Predictive Control of Diabetic Glycemia

Diabetes Mellitus is a chronic disease, where the blood glucose concentration of the patient is elevated. This is either because of missing insulin production due to failure of the β-cells in the pancreas (Type 1) or because of reduced sensitivity of the cells in the body to insulin (Type 2). The therapy for Type 1 diabetic patients usually consists of insulin injections to substitute for the missing insulin. The decision about the amount of insulin to be taken has to be made by the patient, based on empirically developed rules of thumb. To help the patient with this task, advanced mathematical algorithms were used in this thesis to determine intakes of insulin and counteracting glucose that can bring the blood glucose concentration back to normoglycemia. The focus in this work was to determine insulin and glucose intakes around mealtimes. These algorithms used optimization methods together with predictions of the blood glucose concentration and mathematical models describing the patient dynamics to determine the insulin and glucose doses. For evaluation, the control algorithms were tested insilico using a virtual patient and are compared to a simple bolus calculator from the literature. The aim was to increase the time spent in the safe range of blood glucose values of 70 − 180 [mg/dL].

Grantor dc:publisher

Department of Automatic Control, Lund Institute of Technology, Lund University

Year dc:date

2013

Author dc:creator

• Stemmann, Meike

• Control Engineering
• Control
• Diabetes
• Prediction
• Optimization
• Decision Support System

Statement dc:rights

• info:eu-repo/semantics/openAccess

Language dc:language

eng

Repository record dc:identifier

https://lup.lub.lu.se/record/3567702

OAI identifier oai:identifier

oai:lup.lub.lu.se:3b3ef998-234f-4780-ab14-698c83ed2cb8