Data-Driven Robust Control for Type 1 Diabetes Under Meal and Exercise Uncertainties
We present a fully closed-loop design for an artificial pancreas (AP) which regulates the delivery of insulin for the control of Type I diabetes. Our AP controller operates in a fully automated fashion, without requiring any manual interaction (e.g. in the form of meal announcements) with the patient. A major obstacle to achieving closed-loop insulin control is the uncertainty in those aspects of a patient’s daily behavior that significantly affect blood glucose, especially in relation to meals and physical activity. To handle such uncertainties, we develop a data-driven robust model-predictive control framework, where we capture a wide range of individual meal and exercise patterns using uncertainty sets learned from historical data. These sets are then used in the controller and state estimator to achieve automated, precise, and personalized insulin therapy. We provide an extensive in silico evaluation of our robust AP design, demonstrating the potential of this approach, without explicit meal announcements, to support high carbohydrate disturbances and to regulate glucose levels in large clusters of virtual patients learned from population-wide survey data.
Research supported in part by AFOSR Grant FA9550-14-1-0261 and NSF Grants IIS-1447549, CNS-1446832, CNS-1445770, CNS-1445770, CNS-1553273, CNS-1536086, and IIS-1460370.
- 1.Bertsimas, D., Gupta, V., Kallus, N.: Data-driven robust optimization. arXiv preprint arXiv:1401.0212 (2013)
- 2.Centers for Disease Control and Prevention. National Diabetes Statistics Report: Estimates of Diabetes and its Burden in the United States. US Department of Health and Human Services, Atlanta (2014)Google Scholar
- 6.Gondhalekar, R., Dassau, E., Doyle, F.J.: Moving-horizon-like state estimation via continuous glucose monitor feedback in MPC of an artificial pancreas for type 1 diabetes. In: 2014 IEEE 53rd Annual Conference on Decision and Control (CDC), pp. 310–315. IEEE (2014)Google Scholar
- 13.Jacobs, P.G., et al.: Incorporating an exercise detection, grading, and hormone dosing algorithm into the artificial pancreas using accelerometry and heart rate. J. Diab. Sci. Technol., 1932296815609371 (2015)Google Scholar
- 17.Kovatchev, B., et al.: Feasibility of long-term closed-loop control: a multicenter 6-month trial of 24/7 automated insulin delivery. Diab. Technol. Ther. (2017)Google Scholar
- 20.Lee, J.J., Gondhalekar, R., Doyle, F.J.: Design of an artificial pancreas using zone model predictive control with a moving horizon state estimator. In: 2014 IEEE 53rd Annual Conference on Decision and Control (CDC), pp. 6975–6980. IEEE (2014)Google Scholar
- 24.Paoletti, N., Liu, K.S., Smolka, S.A., Lin, S.: Data-driven robust control for type 1 diabetes under meal and exercise uncertainties. CoRR, 1707.02246 (2017)Google Scholar
- 26.Perea, L., How, J., Breger, L., Elosegui, P.: Nonlinearity in sensor fusion: divergence issues in EKF, modified truncated GSF, and UKF. In: AIAA Guidance, Navigation and Control Conference and Exhibit, p. 6514 (2007)Google Scholar
- 28.Resalat, N., El Youssef, J., Reddy, R., Jacobs, P.G.: Design of a dual-hormone model predictive control for artificial pancreas with exercise model. In: 2016 IEEE 38th Annual International Conference of the Engineering in Medicine and Biology Society (EMBC), pp. 2270–2273. IEEE (2016)Google Scholar
- 32.Van Der Merwe, R.: Sigma-point Kalman filters for probabilistic inference in dynamic state-space models. Ph.D. thesis, Oregon Health & Science University (2004)Google Scholar
- 35.Welch, G., Bishop, G.: An Introduction to the Kalman Filter. Technical report, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA (1995)Google Scholar