Glucose signal estimation using high performance ADC and lightprobes

Abstract

A need is felt in recent times to depart from invasive methods of glucose estimation and to adopt noninvasive approaches which are basically harmless in nature. These techniques are more attractive as they surpass the primary fear of the patient i.e. pricking of the body for blood extraction, etc. One of the methods of Glucose estimation is by using a light probe which utilizes the NIR region of electromagnetic radiation. Glucose presents unique absorbance spectra in the NIR region. The authors propose to design an electronic system using LED sources and photodiode detectors as sensors. After the light has passed through the sample, the signal falling on the detector will be used to calculate the absorbance. A trans impedance amplifier will amplify the weak current signal and pass it to a ADC. The system will be equipped with a high precision 32-bit ADC viz.ADS1262.This sigma delta ADC works at 38kSPS and has 10 input channels. These input channels can be used in either in differential mode or singleended mode. This ADC is interfaced using SPI interface and includes additional features such as 50/60 Hz Rejection, GPIO, Oscillator, Programmable gain control and Temperature Sensor. The digitized signal will be processed using multivariate techniques such as PLSR and the model for prediction will be built using the calibration data. Once the model is ready the glucose concentration can be predicted from unknown sample using the prediction model. The accuracy of the glucose estimation depends heavily upon the resolution of the ADC. Therefore the choice of an ADC is of outmost importance in the design and development of instrumentation for glucose estimation.