Diabetes Mellitus indicates to a group of diseases that affect how body uses blood glucose. As an important source of energy for the cells that make up muscles and tissues, glucose is crucial to human body since it is the main source of fuel for every cell in the body. Chronic diabetes conditions include type 1 diabetes mellitus (T1DM) and type 2 diabetes mellitus (T2DM). Diabetes symptoms vary depending on how much your blood sugar is elevated. Some people, especially those with prediabetes or type 2 diabetes, may not experience symptoms initially. In type 1 diabetes, symptoms tend to come on quickly and be more severe. Those who diagnosed with T1DM have little to no insulin secretion from their pancreas in order to reduce high amount of glucose from their blood. Therefore, they need to inject insulin externally. Failure of maintaining blood glucose level (BGL) in appropriate amount can lead serious health problems such as kidney failure, stroke, cardiovascular diseases, nervous system failures.
During the past decade, researchers have developed mathematical models to have a better understanding of insulin-glucose homeostasis. Based on the mathematical models, a control algorithm can be developed to manage BGL within a normal range. This study aims to develop a simple and accurate adaptive control algorithm by using Artificial Neural Network (ANN) based Model Predictive Control (NN-MPC) to perform proper insulin injections according to BGL in T1DM rats. Two recognized mathematical models, the Lombarte model and the Bergman model, are used in this study, however; Wilinska’s subcutaneous insulin kinetic model is applied in those two models to represent T1DM properties since those two models were applied on health rats. The modified models can characterize the glucose-insulin homeostasis in T1DM rats based on inputs including BGL data, insulin and glucose injection. Particle swarm optimization is applied to estimate parameters for the two models.
Experimental data are collected from two T1DM rats. For each subject, parameters for the mathematical models are identified. The NN-MPC is developed based on the identified models and tested by a simulation study. The results show that the NN-MPC controls BGL of T1DM subjects effectively, 67% of the within the normal range with a mean absolute deviation of 31 mg/dl from the reference BGL of 125 mg/dl during 20-hour simulations . The proposed NN-MPC can be exploited as the control algorithm for Artificial Pancreas Systems (APS) to improve the quality of life of T1DM patients.
|Advisor:||Ko, Hoo Sang|
|Commitee:||Lee, H. Felix, Kwon, Guim|
|School:||Southern Illinois University at Edwardsville|
|School Location:||United States -- Illinois|
|Source:||MAI 81/2(E), Masters Abstracts International|
|Keywords:||Diabetes Mellitus, Blood glucose, Insulin-glucose homeostasis|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be