Non-Invasive Blood Glucose Estimation via PPG and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional invasive methods for measuring glycated hemoglobin (HbA1c) are burdensome and provide inaccurate results, especially in cases of short red blood cell lifespan, pregnancy, or kidney disease, necessitating a non-invasive and more accurate estimation method.
Innovation Solution
A method and apparatus utilizing machine learning to estimate glycated hemoglobin or blood glucose by collecting and analyzing bio-signals, specifically PPG signals, through a signal collection stage, feature extraction stage, and machine learning model construction, incorporating features like Zero-Crossing Rate, Power Spectral Density, and external factors like Body Mass Index, to generate accurate glycated hemoglobin or blood glucose estimates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If invasive blood collection method is used to measure glycated hemoglobin, then measurement can be performed, but patient burden increases and measurement accuracy decreases in certain conditions
Solution Approach 1:
The patent replaces the mechanical invasive blood collection system with an optical measurement system using photoplethysmography (PPG). The PPG sensor uses light absorption and scattering properties of blood to detect glycated hemoglobin levels non-invasively, eliminating needles and blood collection while maintaining measurement capability through optical-biological interactions.
Solution Approach 2:
The patent introduces photoplethysmographic signals as an intermediary between the physical blood composition and the measurement device. Instead of directly contacting blood, the system uses PPG waveforms that reflect blood's optical properties as a mediator to infer glycated hemoglobin levels, enabling indirect but accurate measurement.
2Measurement precision
If conventional invasive measurement method is used, then glycated hemoglobin level can be obtained, but the method provides inaccurate results in cases of short red blood cell lifespan, pregnancy, or kidney disease
Solution Approach 1:
The patent changes the measurement parameter from direct hemoglobin concentration (invasive) to photoplethysmographic waveform characteristics (optical properties). By measuring how blood absorbs and scatters light at different wavelengths and analyzing waveform features, the system obtains glycated hemoglobin levels that are not dependent on red blood cell lifespan or kidney function, improving accuracy across diverse patient populations.
Solution Approach 2:
The patent creates a universal measurement system that functions across all patient populations including those with short red blood cell lifespan, pregnancy, and kidney disease. The PPG-based method measures fundamental optical properties of blood that remain valid regardless of these physiological variations, making the system universally applicable where invasive methods fail.
3Measurement precision
If machine learning model with multiple features is constructed, then estimation accuracy improves, but model complexity increases
Solution Approach 1:
The patent segments the machine learning model into distinct functional modules: feature extraction module that processes PPG waveforms, feature selection module that identifies relevant parameters, and prediction module that generates estimates. This modular segmentation manages complexity by organizing the model into manageable, independent components that can be developed and validated separately.
Solution Approach 2:
The patent performs preliminary feature extraction and selection before the main prediction process. By pre-processing the PPG signals to extract relevant waveform characteristics and pre-selecting important features through analysis, the system reduces the dimensionality and complexity of data fed into the prediction model, improving both training efficiency and estimation accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables non-invasive, accurate estimation of glycated hemoglobin or blood glucose, reducing the burden on patients and improving measurement accuracy across various demographics, thereby aiding in diabetes management.
Implementation Method 1
measuring PPG signals based on a change in intensity of the transmitted light or the reflected light
Data Source
AI summary
The present disclosure relates to a method and apparatus for non-invasive estimation of glycated hemoglobin (HbA1c) or blood glucose by using machine learning, the method comprising: a sig nal collection stage of collecting a bio-signal of a measurement subject to be measured; a feature extraction stage of extracting a plurality of features from the bio-signal; a machine learning model construction stage of constructing a machine learning model for estimating glycated hemoglobin or blood glucose by learning training data including the plurality of features; and a glycated hemoglobin/blood glucose estimation stage of generating input data on the basis of the bio-signal extracted from the measurement subject being measured and inputting the input data to the machine learning model, so as to estimate glycated hemoglobin or blood glucose of the measurement subject being measured.