Personalized IMC-PID Controller for Artificial Pancreas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current artificial pancreas systems face challenges in maintaining blood glucose levels within the euglycemic zone for individuals with type 1 diabetes due to inherent delays in subcutaneous sensing and insulin delivery, and existing control algorithms may induce hypoglycemic risk if based on irrelevant models.
Innovation Solution
An internal model-based proportional-integral-derivative (IMC-PID) controller with an insulin feedback scheme, personalized using a lower order discrete model incorporating a subject's basal insulin characteristics and total daily insulin, is developed to adjust controller aggressiveness based on individual insulin sensitivity, combined with a model-predictive controller (MPC) for improved glucose regulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a standardized control algorithm is used for artificial pancreas systems, then the device complexity is reduced and ease of manufacture is improved, but the controller cannot account for individual variations in insulin sensitivity leading to hypoglycemic risk
Solution Approach 1:
The patent applies local quality by personalizing the controller gains for each subject based on their individual insulin sensitivity characteristics. The controller uses subject-specific parameters (basal insulin rate, total daily insulin, insulin sensitivity index) to customize the control action, ensuring that each patient receives tailored glucose control rather than a one-size-fits-all approach. This resolves the contradiction by maintaining standardized controller structure while implementing personalized control parameters.
Solution Approach 2:
The patent changes the controller parameters (proportional, integral, and derivative gains) based on subject-specific clinical characteristics. The controller gains are adjusted according to individual insulin sensitivity, basal insulin requirements, and total daily insulin doses. This parameter customization allows the same controller structure to adapt to different patients, improving safety without increasing fundamental device complexity.
2Measurement precision
If a higher order model is used to capture complex glucose-insulin dynamics, then the measurement precision and control accuracy are improved, but the device complexity and computational burden increase
Solution Approach 1:
The patent extracts only the essential dynamic characteristics of glucose-insulin metabolism that are necessary for effective control. Rather than using a complex high-order model, the invention identifies and utilizes key parameters (insulin sensitivity, basal rate, total daily insulin) that capture the dominant behavior of the system. This extraction approach maintains control accuracy while significantly reducing model complexity and computational requirements.
Solution Approach 2:
The patent employs a dynamic model that adapts to individual subject characteristics through personalized parameter estimation. The model dynamically adjusts controller gains based on subject-specific insulin pharmacokinetics and pharmacodynamics. This dynamic personalization approach achieves high measurement precision and control accuracy using a computationally efficient model structure suitable for real-time implementation.
3Productivity
If the controller aggressiveness is increased to rapidly correct hyperglycemia, then the productivity of glucose regulation is improved, but the object-affected harmful factors increase due to hypoglycemia risk
Solution Approach 1:
The patent dynamically adjusts the controller aggressiveness by changing the proportional, integral, and derivative gains based on the subject's insulin sensitivity and current glucose state. For subjects with high insulin sensitivity, the controller uses more conservative gains to prevent hypoglycemia, while for insulin-resistant subjects, higher gains are appropriate for effective glucose control. This parameter adaptation resolves the contradiction by tailoring the aggressiveness to individual patient characteristics.
Solution Approach 2:
The patent implements feedback mechanisms that continuously monitor glucose levels and adjust controller action accordingly. The feedback loop uses real-time glucose measurements and subject-specific parameters to modulate insulin delivery, preventing overly aggressive corrections that could lead to hypoglycemia. The feedback system ensures that productivity is optimized while maintaining safety by adapting control strength to current physiological conditions.
Data Source
AI summary
A model-based control scheme consisting of either a proportional-integral-derivative (IMC-PID) controller or a model predictive controller (MPC), with an insulin feedback (IFB) scheme personalized based on a priori subject characteristics and comprising a lower order control-relevant model to obtain PID or MPC controller for artificial pancreas (AP) applications.