<p>Maintaining healthy blood glucose levels is a major challenge in diabetes management due to several physiological complexities. Persistent hyperglycemia in diabetic patients is often due to insulin resistance, <InlineEquation ID="IEq1"> <EquationSource Format="TEX">$\beta $</EquationSource> <EquationSource Format="MATHML"><math> <mi>β</mi> </math></EquationSource> </InlineEquation> cell dysfunction, and delays in insulin secretion, absorption, and action. In addition, insulin is a substrate for the insulin-degrading enzyme (IDE), which plays a crucial role in regulating both insulin and glucagon degradation. Studies involving tissue-specific deletion of IDE in liver and pancreatic <InlineEquation ID="IEq2"> <EquationSource Format="TEX">$\beta $</EquationSource> <EquationSource Format="MATHML"><math> <mi>β</mi> </math></EquationSource> </InlineEquation> cells highlight the critical role of IDE in insulin secretion and liver insulin sensitivity. To capture these biological dynamics and time-dependent effects, we adopted a nonlinear delay differential equation model, which incorporates the inhibitory behavior of IDE and reflects temporal disturbances in glucose–insulin interactions. This study presents the development of a fuzzy logic-based artificial pancreas system to regulate blood glucose levels in individuals with Type 1 diabetes, particularly under conditions of high meal variability. The proposed insulin advisory system employs an optimized Mamdani-type fuzzy logic controller to determine insulin dosages in real time, based on continuous feedback from a blood glucose level sensor. The controller operates a micro-pump to deliver insulin doses adaptively, aiming to maintain glucose concentrations within the normoglycemic range (90–120 mg/dL), thus preventing acute hyperglycemia and reducing the risk of long-term complications. To evaluate the effectiveness of the controller, it was tested in five physiologically relevant scenarios: (i) no meal consumption, (ii) multiple scheduled meals, (iii) irregular meal patterns, (iv) sensor noise and disturbances, and (v) uncertainties in model parameters. In all cases, the controller successfully reduced ultradian glucose oscillations and maintained glucose levels to the target range within two hours, demonstrating robustness and adaptability in various metabolic conditions. The proposed fuzzy logic-based framework provides a generalized approach for handling nonlinear, time-delay systems with inherent biological uncertainties in patients. The fuzzy controller is model-free, and the rule-based design enables real-time, adaptive control without relying on exact mathematical representations. This integration of physiological modeling with intelligent control addresses a critical gap in automated diabetes care and contributes to the development of safe, stable, and personalized artificial pancreas systems.</p>

A Fuzzy Logic-Based Artificial Pancreas System for Managing Type 1 Diabetes with High-Meal Variability

期刊论文
发表日期：

2026年5月25日

查看全文

Ankit Sharma,
Nilam

摘要

Maintaining healthy blood glucose levels is a major challenge in diabetes management due to several physiological complexities. Persistent hyperglycemia in diabetic patients is often due to insulin resistance, $\beta $ $β$ cell dysfunction, and delays in insulin secretion, absorption, and action. In addition, insulin is a substrate for the insulin-degrading enzyme (IDE), which plays a crucial role in regulating both insulin and glucagon degradation. Studies involving tissue-specific deletion of IDE in liver and pancreatic $\beta $ $β$ cells highlight the critical role of IDE in insulin secretion and liver insulin sensitivity. To capture these biological dynamics and time-dependent effects, we adopted a nonlinear delay differential equation model, which incorporates the inhibitory behavior of IDE and reflects temporal disturbances in glucose–insulin interactions. This study presents the development of a fuzzy logic-based artificial pancreas system to regulate blood glucose levels in individuals with Type 1 diabetes, particularly under conditions of high meal variability. The proposed insulin advisory system employs an optimized Mamdani-type fuzzy logic controller to determine insulin dosages in real time, based on continuous feedback from a blood glucose level sensor. The controller operates a micro-pump to deliver insulin doses adaptively, aiming to maintain glucose concentrations within the normoglycemic range (90–120 mg/dL), thus preventing acute hyperglycemia and reducing the risk of long-term complications. To evaluate the effectiveness of the controller, it was tested in five physiologically relevant scenarios: (i) no meal consumption, (ii) multiple scheduled meals, (iii) irregular meal patterns, (iv) sensor noise and disturbances, and (v) uncertainties in model parameters. In all cases, the controller successfully reduced ultradian glucose oscillations and maintained glucose levels to the target range within two hours, demonstrating robustness and adaptability in various metabolic conditions. The proposed fuzzy logic-based framework provides a generalized approach for handling nonlinear, time-delay systems with inherent biological uncertainties in patients. The fuzzy controller is model-free, and the rule-based design enables real-time, adaptive control without relying on exact mathematical representations. This integration of physiological modeling with intelligent control addresses a critical gap in automated diabetes care and contributes to the development of safe, stable, and personalized artificial pancreas systems.