Chattering-free blood glucose level control based on ANFIS

Abstract

In the medical field determination of appropriate rate of insulin injection in order to stabilize the blood glucose to a normal level is vital for diabetics. In this paper, an adaptive neuro-fuzzy inference system (ANFIS) based on hybrid blood glucose control data set has been presented. Hybrid blood glucose control employs combination of the fuzzy logic controller optimized by genetic algorithm with well-known Palumbo control method to regulate the blood glucose level in type-1 diabetic mellitus (T1DM) patients. Due to the complexity of the hybrid controller and nonlinear and delayed nature of glucose-insulin mechanism as well as chattering phenomenon, the artificial intelligence based technique, especially the ANFIS method, is proposed in this paper. Finally, the simulation results of the fuzzy control, fuzzy-genetic control, Palumbo control and hybrid control are compared to the new proposed ANFIS control, which indicates the proper functioning of the proposed controller for tracking of desired blood glucose level at the lowest possible chattering error.