PPG and Force-Sensor Blood Pressure Estimation Without Cuffs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing non-invasive methods for estimating blood pressure, such as cuff-based and cuffless techniques, face challenges in accuracy and efficiency, particularly in providing reliable estimates without causing bodily harm.
Innovation Solution
A deep learning-based apparatus utilizing a photoplethysmogram (PPG) sensor and a force sensor, combined with a neural network model, to estimate blood pressure by dividing the pressure range into classes and calculating probability values based on PPG and force signals, employing techniques like convolutional and recurrent neural networks, and optimizing the model with cross-entropy and additive margin softmax loss functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If cuff-based methods are used to measure blood pressure, then measurement accuracy is improved, but the method causes bodily harm through cuff inflation and deflation
Solution Approach 1:
The patent extracts the harmful cuff component from the blood pressure measurement system and replaces it with a PPG sensor that contacts the skin without inflation. The measurement function is transferred from the mechanical cuff to the optical PPG sensor combined with force sensing, eliminating the harmful inflation/deflation cycle while preserving blood pressure estimation capability
Solution Approach 2:
The patent replaces the mechanical cuff-based oscillometric measurement system with an optical-mechanical hybrid system using PPG sensors and force sensors. The mechanical inflation/deflation process is substituted with optical detection of pulse waves combined with force signal analysis, eliminating bodily harm while maintaining measurement functionality
2Measurement precision
If deep learning models with multiple layers are used, then blood pressure estimation accuracy is improved, but device complexity increases
Solution Approach 1:
The patent segments the deep learning model into distinct functional modules: convolutional layers for feature extraction from PPG and force signals, pooling layers for dimensionality reduction, and dense layers for final prediction. This modular segmentation makes the complex model more manageable and interpretable while maintaining high accuracy
Solution Approach 2:
The patent transforms the input signals from time-series data into frequency-domain features through Fourier transforms and wavelet transforms, adding a frequency dimension to the analysis. This dimensional transformation enables the model to capture both temporal and spectral characteristics of pulse waves, improving accuracy without excessively increasing complexity
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
The apparatus provides accurate and non-invasive blood pressure estimation by minimizing prediction errors, offering real-time feedback through displays, audio, or haptic devices, suitable for various electronic devices including smartphones and wearables.
Implementation Method 1
a photoplethysmogram (PPG) sensor configured to measure a PPG signal from an object
Data Source
AI summary
An apparatus for estimating blood pressure includes: a photoplethysmogram (PPG) sensor configured to measure a PPG signal from an object; a force sensor configured to measure a force signal acting between the object and the PPG sensor; and a processor configured to (i) divide a predetermined blood pressure range into a plurality of classes, (ii) input the measured PPG signal and the measured force signal into a blood pressure estimation model to obtain the probability values for each of the classes, and (iii) estimate blood pressure based on the obtained probability values for the respective classes.