Patients with type 1 diabetes would like to enjoy carefree and active lifestyles, conduct physical activities and exercise programs. A closed-loop insulin pump that does not necessitate manual inputs such as meal or physical activity information from the patient can accommodate these wishes. But the interpretation of sensor information and adaptation of the control system to significant metabolic variations is critical. This necessitates mathematical models that can represent the patient's state accurately as her/his metabolic state changes due to a wide spectrum of causes such as meals, physical activity, or stress. Detailed nonlinear models are not attractive for building the closed-loop control systems for miniaturized devices. They are difficult to adjust for each subject and for their metabolic variations over time and they consume significant computational resources. The alternative is simple recursive models that are updated at each sampling time to adapt to the current state of the subject. This project focuses on the development and clinical evaluation of: (1) Recursive patient-specific dynamic models using subcutaneous glucose measurements and physiological data that measure physical activity and stress to provide accurate predictions of blood glucose concentrations;(2) Early warning systems for hypoglycemia;and (3) Adaptive controllers based on these recursive models to manipulate the insulin infusion rate. Young adults in the 18-25 age group will be the focus of the study. Continuous glucose monitors (CGM) will provide glucose concentration information. Physiological signals from an armband body monitoring system will provide the metabolic/physiological information. The elimination of manual inputs entered by patients will reduce the inconveniences that they are experiencing on a daily basis and potential for human errors. Multiple-input (measured glucose concentration and metabolic/physiological information) single-output (predicted glucose concentration) models will be developed for the hypoglycemia warning and closed-loop control system. Generalized predictive controllers (GPC) and self-tuning regulators will be developed for regulating the blood glucose level by manipulating the insulin infusion rate. The performance of the modeling and control techniques will be evaluated by simulation studies using detailed compartmental models as in silico patients and clinical studies conducted at the General Clinical Research Center at Chicago Biomedicine. This project is a collaborative effort between Illinois Institute of Technology, University of Chicago Medical Center (now renamed Chicago Biomedicine), University of Illinois Chicago, and Iowa State University. The collaborative efforts of engineering, medicine and nursing combined with the """"""""bench to bedside"""""""" design of the proposed study is consistent with the goals of translational research.
Closed-loop glucose concentration control systems based on recursive models and adaptive controllers that do not necessitate any manual inputs from the patient would be very appealing to young adults with type 1 diabetes who have active lifestyles. This research will provide patient-specific dynamic models that predict blood glucose concentrations accurately by using subcutaneous glucose measurements and physiological data that measure physical activity and stress, early warning systems for hypoglycemia, and adaptive controllers based on these models that regulate blood glucose concentration by manipulating the insulin infusion rate.
| Turksoy, Kamuran; Quinn, Lauretta T; Littlejohn, Elizabeth et al. (2014) An integrated multivariable artificial pancreas control system. J Diabetes Sci Technol 8:498-507 |
| Turksoy, Kamuran; Cinar, Ali (2014) Adaptive control of artificial pancreas systems - a review. J Healthc Eng 5:1-22 |
| Turksoy, Kamuran; Quinn, Laurie; Littlejohn, Elizabeth et al. (2014) Multivariable adaptive identification and control for artificial pancreas systems. IEEE Trans Biomed Eng 61:883-91 |
| Turksoy, Kamuran; Bayrak, Elif S; Quinn, Lauretta et al. (2013) Hypoglycemia Early Alarm Systems Based On Multivariable Models. Ind Eng Chem Res 52: |
| Turksoy, Kamuran; Bayrak, Elif Seyma; Quinn, Lauretta et al. (2013) Multivariable adaptive closed-loop control of an artificial pancreas without meal and activity announcement. Diabetes Technol Ther 15:386-400 |
| Bayrak, Elif Seyma; Turksoy, Kamuran; Cinar, Ali et al. (2013) Hypoglycemia early alarm systems based on recursive autoregressive partial least squares models. J Diabetes Sci Technol 7:206-14 |
| Eren-Oruklu, Meriyan; Cinar, Ali; Rollins, Derrick K et al. (2012) Adaptive System Identification for Estimating Future Glucose Concentrations and Hypoglycemia Alarms. Automatica (Oxf) 48:1892-1897 |
