| Abstract [eng] |
In this world everywhere, people are affected by diabetes millitus a disease which affects people of all ages. A report by the World Health organisation (WHO) shows that 180 million people are affected by diabetes mellitus, and this rate will increase upto 360 million people in 2030. The pancreas in the human body responsible formaintaining normoglycemia controls the range within 70-120 mg/dl. Beta cells produce the insulin in the pancreas when the blood glucose level is high. Insulin is a hormone that allows glucose to enter into the cells of the human body for providing energy. The conditon, Type 1 Diabetes Mellitus (T1DM) occurs because of the damage of pancreatic beta cells by the body's immune system. Hence, manually controlled insulin therapy is suggested by the physician where three or four times the blood glucose measurements to be taken by finger picking. Then, subcutaneous insulin injection is given based on the daily activities. The automatic blood glucose control is challenged by the disturbance of meal intake and internal system changes. To avoid the manually injecting insulin and to limit the significant changes in blood glucose concentration, an artificial pancreas (using Closed Loop Approach) is developed. The components of the artificial pancreas are glucose monitor, insulin pump, and control algorithm. This controlled framework is called Automatic Blood Glucose- Insulin (ABGI) regulatory system. This thesis proposesthree control strategies for closed loop artificial pancreas and has used three diabetic patient models for simulation that is designed to reduce the risks of hyperglycemia. The proposed control algorithm is implemented for the Bergmann Minimal Model (BMM) as the diabetic patient model. The control algorithm is proposed to develop a closed loop artificial pancreas. The conventional Proportional and Derivative (PD) controller and Internal Model control (IMC) are designed for the BMM model. Then, the PD and IMC control problem is optimized based on control problem to improve the performance of controller. To get optimal controller for the bergmaan minimal model (pso) particle swarm optimization algorithm is used, which should minimize the Integral Square Error (ISE).This research focus on the development of control algorithm using PSO based PD and PSO-IMC. The control problems are studied explicitly and the results show the improved performance of the proposed system. |
