A genetic algorithm tuned optimal controller for glucose regulation in type 1 diabetic subjects

Area of Science:

• Biomedical Engineering
• Control Systems
• Endocrinology

Background:

• Type 1 diabetes requires careful glucose monitoring and insulin management.
• Meal intake significantly impacts blood glucose levels.
• Existing control systems face challenges with noisy data and model inaccuracies.

Purpose of the Study:

• To design an optimal state feedback controller for minimizing post-meal hyperglycemia in Type 1 diabetes.
• To achieve minimal insulin infusion while maintaining normoglycemia.
• To address limitations of noisy measurements, modeling errors, and sensor delays.

Main Methods:

• Utilized linear quadratic regulator (LQR) theory for controller design.
• Employed Kalman filter for state estimation from noisy data.
• Integrated a physiological model with a meal dynamics simulation model.
• Optimized controller and filter parameters using genetic algorithms.

Main Results:

• The developed controller successfully maintained normoglycemia across varying meal carbohydrate content.
• The system demonstrated robustness against noisy output measurements.
• The approach effectively handled modeling errors and sensor measurement delays.

Conclusions:

• An integrated control framework combining evolutionary computing and state-estimated LQR is effective for Type 1 diabetes management.
• This method offers a robust solution for real-time glucose control in diabetic subjects.
• The optimized controller shows promise for improving quality of life for individuals with Type 1 diabetes.