Neural Blood Glucose Estimation for Real-Time Hypoglycemia Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing blood glucose monitoring systems face challenges in accurately and efficiently detecting hypoglycemic conditions due to intermittent measurements, and the computational workload for processing biosignals is high, requiring significant designer intervention.
Innovation Solution
A blood glucose measurement method using a learning model that automates the process of deriving glucose values by training the model with biometric, blood glucose-related, environmental, and health data, employing a neural network structure with multiple hidden layers and varied activation functions to improve accuracy and simplify device configuration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional biosignal processing methods are used, then blood glucose values can be obtained, but the computational workload is high and requires significant designer intervention
Solution Approach 1:
The patent replaces conventional mechanical signal processing methods (filtering, noise removal, feature extraction) with an artificial neural network-based learning model. The neural network automatically learns optimal processing patterns from training data, substituting manual algorithm design and reducing computational complexity while maintaining measurement accuracy.
Solution Approach 2:
The learning model performs self-training by automatically adjusting its internal parameters and weights during the training phase using labeled biosignal data. Once trained, the model independently processes new biosignals without requiring real-time designer intervention or manual configuration, enabling autonomous operation.
2Measurement precision
If conventional biosignal processing methods are used, then blood glucose values can be obtained, but the processing time and designer workload increase
Solution Approach 1:
The patent performs signal processing optimizations in advance during the training phase. The neural network learns optimal feature representations and processing patterns from extensive training data before actual measurement. This preliminary learning eliminates the need for complex real-time processing algorithms, reducing processing time during actual blood glucose measurement while maintaining accuracy.
3Device complexity
If intermittent blood glucose measurements are used, then device simplicity is maintained, but hypoglycemic conditions cannot be accurately detected
Solution Approach 1:
The patent enables continuous blood glucose monitoring by processing biosignals at frequent intervals using the trained neural network. The model efficiently handles continuous data streams, allowing real-time tracking of blood glucose levels and prompt detection of hypoglycemic conditions, transforming intermittent measurement limitations into continuous monitoring capability.
Data Source
AI summary
An embodiment may provide a blood glucose measurement method using a learning model, the method including obtaining a biosignal, generating biometric data from the biosignal, training a learning model with training biometric data, so as to output a blood glucose value, obtaining a blood glucose value corresponding to the biometric data from the biometric data via the learning model when training is completed, and providing, to a user, the blood glucose value corresponding to the biometric data.