Glucose Sensor Sensitivity Modeling for Faster Stabilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing glucose sensors often require a stabilization period of one to three hours before providing accurate readings, which delays the timely monitoring of blood glucose levels.
Innovation Solution
A model mosaic framework is employed to partition sensor electrical properties into subspaces and utilize machine learning models to predict glucose sensitivity, allowing for quicker stabilization and more accurate readings by determining whether to inhibit or utilize sensor data based on these predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a traditional glucose sensor is used, then accurate glucose readings can be obtained, but a stabilization period of one to three hours is required before accurate readings are provided
Solution Approach 1:
The patent segments the sensor response into multiple temporal components (fast component and slow component) and applies different machine learning models to each subspace. By partitioning the sensor data based on electrical property ranges, the system can quickly identify and utilize the fast response component for immediate glucose readings, eliminating the need to wait for the complete stabilization period while maintaining accuracy.
Solution Approach 2:
The patent employs machine learning models that are pre-trained to predict glucose sensitivity and stabilize sensor readings in advance. The system uses historical sensor data and electrical properties to preliminarily adjust and calibrate the sensor output, enabling accurate glucose measurements to be obtained immediately upon sensor insertion without requiring the traditional one to three hour stabilization period.
2Loss of time
If machine learning models are used to predict glucose sensitivity, then the stabilization period can be reduced, but the device complexity increases
Solution Approach 1:
The patent introduces machine learning models as intermediary components that mediate between the raw sensor electrical signals and the final glucose concentration output. These models act as intelligent translators that process sensor data in real-time, predicting glucose sensitivity and compensating for stabilization delays without requiring complex hardware modifications to the sensor itself.
Solution Approach 2:
The patent changes the operational parameters of the sensor system by dynamically adjusting the interpretation of electrical properties based on machine learning predictions. Instead of using fixed threshold values or simple calibration factors, the system continuously adapts sensitivity parameters based on real-time analysis of sensor electrical characteristics, enabling faster stabilization while managing computational complexity through efficient model selection.
Data Source
AI summary
Techniques for determining glucose sensitivity are provided. In some embodiments, the techniques may involve receiving sensor data relating to a sensor electrical property. The techniques may further involve determining a subspace of a plurality of subspaces of an input signal feature space based on a respective range of values associated with the sensor electrical property. The techniques may further involve selecting a machine learning model from a plurality of machine learning models associated with the subspace. The techniques may further involve determining a glucose sensitivity of a glucose sensor device based on the sensor data and the selected machine learning model. The techniques may further involve determining whether to inhibit or utilize glucose readings of the glucose sensor device based on the glucose sensitivity. The techniques may further involve operating the glucose sensor device based on the determination.