Feature extraction of vigilance level based on Heart Rate Variability of Electrocardiogram

This paper presented a simple method of detecting the peak of R wave in Electrocardiogram (ECG) signal and computing the Heart Rate Variability (HRV). Features were extracted from the obtained HRV to analyze the vigilance level during day time short nap sleep. Firstly, the second derivative of ECG signal was calculated and clustered in order to detect the peak of R wave. Secondly, HRV was calculated from the subtraction of two adjoining R waves. Finally, several features were extracted based on HRV in time domain and frequency domain. The variation of vigilance level during the day time short nap was analyzed based on the extracted features. The result showed that the accuracy of peak detection of R wave in ECG signal was about 99.5%. The extracted features can reflect the changing of vigilance level and be useful for sleep stage analysis.