Optimization of PIDD2-FLC for blood glucose level using particle swarm optimization with linearly decreasing weight. (May 2020)

• Record Type: Journal Article
• Authors: Jaradat, Mohammad A.; Sawaqed, Laith S.; Alzgool, Mohammad M.
• Is Part Of: Biomedical signal processing and control. Volume 59(2020)
• Publication Date: 2020-05
• Subjects: Proportional-Integral-Differential plus second order derivative Fuzzy Logic Controller -- PIDD2-FLC -- Glucose insulin regulatory system -- Treatment model -- Diabetes -- Particle swarm optimization; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2020.101922
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: A Proportional-Integral-Differential plus second order derivative Fuzzy Logic Controller (PIDD 2 -FLC) is developed. The new controller is deployed for blood glucose level (BGL) in type I diabetic subjects. Particle Swarm Optimization with Linearly Decreasing Weight (LDW-PSO) is used to optimize the intelligent fuzzy controller. Comparisons are held between the optimized controller and other types of controllers taking into consideration normal and abnormal conditions. Abstract: In this paper, a Proportional-Integral-Differential plus second order derivative Fuzzy Logic Controller (PIDD 2 -FLC) is implemented aiming to maintain blood glucose level (BGL) normal in type I diabetic subjects. In type I diabetes, insulin-secreting cells are destroyed and, hence, patient depends on external insulin to maintain BGL. The suggested controller is responsible for driving a micro-pump that injects the diabetic patient with proper insulin dose, and a continuous BGL sensor is used for feedback. The nonlinear patient model used is the two-delay differential model, and the reference model for BGLs considered as in the treatment model planned using the two-delay differential model with oscillatory behavior. Particle Swarm Optimization with Linearly Decreasing Weight (LDW-PSO) algorithm is used to optimize the proposed controllers to match the reference model performance. Finally, a comparison is held between the optimized controller with other FLC controllers structures taking into consideration the normal and abnormal conditions. PIDD 2 -FLC with seven linguistic fuzzy membership functions (MFs) was found to have the best performance overall under different examining conditions.