive microwave structure to detect changes in complex permittivity associated with glucose concentration in aqueous solutions mimicking blood. Two cost functions are compared to drive the optimization aimed at maximizing sensitivity (sensitivity-driven optimization): the squared error and the cross-correlation. Using a design of experiments approach, the latter proves to be the better candidate for guiding sensitivity optimizations.