Heart Rate Estimation via State Sequence Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Photoplethysmography (PPG) based heart rate estimation during motion is challenging due to motion artifacts, which complicate the identification of the correct heart rate from the PPG signal's frequency spectrum, especially when multiple peaks are present within a realistic heart rate range.
Innovation Solution
A heart rate monitor system that employs state sequence tracking, computing transition pair branch metric values and determining the heart rate estimate based on a maximum path metric, using digital sample values and frequency domain amplitudes to filter out motion artifacts and accurately estimate heart rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If frequency domain analysis is used to estimate heart rate by selecting the highest peak, then the estimation process is simple and fast, but the accuracy deteriorates when multiple peaks are present due to motion artifacts
Solution Approach 1:
The patent introduces an intermediary selection process that evaluates multiple frequency peaks using criteria beyond simple amplitude magnitude. This intermediary step considers physiological plausibility, peak prominence, and spectral characteristics to mediate between the simple peak selection approach and accurate heart rate determination, resolving the contradiction by adding intelligent filtering without excessive complexity
Solution Approach 2:
The patent changes the selection parameters from单一的amplitude magnitude to a multi-parameter evaluation system that includes peak prominence, spectral width, and physiological range constraints. This parameter transformation allows the system to distinguish true heart rate peaks from motion artifact peaks, improving accuracy while maintaining computational efficiency through optimized parameter thresholds
2Measurement precision
If motion compensation techniques are applied to remove motion artifacts, then signal fidelity is improved, but device complexity increases due to additional sensors and processing
Solution Approach 1:
The patent applies self-service by using the PPG signal's own spectral characteristics and temporal patterns to identify and filter motion artifacts without requiring external motion sensors. The system serves itself by detecting artifact signatures within the PPG data and applying appropriate filtering, thereby improving signal fidelity while avoiding the complexity of additional accelerometer or gyroscope hardware
Solution Approach 2:
The patent extracts and removes motion artifact components from the PPG signal by identifying their distinctive spectral features and separating them from the true heart rate signal. This extraction approach selectively eliminates harmful motion-related frequency components while preserving the physiological signal, achieving improved fidelity without the need for complex multi-sensor fusion systems
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 system effectively filters out motion artifacts and accurately determines heart rate by leveraging state sequence optimization, even in the presence of multiple frequency domain peaks, providing a reliable heart rate estimation.
Implementation Method 1
A photodiode or other optical sensor generates the PPG signal indicating the measured light transmission or reflection
Data Source
AI summary
Disclosed examples include heart rate monitor systems and methods to estimate a patient heart rate or rate of another pulsed signal, in which rate hypotheses or states, are identified for a current time window according to digital sample values of the pulsed signal for a current sample time window, and a rate change value is computed for potential rate transitions between states of the current and previous time windows. Transition pair branch metric values are computed as a function of the rate change value and a frequency domain amplitude of the corresponding rate hypothesis for the current time window, and the pulsed signal rate estimate is determined according to a maximum path metric computed according to the branch metric value and a corresponding path metric value for the previous time window.