276430 Modelling and Robust Model Predictive Control of Insulin Delivery in Type 1 DM

Wednesday, October 31, 2012: 9:05 AM
Pennsylvania East (Westin )
Stamatina Zavitsanou1, Nicki Panoskaltsis2, Athanasios Mantalaris1, Michael C. Georgiadis1 and Efstratios N. Pistikopoulos3, (1)Department of Chemical Engineering, Centre for Process Systems Engineering, Imperial College, London, United Kingdom, (2)Department of Haematology, Northwick Park Hospital, London, United Kingdom, (3)Chemical Engineering, Imperial College London, Centre for Process Systems Engineering, London, United Kingdom

The determination of effective strategies to treat Type 1 Diabetes Mellitus is of paramount importance since it can be characterised as one of the most prevalent severe chronic diseases in childhood. Motivated by the challenge to ensure the patients’ safety and to improve their quality of life the development of a closed loop system of optimal insulin delivery has been the most widely studied approach in the field (Kovatchev et al., 2010), (Grosman et al., 2010). In this framework, a detailed physiologically based compartmental model is presented, extended from previous studies (Zavitsanou, et al. 2011). It represents personalised glucose-insulin interactions in the fasting, prandial and postprandial state while insulin is administered exogenously with an insulin pump and mimics the mechanism of pancreatic response. Glucose coming from food consumption and endogenously produced is distributed in blood and tissue compartments, representing the involved organs and glucose regulation is achieved with the appropriate subcutaneous insulin infusion to compensate for high blood glucose concentrations. Patient specific characteristics such as age, gender, weight, height, hematocrit and insulin sensitivity are included and the model can represent a virtual patient. The individual insulin delivery (dosing and scheduling) is optimised in terms of maintaining blood glucose concentration within the safe range. Closed loop studies in the context of explicit/multi-parametric model predictive control (Dua et al., 2006) are presented to evaluate the performance of the controller that takes into account safety constraints of desired blood glucose concentration and operational constraints of available insulin delivery. Case studies of closed loop insulin delivery of patients subjected to external disturbances are presented to highlight the challenges of the system such as high intra patient variability and time delays.      

Dua, P., Doyle, F. J., & Pistikopoulos, E. N. (2006). Model-based blood glucose control for Type 1 diabetes via parametric programming. IEEE transactions on bio-medical engineering, 53(8), 1478-91.

Grosman, B., Dassau, E., Zisser, H. C., Jovanovič, L., & Doyle, F. J. (2010). Zone model predictive control: a strategy to minimize hyper- and hypoglycemic events. Journal of diabetes science and technology, 4(4), 961-75.

Kovatchev, B., D, P., Cobelli, C., Renard, E., Anderson, S., Breton, M., et al. (2010). Multinational study of subcutaneous model-predictive closed-loop control in type 1 diabetes mellitus: summary of the results. Journal of diabetes science and technology, 4(6), 1374-81.

Zavitsanou, S., Panoskaltsi, N., Mantalaris, A., Georgiadis, C, M., & Pistikopoulos, E. N. (2011). Modelling of the Insulin Delivery System for patients with Type 1 Diabetes Mellitus. 21st European Symposium on Computer Aides Process Engineering.

Extended Abstract: File Not Uploaded