Design of Optimal Self-Regulation Mamdani-Type Fuzzy Inference Controller for Type I Diabetes Mellitus

Abstract

The control of blood glucose is a significant topic for Type I diabetic patient. The controller design for hypoglycaemia condition is forever challenging one. The main goal of this paper is to design an optimal fuzzy-PI controller based on Mamdani-type structure for blood glucose in diabetic patients. In the core of the proposed controller, a heuristic algorithm, namely particle swarm optimization with linearly decreasing weight, is utilized to optimize the membership functions, PI controller and insulin infusion signal simultaneously. To evaluate the performance of the proposed controller, it is tested by three different sets of parameters that relate to three different patients. With regard to the existing limitations in insulin injection rate, the proposed controller has minimized the insulin injection rate and time margin. The simulation results depict that the proposed controller has much better potential in terms of solution accuracy and better convergence speed in comparison with other methods despite existent uncertainties in parameters of model.