Blood glucose level monitoring is of great importance, especially for subjects experiencing type 1 diabetes. Accurate monitoring of their blood glucose level prevents dangerous and life-threatening situations that might be experienced by those subjects. In addition, precise monitoring of blood glucose levels over long periods of time helps establishing knowledge about the daily mealtime routine which aids the medical staff to monitor subjects and properly intervene in hazardous cases such as hypo- or hyperglycemia. Establishing such knowledge will play a potential role when designing proper treatment intervention plan. In this research, we present a complete IoT framework, starting from hardware acquisition system to data analysis approaches that gives a hand for medical staff when long periods of blood glucose monitoring are essential for subjects. Also, this framework is validated with real-time data collection from 7 subjects over 10 successive days with temporal resolution of 5 minutes allowing for near real-time monitoring and analysis. Our results show the precisely estimated daily mealtime routines for 4 subjects out of the 7 with discard of 3 subjects due to huge data loss mainly. The daily mealtime routines for the 4 subjects are found to be matching to have a pattern of 4 periods of blood glucose level changes corresponding to the breakfast around 8 AM, the lunch around 5 PM, the dinner around 8 PM, and finally a within-day snack around 12 PM. The research shows the potential of IoT ecosystem in support for medically related studies.

翻译：暂无翻译