臺灣博碩士論文加值系統

為了讓「人工胰臟」能夠實際應用於不需臥病在床的糖尿病患者，發展一個自動調節胰島素注射速率的閉環路控制策略是必要的。所以，本論文以模式預測控制器 (MPC) 來維持第一型糖尿病患者的血糖濃度在其正常範圍，其中MCP利用一個預測模式估計未來的血糖濃度以決定下一步的胰島素注射速率，因此，該預測模式對於MPC血糖控制的效能有密切關聯。文獻中已提出許多用來描述血糖-胰島素動態的生理模式，這些生理模式都具有非線性動態，但用來描述第一型糖尿病的經驗模式則相對較少，然而一般來說，經驗模式在控制器的設計上會比較簡單，故經驗模式比生理模式更適合用於控制器的設計上，再者，由於非線性經驗模式通常會比線性經驗模式更能夠準確地描述血糖-胰島素之非線性動態，因此，本論文使用方塊導向非線性模式，如Hammerstein模式、Wiener模式來對於血糖-胰島素動態進行建模並提出其識別方法，分別以兩種生理模式：Bergman模式、Sorensen模式所模擬得到的血糖-胰島素動態資料來建立其方塊導向非線性模式，並且探討Hammerstein模式、Wiener模式用於血糖預測之準確度。根據方塊導向非線性模式，可設計一個簡單的非線性模式預測控制架構來進行血糖控制，而使用方塊導向模式的主要優點為控制器只需針對線性系統部分進行設計，大幅降低控制器設計與計算之複雜度，由餐後血糖控制的模擬結果發現搭配Wiener模式的MPC控制器可維持血糖處於正常範圍且控制效能優於搭配線性模式的MPC控制器，最後說明本論文所提方法在實際應用的可能性。

In order for an “artificial pancreas” to become practical for ambulatory use, a closes-loop control strategy for automated insulin delivery must be developed. In this thesis, model predictive controller (MPC) is considered for this purpose to regulating the blood glucose in subjects with Type I diabetes. The MPC utilizes a model to estimate the future glucose concentration based on a series insulin infusion rates, so that this model is very crucial for the performance of glucose control. Many physiological models have been proposed to describe the glucose-insulin dynamics which are nonlinear in nature. On the other hand, empirical models for Type I diabetes have been far less prevalent in the literature. However, empirical models are more attractive than physiological models in the sense that they are often much simpler and thus suitable for control design. In addition, nonlinear empirical models are expected to describe the nonlinear glucose-insulin dynamics better than the linear one. Therefore, this study uses block-oriented nonlinear models, i.e., Hammerstein and Wiener systems, for modeling the glucose-insulin dynamics. Algorithms for the identification of block-oriented nonlinear models based on the simulation data from physiological models are presented. Two physiological models, Bergman minimal model and Sorensen model, have been considered for these simulations. The predictive capabilities of glucose concentration using Hammerstein and Wiener models are investigated. Based on the nonlinear model, a nonlinear MPC scheme can be designed for blood glucose control. The main advantage of using block-oriented nonlinear model is that the controllers can be designed if a linear system is being controlled. Simulation results show that the Wiener model based MPC can keep blood glucose level safely within normoglycemic ranges after a meal disturbance, and can have better control performance than that of linear model based MPC. Finally, the possibility for practical application of the proposed method is also demonstrated.

中文摘要 i
英文摘要 iii
誌謝 v
目錄 vi
表目錄 viii
圖目錄 ix
第一章 緒論 1
1.1 糖尿病介紹 1
1.2 文獻回顧 3
1.3 研究動機 5
1.4 章節安排 6
第二章 第一型糖尿病之血糖生理模式 8
2.1 生理模式簡介 8
2.2 Bergman模式 9
2.3 Sorensen模式 12
第三章 血糖－胰島素動態之非線性建模 19
3.1 方塊導向(block-oriented)非線性模式簡介 19
3.2 模式識別方法 20
3.2.1. Hammerstein系統 21
3.2.2. Wiener系統 23
3.3 Bergman模式之識別結果與討論 26
3.3.1. Hammerstein模式 27
3.3.2. Wiener模式 36
3.3.3. 線性模式 48
3.3.4. 比較三種經驗模式的血糖預測能力 52
3.4 Sorensen模式之識別結果與討論 57
3.4.1. Hammerstein模式 59
3.4.2. Wiener模式 67
3.4.3. 線性模式 80
3.4.4. 比較三種經驗模式的血糖預測能力 85
第四章 血糖模式預測控制 90
4.1 模式預測控制簡介 90
4.2 以非線性方塊導向模式為基礎之控制架構 90
4.3 血糖模式預測控制結果與討論 92
4.3.1. Bergman模式 96
4.3.2. Sorensen模式 96
4.4 血糖控制之實際應用策略 99
第五章 結論與未來展望 107
5.1 結論 107
5.2 未來展望 108
參考文獻 109
符號彙編 111

