SEMG signal processing and analysis using wavelet transform and higher order statistics to characterize muscle force

Hussain, M. S. and Reaz, Mamun Ibn and Ibrahimy, Muhammad Ibn (2008) SEMG signal processing and analysis using wavelet transform and higher order statistics to characterize muscle force. In: 12th WSEAS International Conference on SYSTEMS, 22-24 July 2008, Heraklion, Greece.

PDF - Published Version Restricted to Repository staff only Download (298kB) | Request a copy

Abstract

An algorithm is proposed for processing and analyzing surface electromyography (SEMG) signals using wavelet transform and Higher Order Statistics (HOS). EMG signal acquires noise while travelling though different media. Wavelet denoising is performed in this research for initial EMG signal processing. With the appropriate choice of the Wavelet Function (WF), it is possible to remove interference noise effectively. Root Mean Square (RMS) difference and Signal to Noise Ratio (SNR) values are calculated to determine the most suitable WF. Results show that WF db2 performs denoising best among the other wavelets. Power spectrum analysis is performed to the denoised SEMG to indicate changes in muscle contraction. Furthermore, HOS method is applied for further efficient processing due to the unique properties of HOS applied to random time series. Gaussianity and linearity tests are conducted as part of HOS which shows that SEMG signal becomes less gaussian and more linear with increased force.

Item Type:	Conference or Workshop Item (Full Paper)
Additional Information:	4637/36225
Uncontrolled Keywords:	SEMG, wavelet transform, denoising, mean power frequency, HOS, bispectrum
Subjects:	T Technology > T Technology (General)
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Engineering > Department of Electrical and Computer Engineering
Depositing User:	Dr Muhammad Ibrahimy
Date Deposited:	03 Apr 2014 08:58
Last Modified:	16 Apr 2014 09:58
URI:	http://irep.iium.edu.my/id/eprint/36225

Actions (login required)

View Item