Closed-loop insulin delivery with sensor error model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current closed-loop systems for managing insulin delivery in diabetes patients face challenges due to inaccuracies in glucose monitoring, leading to risks of hypoglycemia and hyperglycemia, as they rely on mathematical models with random data rather than actual sensor performance, and lack automation for continuous glucose monitoring and controlled insulin delivery.
Innovation Solution
A system that uses a glucose sensor to provide measurement signals, an insulin delivery device, and a controller programmed with a glucose measurement error model derived from actual sensor data, employing model predictive control to adjust insulin delivery based on real-time glucose levels, weight, daily insulin dose, and basal insulin profile, while incorporating safety checks to prevent over- or under-dosing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If mathematical models with random data are used for glucose monitoring, then the system can operate without actual sensor data, but the accuracy of glucose measurement deteriorates leading to risks of hypoglycemia andhyperglycemia
Solution Approach 1:
The patent creates a virtual copy of the glucose sensor system through mathematical modeling. The in silico sensor replicates the functionality of a physical sensor by using mathematical models that simulate glucose measurement behavior, allowing the closed-loop system to operate with virtual sensor data that mirrors real sensor characteristics without requiring actual physical sensors.
Solution Approach 2:
The patent transforms the glucose measurement process by changing from physical sensor parameters to mathematical model parameters. The system uses adjustable mathematical parameters to represent sensor behavior, enabling flexible simulation of different sensor conditions and improving measurement accuracy through parameter optimization rather than relying on random data.
2Device complexity
If patient intervention is required to calculate and control insulin delivery, then the system can use simple monitoring devices, but the system cannot provide continuous control during periods when the patient is unable to intervene
Solution Approach 1:
The patent implements self-service through the automated closed-loop control system. The system automatically calculates insulin delivery requirements and executes delivery without patient intervention. The controller continuously monitors glucose levels and autonomously adjusts insulin delivery, enabling the system to provide continuous control even when the patient is sleeping or unable to intervene.
Solution Approach 2:
The patent establishes a closed-loop feedback system where glucose sensor data continuously feeds back to the controller, which then adjusts insulin delivery accordingly. This feedback mechanism enables automatic continuous control by constantly monitoring glucose levels and making real-time adjustments to insulin delivery without requiring patient intervention.
3Extent of automation
If integrated automated system is implemented for continuous glucose monitoring and insulin delivery, then continuous control is achieved, but the device complexity increases
Solution Approach 1:
The patent merges multiple functions into an integrated system. The glucose sensor, controller, and insulin delivery device are combined into a unified closed-loop system that automatically performs continuous glucose monitoring and insulin delivery. This merging reduces the need for separate devices and manual coordination, managing complexity through integration rather than proliferation of separate components.
Data Source
AI summary
A closed-loop system for insulin infusion overnight uses a model predictive control algorithm (“MPC”). Used with the MPC is a glucose measurement error model which was derived from actual glucose sensor error data. That sensor error data included both a sensor artifacts component, including dropouts, and a persistent error component, including calibration error, all of which was obtained experimentally from living subjects. The MPC algorithm advised on insulin infusion every fifteen minutes. Sensor glucose input to the MPC was obtained by combining model-calculated, noise-free interstitial glucose with experimentally-derived transient and persistent sensor artifacts associated with the FreeStyle Navigator® Continuous Glucose Monitor System (“FSN”). The incidence of severe and significant hypoglycemia reduced 2300- and 200-fold, respectively, during simulated overnight closed-loop control with the MPC algorithm using the glucose measurement error model suggesting that the continuous glucose monitoring technologies facilitate safe closed-loop insulin delivery.