Closed Loop Blood Glucose Control Algorithm Testing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Developing a closed loop blood glucose control algorithm is challenging due to noise, delays, and individual variability in lifestyle and physiology, making traditional testing methods expensive, time-consuming, and risky, or insufficiently representative.
Innovation Solution
A method that processes continuous glucose data to generate a hypothetical uncontrolled blood glucose excursion, which can be used as a test input to aid in developing, testing, or tuning a closed loop blood glucose control algorithm, allowing for customization to individual requirements without risking human subjects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If testing is performed on living diabetic subjects, then the testing captures real physiological complexity, but the testing becomes expensive, time-consuming, and poses significant health risks
Solution Approach 1:
The patent creates a computational model that copies and simulates diabetic physiology instead of using living subjects. The model reproduces key physiological characteristics including glucose-insulin dynamics, noise, and delays, allowing safe yet realistic algorithm testing without exposing human or animal subjects to health risks
Solution Approach 2:
The patent replaces the mechanical/biological testing system (living subjects) with a computational/mathematical system. The physiological processes are modeled through equations and algorithms that simulate glucose metabolism, insulin action, and sensor behavior, substituting physical experimentation with virtual simulation
2Productivity
If testing is performed with mathematical models, then the testing is fast and inexpensive, but human physiology is too complex to be sufficiently represented
Solution Approach 1:
The patent employs adjustable parameters within the computational model that can be tuned to match individual patient characteristics such as insulin sensitivity, glucose production rates, and sensor noise levels. This allows the model to adapt to different physiological scenarios while maintaining computational efficiency for rapid testing
Solution Approach 2:
The patent implements dynamic behavior in the computational model through time-varying parameters and stochastic elements that simulate the unpredictable nature of physiology. The model includes random noise components and delay elements that capture the dynamic, non-stationary characteristics of blood glucose regulation
3Reliability
If the control algorithm is personalized to each individual user, then the control effectiveness is improved, but the development and testing complexity increases
Solution Approach 1:
The patent performs preliminary characterization of individual patient physiology through the computational model before actual control implementation. By pre-simulating various scenarios and tuning parameters based on patient-specific data, the system prepares personalized control parameters in advance, reducing the complexity of real-time customization
Data Source
AI summary
Methods and devices to generate a tool for testing, simulating and/or modifying a closed loop control algorithm are provided. Embodiments include receiving glucose data for a predetermined time period, determining a variation in the glucose level based on the received glucose data, filtering a received glucose data based on the determined variation, substituting a negative change in the glucose data value with a predetermined value to generate a sequence of modified glucose values, and integrating the sequence of modified glucose values to determine an uncontrolled blood glucose excursion condition.