Drowsiness detection based on wavelet analysis of ECG and pulse signals

Author

Aihua Zhang ; Fenghua Liu

Author_Institution

Coll. of Electr. & Inf. Eng., Lanzhou Univ. of Technol., Lanzhou, China

fYear

2012

fDate

16-18 Oct. 2012

Firstpage

491

Lastpage

495

Abstract

The purpose of this study is to explore the impact of drowsiness state on ECG and pulse signals and seek a more convenient and effective method for detecting drowsiness. Different frequency bands of ECG and pulse signals are selected to calculate the wavelet packet energy and wavelet entropy based on wavelet analysis. The results show that the wavelet packet energy of ECG whose frequency ranges are from 7.8Hz to 23.4Hz and from 23.4Hz to 62.5Hz respectively, and the wavelet entropy of pulse signal whose frequency range is from 0.1Hz to 31.25Hz are significantly decreased (p<;0.01) in the drowsiness state compared with that in the waking state. The accuracy rate of classification for these three features can reach 100% by using Support Vector Machines (SVM).

Keywords

electrocardiography; feature extraction; medical signal processing; signal classification; sleep; support vector machines; wavelet transforms; ECG; SVM; drowsiness detection; drowsiness state; feature classification; frequency 23.4 Hz to 62.5 Hz; frequency 7.8 Hz to 23.4 Hz; frequency bands; pulse signals; support vector machines; waking state; wavelet analysis; wavelet entropy; wavelet packet energy; ECG; drowsiness detection; pulse signal; support vector machine (SVM); wavelet analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Biomedical Engineering and Informatics (BMEI), 2012 5th International Conference on

Conference_Location

Chongqing

Print_ISBN

978-1-4673-1183-0

Type

conf

DOI

10.1109/BMEI.2012.6513058

Filename

6513058