Insulin Delivery Control via Segmented Feedback and Meal Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current insulin delivery systems for diabetic patients, such as the artificial pancreas, face challenges in stabilizing blood glucose levels due to physiological delays, saturation limits of insulin pumps, and the lack of reliable individual patient models, leading to limitations in feedback time constants and instability in glycemic control, especially during meal-related glycemic excursions.
Innovation Solution
A method for controlling insulin delivery that separates actions between feedback and meal compensation paths, using a linear model predictive controller and a Kalman filter to optimize insulin dosing based on predicted meal intake and continuous glucose monitoring data, while incorporating conventional therapeutic rules and empirical models to improve robustness and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a standard feedback control method is used to regulate glycemia, then the system can maintain stability under normal conditions, but it cannot respond quickly enough to meal-related glycemic excursions due to physiological delays and saturation limits
Solution Approach 1:
The control method is segmented into two independent paths: a feedback path that maintains stability under normal conditions, and a meal compensation path that provides rapid response to meal-related glycemic excursions. This segmentation allows each path to be optimized for its specific function without compromising the other.
Solution Approach 2:
The meal compensation path uses predictive models to estimate future glycemic excursions based on anticipated meals, allowing the system to prepare and deliver insulin proactively before the actual glycemic spike occurs, thereby improving response speed while maintaining stability through the separate feedback path.
2Measurement precision
If a complex individual patient model is used to improve accuracy of insulin dosing, then precision can be improved, but the system becomes less robust due to lack of reliable models and increased complexity
Solution Approach 1:
The system introduces an intermediary layer of predictive modeling that translates complex patient-specific physiological parameters into simplified control signals. The predictive model acts as a mediator between the complex patient physiology and the simple insulin delivery mechanism, maintaining precision without requiring the control system to directly handle complex individual models.
Solution Approach 2:
The system changes parameters from complex individual patient models to population-based predictive models with adjustable parameters. This allows the system to maintain precision by adapting to individual patients through parameter adjustment rather than through complex individual modeling, thereby reducing system complexity while preserving dosing accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method (400) for controlling the delivery of insulin in a diabetic patient (P) comprising steps of: providing data (d) representative of at least a fraction of a meal ( m(k + i) ) that the patient (P) will consume; providing, from a block (R) representative of conventional therapy or open loop rule that the patient (P) is subject to, based on the data (d) representative of at least a fraction of the foreseen meal (m(k + i) ) that the patient (P) will consume, a reference insulin value (u0); providing data representative of the difference between input data ( y), representative of a reference glycemic level, and feedback data (yCGM) representative of the glycemic level detected in the patient (P); providing, by a control module (301; 401), based on the data representative of the difference between the input data ( y ), representative of the reference glycemic level, and the feedback data (yCGM), representative of the glycemic level detected in the patient (P) and the data (d) representative of at least a fraction of the foreseen meal ( m(k + i) ) that the patient (P) will consume, data (uMPC) representative of a first insulin variation value; providing, based on the data (uMPC), representative of the first insulin variation value, and data (ub), representative of a basal insulin value, a value of insulin (i) to be delivered to the patient (P).