Electrocardiogram signal processing method for exact Heart Rate detection in physical activity monitoring system: Wavelet approach

Abstract

Physical Activity Monitoring is a device that can measure the human activity quantity quantitatively through Heart Rate detection in real time. R-Spike detection of ECG is required for this Heart Rate detection. Since Physical Activity Monitoring System is usually used during activity or exercise, however, signal measured in ECG System is contaminated by diverse noises. Diverse noises become the factors of failure in R-Spike detection. Such factors impede the exact HR detection. This paper suggests method to convolute wavelet function and scaling function as the optimum signal disposition method for optimum R-Spike detection. This method was compared with the R-Spike detection method that uses quadratic spline wavelet presented before. To verify performance of signal disposition method suggested in this paper, the ECG of noise stress test database (NSTDB) and MIT-Database were tested in combination. Then, the sensitivity of R-Spike detection rate for noise was also additionally tested by gradually lowering SNR of NSTDB. Then, it was verified through ECG signal that was actually measured in physical activity monitoring.