[1]
Wild, S.; Roglic, G.; Green, A.; Sicree, R.; King, H. Global Prevalence of Diabetes: Estimates for the year 2000 and projections for 2030. Diabetes Care. 2004, 27, 1047-1053.
[2]
歐陽振宇，糖尿病治療全書，臺北市：憲業企管，2008。
[3]
Gurny, R.; Junginger, H.; Peppas, N. A. Pulsatile Drug Delivery: Current Applications and Future Trends. Wissenschaftliche: Stuttgart, 1993.
[4]
Parker, R. S.; Doyle III, J. F.; Harting, J. E.; Peppas, N. A. Model predictive control for infusion pump insulin delivery. The 18th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Amsterdam, The Netherlands, 1996, 5, 1822-1823.
[5]
Wilinska, M. E.; Hovorka, R. Simulation models for in silico testing of closed-loop glucose controllers in type 1 diabetes. Drug Discovery Today: Disease Models. 2008, 5, 289-298.
[6]
Bergman, R. N.; Phillips, L. S.; Cobelli, C. Physiologic evaluation of factors controlling glucose tolerance in man: measurement of insulin sensitivity and beta-cell glucose sensitivity from the response to intravenous glucose. The Journal of Clinical Investigation. 1981, 68, 1456-1467.
[7]
Sorensen, J. T. A physiologic model of glucose metabolism in man and its use to design and assess improved insulin therapies for diabetes. Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, U. S., 1985.
[8]
Hovorka, R.; Canonico, V.; Chassin, L. J.; Haueter, U.; Massi-Benedetti, M.; Federici, M. O.; Pieber, T. R.; Schaller, H. C.; Schaupp, L.; Vering, T.; Wilinska, M. E. Nonlinear model predictive control of glucose concentration in subjects with type 1 diabetes. Physiological Measurement. 2004, 25, 905-920.
[9]
Fabietti, P. G.; Canonico, V.; Federici, M. O.; Benedetti, M. M.; Sarti, E. Control oriented model of insulin and glucose dynamics in type 1 diabetics. Medical and Biological Engineering and Computing. 2006, 44, 69-78.
[10]
Yasini, S.; Karimpour, A.; Naghibi-Sistani, M. B.; Ghareh, S. An Automatic Insulin Infusion System based on H-Infinity Control Technique. The 2nd International Conference on Bioinformatics and Biomedical Engineering. Cairo, Egypt, 2008, 1-5.
[11]
Lynch, S. M.; Bequette, B. W. Model predictive control of blood glucose in type I diabetics using subcutaneous glucose measurements. Proceedings of the 2002 American Control Conference. Anchorage, Alaska, USA, 2002, 5, 4039-4043.
[12]
Jeng, J. C.; Huang, H. P. Nonparametric identification for control of MIMO Hammerstein systems. Industrial and Engineering Chemistry Research. 2008, 47, 6640-6647.
[13]
陳永論，第ㄧ型糖尿病之血糖動態模式與胰島素用藥策略研究，碩士論文，國立臺灣大學化學工程研究所，臺北市，2007。
[14]
Fisher, M. E. A semiclosed-loop algorithm for the control of blood glucose levels in diabetics. IEEE Transactions on Biomedical Engineering. 1991, 38, 57-61.
[15]
Farmer, T. G.; Edgar, T. F.; Peppas, N. A. Effectiveness of intravenous infusion algorithms for glucose control in diabetic patients using different Simulation models. Industrial and Engineering Chemistry Research. 2009, 48, 4402-4414.
[16]
Lehmann, E.; Deutsch, T. A physiological model of glucose-insulin interaction in type 1 diabetes mellitus. Journal of Biomedical Engineering. 1992, 14, 235-242.
[17]
Ruiz-Velázquez, E.; Femat, R.; Campos-Delgado, D. U. Blood glucose control for type I diabetes mellitus: A robust tracking H∞ problem. Control Engineering Practice. 2004, 12, 1179-1195.
[18]
Parker, R. S.; Doyle III, F. J.; Ward, J. H.; Peppas, N. A. Robust H∞ glucose control in diabetes using a physiological model. AICHE Journal. 2000, 46, 2537-2546.