ECG PQRST complex detector and heart rate variability analysis using temporal characteristics of fiducial points

Abstract The detection accuracy of fiducial points related to the main waves of electrocardiogram (ECG)—the PQRST complex—considerably affects the heart rate variability (HRV) accuracy in detecting heart abnormalities or cardiac arrhythmia. However, the deformations of R-, P-, and T-waves make it difficult to detect each wave’s fiducial points. Herein, a ECG PQRST complex detector considering the deformation of main waves is presented for real-time ECG analysis, and HRV analysis is performed based on the detected fiducial points. It detects R-peak and makes a refined signal for detecting the PQ-ST fiducial points, utilizing the connecting line of the concave point (CLCCP) and the connecting line of the convex point (CLCVP) generated from the original signal. The QRS interval is first detected by applying a threshold to the sum of the difference between CLCCP and CLCVP. The refined signal for detecting the fiducial points of P- and T-waves is obtained by averaging the CLCVP and CLCCP for the non-QRS interval. The P- and T-points are detected using a symmetrical feature, and the associated onset and offset points are detected using an asymmetrical feature. The method achieves an R-detection performance sensitivity of 99.82%, a positive prediction of 99.81%, and a detection error rate of 0.36% against the MIT-DB and PQRST with the mean and standard deviation of 0.7 ± 7.4 ms and a root mean square error of 8.4 for QT-DB. Through detection simulation and HRV analysis, we confirm that the method accurately detects the fiducial points for deformed R, P, and T-waves.