Microstrip Antenna Design for a non-Invasive Glucose Sensor

Diabetes Mellitus (DM) is one of the diseases with many patients in the world. The current method used by people with diabetes was an invasive approach using blood samples to monitor their blood glucose levels. The non-invasive approach that does not require blood samples would increase ease-of-use for diabetics. This research presents a microstrip antenna to monitor glucose levels in human blood non-invasively. The background concept of this approach was blood glucose variation causes variations of blood permittivity accordingly. This blood permittivity will affect the resonant frequency of the microstrip antenna. Antenna design and simulation were performed using CST Microwave Studio software. The antenna is designed on FR4 (εr = 4.3) dan NPC H220 (εr = 2.17) substrate with h = 1.6mm. A patch of microstrip antenna placed on one side and ground plane on another side, they are connected using SMA connector. The fabrication result of the designed antenna has an operating frequency of 2.17 GHz. In this research, the correlation between blood glucose and changing the microstrip antenna’s resonant frequency was observed. For this noninvasive method, the microstrip antenna is placed facing the forearm. Experimental results of both approaches, antenna placement with and without spacing to the forearm, show the shift of resonant frequency according to blood glucose changes. Based on this research’s linearity, accuracy, and sensitivity result, the microstrip antenna shows potential as a glucosesensing device for a non-invasive method with further development.


INTRODUCTION
Noninvasive determination of biological parameters is a research challenge now a day (Sen & Anand, 2021). Diabetes is a disease that affects body metabolism and already afflicts people worldwide. It is caused by impaired production of insulin; this increases the glucose level in the blood and leads to chronic hyperglycemia (Kharroubi, 2015). Diabetes Mellitus has risen dramatically in the last decade, and mainly due to changing lifestyle, increased prevalence of obesity and longevity and global projection shows diabetics increased by 2010 and will double in 2025 (Saeedi, et al. 2019) (Arinzon, et al. 2008) (Bethel, et al. 2007). The diabetic patients use invasive technique to monitor their Blood Glucose Level (BGL) on a regular basis .
To determine people with diabetes is to test sugar levels in the blood. Most test devices currently use invasive methods that use blood samples drawn by syringe and cause pain during the test, and it's not suitable for long-term monitoring (Omer, et al. 2020) (Turgul & Kale, 2018).
Therefore, a non-invasive method is desirable to reduce pain and a more comfortable way because it does not require blood extraction (Sains, et al. 2015). Several studies show non-invasive methods, like impedance spectroscopy, reverse iontophoresis, infrared spectroscopy, and many others (Kumar & Jayanthy, 2020).
The present work used microstrip antennas for non-invasive glucose sensing and fabricated using two different substrates FR4 and NPC H220 (Turgul & Kale, 2018). This paper will be based on changes in dielectric properties of blood as shown in fig.1 and lead to shifting the resonant frequency of antenna (Freer & Venkataraman, 2010) (Hayashi, 2003) (Cespedes, 2017). The remaining chapters of this paper are organized as follows. Chapter II describes the proposed antenna design in detail. Chapter III explains the results and analysis.

ANTENNA DESIGN
The proposed antenna is a microstrip antenna that is designed and simulated using CST Microwave Studio software. Microstrip antenna has a shape like a strip with small size and thin, and it consists of two side conductors known as radiating patch and ground plane and separated by dielectric substrate (Milligan, 2005) (Garg, et al). The main reason to choose this antenna model is due to the size and easy to manufacture, Based on the dielectric properties of blood present on (Freer & Venkataraman, 2010) (Hayashi, 2003) (Cespedes, 2017), the proposed microstrip antenna is required to able to penetrate multiple layers of the body through the blood. An electromagnetic wave with a wavelength of 1 cm is necessary to reach up to the blood, and a suitable frequency choice varies from 1 GHz to 5 GHz . Therefore, for our antenna, we choose around 1 to 2 GHz for the antenna working frequency.   The microstrip antenna in this work consists of two different substrate materials, FR4 (εr = 4,3) and NPC H220 (εr = 2.17), with a height of 1.6mm. Table 1 shows the overall final dimension of the microstrip antenna. Size of the substrate equal to the ground plane size, 60*70mm for FR4 and 70*80mm for NPC H220. Both are feed using feed line 50 Ω characteristic impedance.  Fig. 5 show the return loss results at the designed microstrip with two substrate materials FR4 and NPC H220, using a vector network analyzer (VNA). It can be observed return loss of FR4 is -31.14 dB at frequency 2.17 GHz, and NPC H220 is -42.23 dB at frequency 1.65 GHz.

RESULT AND DISCUSSION
Both the proposed microstrip antenna is analyzed by determining the correlation of shift frequency to glucose level. The human subject's glucose level is measured using the invasive method and wrote down. And using our microstrip antenna interfaced to VNA and placed on the antenna on the forearm as shown in Fig. 6 simultaneously, the return loss and frequency are observed. The process repeated for various blood glucose levels and created a dataset corresponding to shift frequency and invasive blood glucose. Each antenna is measured on two conditions, one with a ±1 cm gap from the forearm and antenna and one without a gap.   TEKNIK: Jurnal Ilmiah Ilmu-Ilmu Teknik Vol. 7, No. 1, Maret 2022p-ISSN 2656-7288, e-ISSN 2656 `5 Fig. 8. FR4 (with a distance) antenna response to BGL variation As shown in Fig. 7 and 8, a dataset of microstrip antennas with the FR4 substrate material shows changes in the antenna's frequency correlated to glucose level changes. The measurement without a gap shows varying frequency from 1,88000 GHz -1,91625 GHz with 86 mg/dL -153 mg/dL glucose levels, and for measurement with a gap shows 2,11500 GHz -2,14125 GHz with 96 mg/dL-160 mg/dL for glucose. It shifted around 250 MHz to 290 MHz of FR4 antenna for measurement without gap and around 20 MHz to 50MHz for measurement with a gap from the initial return loss. These datasets are then evaluated using simple linear regression y = ax + b and shown as a yellow line at all graphics above. Error value calculated by comparing blood glucose using linear regression and invasive method. For FR4 without a gap shows 1,808% -5,063% and with a gap around 1,678% -16,204%, and NPC H220 without a gap error value vary from 2,892% -13,378% and with a gap around 0,161% -16,192%. Table 2 shows R square from linear regression and data validation from both antennas with error average and sensitivity.

CONCLUSIONS
In this paper, the relation between glucose and frequency is observed. The resonant frequency of both antennas shows increasing value according to decrease blood glucose level. The result shows decent accuracy and sensitivity using a microstrip antenna as a sensor. In future work, collecting sufficient datasets and optimizing the existing model can be used to estimate blood glucose level non-invasively.

ACKNOWLEDGMENT
The preferred spelling of the word "acknowledgment" in America is without an "e" after the "g". Avoid the stilted expression "one of us (R. B. G.) thanks ...". Instead, try "R.B. G. thanks...". Put sponsor acknowledgments in the unnumbered footnote on the first page.