Executive Summary : | A prototype of a non-invasive photoacoustic based continuous blood glucose measurement system has already been developed and validated on glucose solutions and partially on humans. The project is aimed to design & development a non-invasive continuous blood glucose measuring SoC based on photoacoustic spectroscopy. The project will involve the characterization of the effects of biological constituents of tissues and physiological factors on the prototype measurements, followed by accuracy and reliability testing. According to WHO estimates, India has the largest diabetic population in the world, with the number of diabetes cases projected to rise from 32 million in 2000 to 80 million by 2030, and a present mortality rate of 22.4 per 100,000. The sudden outbreak of the novel Coronavirus makes especially diabetic people more vulnerable. There is no such permanent cure for diabetes as of now, but continuous monitoring of blood glucose and its tight control leads to a 40-80% reduction in the development and progression of associated disorders such as retinopathy, nephropathy, and neuropathy. Though diabetes mellitus can be more accurately monitored using glycated hemoglobin (HbA1c) meter, people are reluctant to use it due to its invasive nature. Painless, sample-free, low-cost, continuous, and non-invasive measurement is thus the possible way to combat this unrelenting disease. Amongst various non-invasive measurement techniques, Photoacoustic Spectroscopy (PAS) is widely practiced for blood glucose concentration determination. Prediction of blood glucose values from the turbid blood in the human body is a challenging task to perform. Every measurement is needed to be verified with hospital data. The thorough analysis of the results helps us to develop an algorithm to nullify the effect of unwanted factors causing erroneous results. Unlike any other system design, the development of a non-invasive bio-analytes monitoring system would require extensive characterization of different associated biological constituents including physiological factors to provide a reliable and accurate picture of the desired component in blood. So, all factors that appear to impede the accuracy of blood glucose information like arterial oxygen saturation, tissue character at the test site, melanin content of the skin, the natural vibration of the body, changes in ambient and operating conditions, etc. will be characterized using different optical methods at multiple wavelengths, impedance spectroscopy, the mechanical design of the housing, etc. A suitable Machine Learning algorithm will be developed keeping the accuracy, repeatability, and cost factor into consideration. After feature computation and optimal feature selection followed by calibration of PA measurements, a calibration algorithm will be developed and implemented on SoC. Finally, the display and blue-tooth/Wi-Fi module will be integrated with the SoC to visualize and access the data with waveform remotely. |