Cloud Recording for Diabetes Regulation of Blood Glucose Concentrations

摘要

Treatment of type 1 diabetes has improved with the application of continuous glucose monitoring and insulin pumps. Human input is still typically required for estimation of food-intake and level of physical activities. Many individuals find accurate estimation and reporting difficult, resulting in a mismatch between the levels of insulin injection and the rise in concentration of blood glucose, which the insulin is meant to counter. Machine learning algorithms could help find historical patterns and interpret manual estimations of food-intake and physical activity, as well as sensor input from continuous glucose monitoring devices and insulin pump devices. Development of machine learning models requires data having sufficient volume and reliability. The Internet of Things (IoT) could be used to transfer diabetes related data to cloud databases, which could then be analyzed. This study has started development of such a system. Google Sheets was selected for the prototype cloud storage using Google Script API for HTTP RESTful data transfer. A simulation of food-intake, digestion and blood glucose regulation was used for development of the prototype system instead of physical devices and human subjects. The initial results show promise, but more development and testing is required.