Artificial intelligent pancreas for type 1 diabetic patients using adaptive type 3 fuzzy fault tolerant predictive control

In this paper, the design methodology of artificial intelligent pancreas is presented. Accurate regulation of blood glucose levels in type 1 diabetic patients is of great importance in the presence of possible faults caused by sensor measurements. Regulation of blood glucose levels using a type 3 fuzzy predictive controller in type 1 diabetic patients in the presence of sensor faults is considered. The proposed structure includes a main control structure and a virtual dynamic, in which the main structure includes a fuzzy identifier, predictive controller, and an adaptive compensator, and the virtual structure is used to identify the sensor faults. Glucose is unknown in the dynamics of type 1 diabetes and is estimated on-line using a type 3 fuzzy system. Also, Lyapunov stability analysis is used to design the adaptive compensator to ensure the stability of the closed-loop system. The proposed methodology is evaluated based on Bergman's minimum model for different patients under various parametric uncertainties and disturbances.