Permittivity Extraction of Glucose Solutions Through Artificial Neural Networks and Non-invasive Microwave Glucose Sensing Turgul, V. and Kale, I.

obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.The WestminsterResearch online digital archive at the University of Westminster aims to make the research output of the University available to a wider audience. Copyright and Moral Rights remain with the authors and/or copyright owners.Whilst further distribution of specific materials from within this archive is forbidden, you may freely distribute the URL of WestminsterResearch: ((http://westminsterresearch.wmin.ac.uk/).In case of abuse or copyright appearing without permission e-mail repository@westminster.ac.uk Sensors-18665-2017 1 Abstract-The non-invasive measurement of blood glucose is a popular research topic where RF/microwave sensing of glucose is one of the promising methods in this area. From the many available measurement sites in the human body, fingertips appear to be a good choice due to a good amount of fresh blood supply and homogeneity in terms of biological layers present. The non-invasive RF measurement of blood glucose relies on the detection of the change in the permittivity of the blood using a resonator as a sensor. However, the change in the permittivity of blood due to the variation in glucose content has a limited range resulting in a very small shift in the sensor's frequency response. Any inconsistency between measurements may hinder the measurement results. These inconsistencies mostly arise from the varied thickness of the biological layers and variation of fingerprints that are unique to every human. Therefore, the effects of biological layers and fingerprints in fingertips were studied in detail and are reported in this paper.

Non-invasive blood glucose measurement has attracted great interest from researchers deploying various techniques where microwave sensing is one of them. Microwave resonators are utilised as sensors for measuring the glucose levels. A body part is placed on the sensor for a reading and the measurement principle lies in the change of the dielectric properties of blood with varying levels of glucose. The fingertip is a popular measurement site as there is a good amount of fresh blood supply. The position of the fingertip on the sensor has an effect on the sensor response due to the change in the propagation path of the EM field inside the finger. Moreover, fingerprints also affect the sensor response as the irregular ridges and valleys in the fingerprint introduce air gaps altering the effective permittivity seen by the sensor. The effects of fingerprints as well as finger positioning on the sensor is explored and explained.Introduction: Diabetes mellitus is a well-known disease associated with high blood glucose levels (BGLs), affecting over 400 million people worldwide and the number keeps increasing [1]. At the moment, the conventional BGL measurement method is the use of portable measurement devices where a small blood sample is needed; therefore, this method is uncomfortable. Hence, there is a great deal of research conducted on non-invasive BGL measurement methods [2]. Microwave sensing is a promising method where microwave resonators are used as the sensing element. A variation in the concentration of the glucose in blood induces a change in the overall dielectric properties of the blood. The sensor's frequency response changes accordingly as a result of this. Preliminary studies show that a frequency shift of <8 MHz is expected with varying levels of glucose in the blood [3]. The fingertip is the most common site to obtain the blood sample for the conventional method and it is also a desirable measurement site for the non-invasive microwave BGL measurements due to having a lot of fresh blood supply. Furthermore, it is easy to position the finger on the sensor and the fingertip is homogenous in terms of biological layers present. However, a positioning error may introduce more frequency shift than the shift caused by the change in the glucose content. Moreover, the irregular shape of the fingerprint in the fingertip adds to this error. To demonstrate this, simulations were performed with the conventional 4-layer fingertip model as well as a novel model with a fingerprint imprinted. Additionally, the effect of positioning of the finger on the sensor was investigated. There are other irregularities on whichever measurement site used possibly contributing to the measurement error such as hair follicles on skin, sweat ducts and moisture due to sweat which also need to be considered; however, these are not in the scope of this work.

No abstract

No abstract