Omer, Ala Eldin2022-05-162022-05-162022-05-162022-04-18http://hdl.handle.net/10012/18282Monitoring glycemia levels in people with diabetes has developed rapidly over the last decade. A broad range of easy-to-use systems of reliable accuracies are now deployed in the market following the introduction of the invasive self-monitoring blood glucose meters (i.e., Glucometers) that utilize the capillary blood samples from the fingertips of diabetic patients. Besides, semi-invasive continuous monitors (CGM) are currently being used to quantify the glucose analyte in interstitial fluids (ISF) using an implantable needle-like electrochemical sensors. However, the limitations and discomforts associated with these finger-pricking and implantable point-of-care devices have established a new demand for complete non-invasive pain-free and low-cost blood glucose monitors to allow for more frequent and convenient glucose checks and thereby contribute more generously to diabetes care and prevention. Towards that goal, researchers have been developing alternative techniques that are more convenient, affordable, pain-free, and can be used for continuous non-invasive blood glucose monitoring. In this research, a variety of electromagnetic sensing techniques were developed for reliably monitoring the blood glucose levels of clinical relevance to diabetes using the non-ionizing electromagnetic radiations of no hazards when penetrating the body. The sensing structures and devices introduced in this study were designed to operate in specific frequency spectrums that promise a reliable and sensitive glucose detection from centimeter- to millimeter-wave bands. Particularly, three different technologies were proposed and investigated at the Centre for Intelligent Antenna and Radio Systems (CIARS): Complementary Split-Ring Resonators (CSRRs), Whispering Gallery Modes (WGMs) sensors, and Frequency-Modulated Continuous-Wave (FMCW) millimeter-Wave Radars. Multiple sensing devices were developed using those proposed technologies in the micro/millimeter-wave spectrums of interest. A comprehensive study was conducted for the functionality, sensitivity, and repeatability analysis of each sensing device. Particularly, the sensors were thoroughly designed, optimized, fabricated, and practically tested in the laboratory with the desired glucose sensitivity performance. Different topologies and configurations of the proposed sensors were studied and compared in sensitivity using experimental and numerical analysis tools. Besides, machine learning and signal processing tools were intelligently applied to analyze the frequency responses of the sensors and reliably identify different glucose levels. The developed glucose sensors were coupled with frequency-compatible radar boards to realize small mobile glucose sensing systems of reduced cost. The proposed sensors, beside their impressive detection capability of the diabetes-spectrum glucose concentrations, are endowed with favourable advantages of simple fabrication, low-power consumption, miniaturized compact sizing, non-ionizing radiation, and minimum health risk or impact for human beings. Such attractive features promote the proposed sensors as possible candidates for development as mobile, portable/wearable gadgets for affordable non-invasive blood glucose monitoring for diabetes. The introduced sensing structures could also be employed for other vital sensing applications such as liquid type/quantity identification, oil adulteration detection, milk quality control, and virus/bacteria detection. Another focus of this thesis is to investigate the electromagnetic behavior of the glucose in blood mimicking tissues across the microwave spectrum from 200 MHz to 67 GHz using a commercial characterization system (DAK-TL) developed by SPEAG. This is beneficial to locate the promising frequency spectrums that are most responsive to slight variations in glucose concentrations, and to identify the amount of change in the dielectric properties due to different concentrations of interest. Besides, the effect of the blood typing and medication was also investigated by measuring the dielectric properties of synthetic “artificial” as well as authentic “human” blood samples of different ABO-Rh types and with different medications. Measured results have posed for other factors that may impact the developed microwave sensors accuracy and sensitivity including the patient’s blood type, pre-existing medical conditions, or other illnesses.encomplementary split-ring resonator (CSRR)whispering gallery mode (WGM)dielectric characterizationnon-invasive glucose sensingwearable CGMmachine learningblood typingartificial intelligenceFMCW radarmm-waves sensingflexible antennapoint-of-care devicesignal processingelectromagnetic sensorsdiabetes managementnumerical analysisDoctoral Thesis