Design and Fabrication of a Non-Invasive Near Infrared Meter for Glucose Monitoring
Loading...
Date
2025-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Kenyatta University
Abstract
In the current times, there is an alarming increase of diabetes in the world, which has largely been linked to changes in lifestyle. Diagnosing and controlling blood sugar levels are necessary for people with diabetes. The reason is to ensure that adequate action is taken to maintain insulin at acceptable levels. The methods used currently to measure blood sugar levels are frequently invasive or minimally invasive. These methods are effective and successfully used, but they are painful, costly, and carry a risk of infection. Thus, non-invasive, painless, infection-free and cost-effective alternative methods are currently being pursued. Numerous non-invasive techniques have been thoroughly studied. Notwithstanding the challenges of accuracy, calibration and bulkiness, these techniques have shown promising results for the development of non-invasive glucose meters. Given these challenges, the primary objective of this study was to develop an accurate Near Infrared (NIR) meter for non-invasive glucose monitoring that can identify both high and low blood sugar levels. Specifically, the research endeavored to design, develop, test and validate a NIR blood sugar level detector system. The NIR meter utilizes the transmittance principle by using a low-power infrared light-emitting diode (LED) with an emission wavelength of 1150 nm as a transmitter and a photodiode as a receiver. The signal is then filtered, linearized, stabilized and passed to a microcontroller. The Microcontroller uses Arduino programming language fitted with a regression equation to transform the signal and give an output. The output of the detector was first passed through the trans-impedance amplifier, a filter circuit, amplifier and a linearizer before passing to a microcontroller. In order to compute glucose concentrations, a linear regression equation was fitted into the software to calculate equivalent glucose levels of the voltages obtained. The results were displayed on the liquid crystal display (LCD) and transmitted to a remote device for monitoring via the Global System for Mobile communication (GSM) for monitoring and storage. An In-vitro approach was adopted to test the system. Standard glucose concentrations within all the sugar levels were prepared for the test. The samples were then set between the probe beam and the detector. During system testing, it was observed that the output voltages directly vary with glucose concentrations. Therefore, the voltage output is directly proportional to glucose concentrations. Thus, the results proved that an 1150nm infrared LED can be used to measure glucose levels non-invasively. The system was validated using Clark Error Grid and Bland Altman analysis. It was ascertained from the Clark error grid that the glucose concentration of the tested system falls within region A and region B of the Clark’s Error Grid. The majority of the values fell within both limits of agreement when Bland Altman's analysis was used. It implies that the device is suitable for measuring glucose concentration non-invasively. The residual plot for 1150nm wavelength suggests that the model satisfies the requirements of linear regression and offers an acceptable fit to the data and the Pearson correlation coefficient suggests a very strong positive linear relationship within the data set, with the measurements showing high consistency and predictability. However, since the scope of the study was limited to in-vitro measurements, it is recommended that more validation be done on a large number of data samples specifically with the use of in-vivo measurements and compare the results with commercially available glucometers.
Description
A Thesis Submitted in Partial Fulfilment of the Requirements for the Award of the Degree of Master of Science (Electronics and Instrumentation) in the School of Pure and Applied Sciences of Kenyatta University, June 2025.
Supervisors
1. Dr. Mathew K. Munji
Dr. Raphael L. Nyenge