Development of peak detection algorithm using window variance for physiological variability monitoring

Abstract

The proposed invention provides a novel method for real-time detection and validation of systolic peaks in photoplethysmogram (PPG) signals using a window variance approach. This method utilizes a moving window of the PPG signal to compute variance and identify peaks, while eliminating false positives through a peak validation process based on consistent CCintervals (time between consecutive systolic peaks). The algorithm is designed to operate in real time, enabling continuous, accurate heart rate variability (HRV) monitoring with minimal computational requirements. The system is highly effective in noise-contaminated signals, demonstrating fewer than 0.16% false negatives, and no false positives, even when the PPG signal is corrupted with random noise. This makes it suitable for wearable devices and health monitoring systems that require robust peak detection, enabling precise tracking of cardiovascular health with immediate feedback. The simplicity and efficiency of the algorithm make it ideal for use in continuous healthcare applications, fitness monitoring, and telemedicine.

Specification

Description:The present invention relates to the detection and validation of peaks in physiological signals, particularly those obtained from photo-plethysmography (PPG) devices used in continuous health monitoring systems. More specifically, it proposes a novel method for detecting systolic peaks in real-time without requiring manual intervention, which is essential for the analysis of physiological variability. The invention introduces a window variance algorithm that uses a moving window approach to detect peaks in PPG signals with high accuracy while minimizing false positives and false negatives. The algorithm efficiently handles signal noise and artifacts by applying a peak validation process based on the consistency of the detected peaks' intervals. This method ensures robust and precise peak detection, even in the presence of motion artifacts or random noise, which is crucial for the continuous monitoring of autonomic function and heart rate variability. The invention is applicable in wearable health devices, rem , Claims:1. A method for real-time peak detection in a photoplethysmogram (PPG) signal, comprising acquiring the PPG signal, applying a moving window to compute the variance and maximum value, detecting peaks based on the variance, and validating the detected peaks based on CC-interval consistency between consecutive systolic peaks.
2. The method of claim 1, wherein the moving window comprises 120 samples with a 10-sample shift, and the variance is used to detect peaks by identifying maxima, which are validated using the peak validation process to minimize false positives and false negatives.
3. The method of claim 2, wherein the peak validation process includes confirming that the CC-interval between detected peaks is within a range of 0.7 to 1.3 times the average of previous CC-intervals, ensuring that only legitimate systolic peaks are retained and false positives are eliminated.
4. The method of claim 1, wherein the PPG signal is acquired usi

Documents