A Novel R Peak Detection Method for Mobile Environments

With the development of the portable electrocardiogram (ECG) sensor, R peaks can be monitored during various physical activities in mobile environment. However, such ECG signals contain real-world noise, complicating the accurate detection of R peak. In this paper, we propose a novel approach for R...

Full description

Saved in:
Bibliographic Details
Published in:IEEE access 2018-01, Vol.6, p.51227-51237
Main Authors: Lee, Miran, Park, Dajeong, Dong, Suh-Yeon, Youn, Inchan
Format: Article
Language:English
Subjects:
Band-pass filters
Cardiac stress tests
electrocardiogram
Electrocardiography
Filtering algorithms
Heart rate
Legged locomotion
Low pass filters
Noise
R peak detection
Shannon energy envelope
signal envelop filtering
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:With the development of the portable electrocardiogram (ECG) sensor, R peaks can be monitored during various physical activities in mobile environment. However, such ECG signals contain real-world noise, complicating the accurate detection of R peak. In this paper, we propose a novel approach for R peak detection in ECG signals with real-world noise by using signal envelop filtering (SEF) and Shannon energy envelope with the Savitzky-Golay filter (SEE-SG). SEF was designed to diminish abrupt peaks due to movement by adjusting its amplitude based on an automatically determined threshold. SEE-SG was used to enhance the QRS-complex. We first designed the experimental protocol for physical activity (sit, walk, and ascend), namely, normal ECG in a physical activity database, and validated our method by using the ECG data of 23 subjects. Moreover, to avoid overtuning on our database, we validated the proposed method using two public databases: the MIT-BIH QT database (QTDB) and Noise Stress Test Database (NSTDB). The experimental results show promising performance on QTDB (SE and P+: 99.96% and ER: 0.08%), NSTDB (SE: 94.07%, P+: 93.45, and ER: 12.57%), and NPADB (SE: 99.95%, P+: 99.78%, and ER: 0.27%). We also assessed our method by comparing the absolute differences of R-R intervals and heart rates between actual and detected R peaks. In conclusion, as the proposed method achieved noise robustness and reliable performance on R peak detection during physical activities, it can be used to monitor a user's ECG and cardiac health information accurately in daily life.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2018.2867329