Supported by National Institutes of Health – NIH, and Juvenile Diabetes Research Foundation - JDRF

Multivariable artificial pancreas systems for people with diabetes

Development of artificial pancreas systems that use glucose concentration and physiological variables from wearable devices to provide feedback and feedforward model predictive control. Development of recursive models to predict glucose concentration variations from glucose concentration and physiological variables measurements. Development of predictive hypoglycemia warning systems, physical activity, acute psychological stress and sleep assessment modules, estimators for real-time meal size and plasma insulin concentration and adaptive control techniques for artificial pancreas systems that can mitigate the effects of meals, exercise, sleep and stress (MESS). Extension of these algorithms in advisory systems for insulin pen users.

Simulation systems for glucose, insulin and physiological variable dynamics in Type 1 and Type 2 diabetes

Development of multivariable simulators that can perform simulation up to one year with random variations in the amounts and times of meals, and types, intensities and times of physical activities. In addition to glucose concentration and continuous glucose monitoring (CGM) values, heart rate, skin temperature , blood volume pulse, accelerometer readings and galvanic skin response are also predicted to the multivariable artificial pancreas systems . The simulator is extended to Type 2 diabetes to assess the efficacy of various drugs and lifestyle changes in regulating blood glucose levels of people with Type 2 diabetes.

Development of an agent-based model to simulate the destruction of insulin producing beta cells in the pancreas by the immune system and the onset of diabetes

Simulation of T cell – beta cell interactions in the pancreas to investigate various hypotheses for the destruction of beta cells and provide information for designing experiments by medical researchers.