Dynamic Systems Modeling of Insulin to Glucose in Diabetic Pregnant Females

This project aims to find dynamic systems models of pregnant females exhibiting Type 1 Diabetes for their insulin to glucose behavior. Once a model is identified, this will enable development of an automatic feedback control scheme as a standalone device. This will greatly improve the quality of life of the individual and provide a margin of safety for the fetus during gestation. The project will collect data of injected insulin over time from an automated injection system and also gather measured glucose levels at selected intervals. This data can be parsed for time periods such as day, night, active, sedentary, and during and after meals. Moreover, data will be collected from individuals representing various body types and shapes.Collected data will be processed using carefully chosen mathematical techniques including optimization and machine learning methods. This will lead to dynamic systems models that can be incorporated into control system design approaches for automatic feedback control and eventual production of a device that will enable automated insulin delivery. Iman Erfan will study the efficacy and applicability of derived models for the range of patient data. This work will be under the supervision of medical professionals including the industrial sponsor, Inner Analytics.

Iman Erfan
Faculty Supervisor: 
Jeffery Kurt Pieper
Partner University: