7 research outputs found
Design and development of an optical blood glucose measurement for infrared and near-infrared testing / Sarah Addyani Shamsuddin
No full textYearly, four million people to die because of diabetes and it also leads people to other serious disease. Hence, the existence of the portable blood glucose self check device is very helpful to the patients and others who concern to know their blood glucose reading. Non-invasive method is more preferable since it should be painless compared to conventional finger pricking device. Besides, patients who need to do self test often need to refer to the procedures and storing surrounding condition which can be hassle to some people with language restriction. Alternatively, using non-invasive will not be wasteful and the measurement or reading can be done any time and numbers of time. Previously, many researches had been done on non-invasive using near-infrared sensing. From previous research by Sia, he had investigated near-infrared sensing using signal penetrating finger method. However, by using finger penetration, there are no results obtained. He only obtained signal using glucose concentration. Therefore the objectives of this research are to investigate the performance of three different wavelength of sensors; infrared 940nm and infrared 950nm and also near infrared 1450nm. Sensor that gave the best output had been chosen to achieve the second objective of this project which is to design non-invasive blood glucose measurement device based on optical sensing and to develop prototype of a blood glucose optical sensing instrumentation with acceptable accuracy and repeatability. Generally, the overall system consists of three parts including sensor part, signal conditioning circuit, and also numerical display. The initial design tested by considering initial voltage 1.616Vto 1.68V which referred to previous research by Sia as the output of the sensor. Then proceed by using test tube which contains various percentage of glucose concentration. The same methods had been used to the human samples fingers instead of test tube. From the experiment, output graph of the 950nm shows more consistent pattern compared to the 940nm. 950nm also has a larger range scale for voltage which from 5.016V to 5.4633V compare to the 940nm voltage range scale which from 5.0327V to 5.4201V. Further test on human finger had been done by using 950nm infrared but the output voltages were too small. The performance of the measurement can be improved by controlling the surrounding condition and fixed the path length between transmitter and receiver. Test using test tube showed that the near infrared and infrared were capable to predict different glucose concentration. By comparing the performance of infrared and near-infrared, near-infrared gave better performance since near-infrared had higher output voltage range which from 0.6 to 3.4174V compared to infrared. Graph near-infrared output voltage shows that the voltage is almost directly proportional to the percentage of glucose concentrations. By using circuit designed, it can be seen that the voltage reading became higher compared to before meal which shows that there were increment in glucose reading from before to after meal. Therefore, it can be concluded that the circuit design functions accordingly and non-invasively. During human sample test, increment pattern can be seen from fasting to non-fasting condition but the main effect is all samples have different fingers' diameter which each of user needs to be calibrated
Comparative investigation using GaAs(950nm), GaAIAs (940nm) and InGaAsP (1450nm) sensors for development of non-invasive optical blood glucose measurement system
No full text
Design simulation and theoretical analysis of glucose sensing using Polysilicon-based CMOS micromachined Piezoresistive Microcantilever
No full textInitial quantitative comparison of 940nm and 950nm infrared sensor performance for measuring glucose non-invasively
No full textSystematic literature review of machine learning methods in insulin secretion model analysis / Mohd Hussaini Abbas ... [et al.]
No full textEndogenous insulin secretion (UN) plays a critical role in maintaining glucose homeostasis. Pathological changes in UN enable early detection of metabolic inefficiency prior to the onset of diabetes mellitus (DM). Numerous researches have been carried out to establish the most effective method for assessing the participant’s glycemic state by identifying their UN profile. In contrast to insulin sensitivity (SI), there is no gold standard for UN profile. Thus, the deconvolution of C-peptide measurements is used in the majority of research to identify the UN profile. Due to the fact that C-peptide and insulin are co-secreted equimolarly from pancreatic β-cells, the latter method is shown to be accurate. Although studies have shown that the machine learning-based strategies can yield very positive outcomes in other areas of DM diagnosis, there is currently little research that employing machine learning for quantifying the UN profile to enable early diagnosis of metabolic dysfunction. Hence, the main objective of this study is to conduct a thorough search on machine learning-based modelling strategies that were used to identify the individual specific UN profile through the development of a UN model. Additionally, this study will investigate whether the data acquired from the UN model can be used to quantify a person’s metabolic condition (either normal, pre-diabetic or T2D). The literature search turned up prospective studies linking machine learning and UN in its search and analysis. Meta-analyses summarize the available data and highlight various methodological stances. Thus, the exploratory of machine learning classification and regression technique can be portrayed in 3 different scenarios during the identification of UN profile. The 3 scenarios are: the study of insulin secretion through analyzing the insulin sensitivity, the study of UN without taking into considerations or in-depth study of U1 and U2, and the study of insulin secretion using deconvolution of plasma C-peptide concentrations. It is evident that while Decision Tree (DT) is ideal for the first scenario, Random Forest (RF) is the better option for the other two scenarios. Further optimization can be implemented with the use of these techniques under supervised learning to improve diagnosis and comprehend the pathogenesis of diabetes, particularly in UN
