Non-Invasive Blood Glucose Monitoring System

The objective of this project is to create a non-invasive hypoglycemic alert system that will detect a drop in blood sugar in type 1 diabetics during sleep. This will be achieved by creating an algorithm that couples heart rate variability with skin conductance to increase the accuracy of hypoglycemia detection. The device will be housed in a torso strap that will include: electrodes located over the user’s rib cage, a skin conductance sensor placed in the user’s armpit, a microcontroller to collect and process the data, and vibrating motors that will awaken the patient if hypoglycemia is detected. Integrating the ECG leads into the torso strap incorporates a capacitive circuit that reduces reverberation due to lead placement over the rib cage while also increasing user safety and accuracy of R-wave detection. This is in contrast to the standard bipolar three ECG lead arrangement. This technique was discovered after realizing the need for a more ergonomic design to allow for full range of motion for the user. Skin conductance will be measured through a sensor made of conductive fabric that will be placed in the patient’s armpit due to the high concentration of sweat glands while maintaining the ergonomics of the design.A LillyPad microcontroller will be programmed to collect and process the signals using Arduino Software and will include a SD Card for storage. The ECG signal will be amplified, filtered, and the R-wave will be detected. A timer within the system will determine the intervals of the R-waves which will create a plot of time versus the index number. This data will be saved on the SD Card. Welch’s Method of averaging Discrete Fourier Transforms (DFT) will determine the power of the low frequency band of the signal in order to compute spectral components. Previous research has shown that the power of the low frequency range (0.04-0.15Hz) of the ECG is related to hypoglycemia. Skin conductance will be measured using a low level constant current which will measure a change in conductivity of the skin via the conductive fabric located in the armpit that will be attached to the torso strap. If skin conductance increases along with a decrease in the power of the low frequency component of the ECG signal, the diabetic will be alerted via a vibrating motor in the torso strap.A finalized material and budget list as well as a finalized conceptual model were created for the Sternheimer Grant application. Materials for fabrication are in the process of being ordered and will be ready to begin prototyping in mid-January.