HRV Data Preprocessing with Variance and RR Interval Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional HRV signal data sources contain abnormal data due to issues like device precision deficiencies and emotional or movement-related fluctuations, leading to inaccurate feature extraction and disease prediction.
Innovation Solution
An HRV-data preprocessing method that filters HRV data using variance and RR interval ranges to exclude abnormal data, followed by amplitude adjustment and anomaly detection to ensure data accuracy, and extracts time-, frequency-, and nonlinear features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature extraction is directly performed on raw HRV data, then the processing speed is fast, but the accuracy of extracted features is low due to abnormal data and interference
Solution Approach 1:
The patent segments the HRV data processing into distinct stages: raw data acquisition, quality assessment (calculating signal quality index SQI and noise level NL), abnormal data identification, and feature extraction. This segmentation allows systematic filtering of abnormal data points before feature extraction, improving feature accuracy without creating an overly complex monolithic process. Each segment handles a specific aspect of data quality control.
Solution Approach 2:
The patent performs preliminary quality assessment and abnormal data identification before feature extraction. By calculating SQI and NL metrics in advance and identifying abnormal data points beforehand, the system prepares clean data for subsequent feature extraction. This preliminary action ensures that only quality-assured data undergoes feature extraction, significantly improving the accuracy of extracted features.
2Reliability
If conventional filtering methods are used on HRV data, then the processing is simple, but abnormal data due to device precision and emotional/movement fluctuations cannot be effectively removed
Solution Approach 1:
The patent changes the approach from simple amplitude-based filtering to a multi-parameter quality assessment system. It calculates signal quality index (SQI) combining multiple factors, evaluates noise levels (NL) through spectral analysis, and uses RR interval variability metrics. These parameter changes enable comprehensive identification of abnormal data caused by device precision issues, emotional fluctuations, and movement artifacts, significantly improving HRV data reliability.
Solution Approach 2:
The patent introduces intermediate quality metrics (SQI, NL, RR interval analysis) as mediators between raw HRV data and final feature extraction. These intermediary assessments act as filters that evaluate data quality without directly modifying the original signal, allowing systematic identification and removal of abnormal data points while preserving valid physiological information.
3Measurement precision
If multiple rounds of filtering are performed on HRV data, then the data accuracy is improved, but the processing time increases
Solution Approach 1:
The patent applies partial filtering by focusing computational resources on identifying and removing only the most critical abnormal data points rather than performing exhaustive multiple-round filtering. The quality assessment calculates key metrics (SQI, NL) and identifies abnormal segments, then removes only those specific problematic portions. This partial action achieves sufficient data accuracy without the time cost of repeated comprehensive filtering rounds.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
The present disclosure relates to an HRV-data preprocessing method and apparatus, and an electronic device. The method includes: obtaining HRV data corresponding to a sliding window; determining a variance of a plurality of target peaks corresponding to the HRV data, and determining HRV data having the variance that falls within a predetermined variance range as first target HRV data; determining an RR interval sequence of the first target HRV data based on the plurality of target peaks, determining whether each RR interval in the RR interval sequence falls within a predetermined heartbeat time interval range, and determining second target HRV data; and extracting a time-domain feature, a frequency-domain feature, and a nonlinear feature of the second target HRV data. By filtering out abnormal data twice, the second target HRV data with high accuracy and in line with a human heartbeat characteristic is obtained, enabling an extraction of precise HRV features.