Enhancement of the Blood Glucose Level for Diabetic Patients Based on an Adaptive Auto-Tuned PID Controller via Meta-Heuristic Methods.

This work proposes an intelligent method that uses an adaptive auto-tuned proportional-integral-derivative (PID) controller to track and regulate the blood glucose level of diabetic patients. The suggested controller seeks to provide the optimal insulin control action, which is in charge of swiftly, precisely, and accurately managing the blood glucose level. To train this controller, two meta-heuristic techniques are employed. The first technique is the particle swarm optimization (PSO), which has been widely used in both data estimation and training because of its quick computing speed, while the second one is an intrusion detection technique called grey wolf optimization (GWO), which was created to categorize data and effectively find multiple intrusions. The mean square error performance index is used in the two distinct meta-heuristic algorithm types to determine and optimize the optimal or nearly optimal gain parameters of the adaptive PID controller. The results of the MATLAB simulations for three different patients demonstrated the efficacy and resilience of the proposed control algorithm in tracking the dynamic behaviour of the diabetic patients' blood glucose levels by minimizing overshoot in the transient state to zero value, maintaining the steady-state blood glucose level in the normal physiological level of (60-120) mg/dl. These characteristics were particularly evident when we added a meal as a disturbance effect. Moreover, the comparison results showed that the proposed PID-GWO and PID-PSO algorithms enhanced the time (96 and 88) minutes to reach the blood glucose level at a normal physiological level by 4% and 12%, respectively, when compared to the fractional order PID and fuzzy logic control algorithms that the blood glucose level reached at a normal physiological level at 100 minutes and improved the time by 20% and 27%, respectively, when compared to the type-2 fuzzy control algorithm that the blood glucose level reached at a normal level at 120 minutes. In particular, the blood glucose level was kept at the desired normal physiological level without any oscillation. [ABSTRACT FROM AUTHOR]