Artifact Removal in Ecg Signals Using Modified Data Normalization Based Signal Enhancement Units for Health Care Monitoring Systems 1

Content Provider	Semantic Scholar
Author	Mantravadi, Nagesh Prasad, S. V. A. V. Rahman, Zia Ur
Copyright Year	2016
Abstract	Low complexity noise cancellation structures are needed for reliable transmission of ECG signals at real time environments. These low complexity structures can be developed with the help of the partial update techniques for better convergence and complexity. In this paper the same idea is used to derive several structures which are good at convergence and complexity. Based on partial update mechanism of the coefficients of the adaptive filter, we upgraded the conventional normalized least mean square (NLMS) algorithm. This modified algorithm updates only some coefficients of the taps where the signal characteristics widely deviate from the previous iteration. The modified NLMS (MNLMS) based on partial update mechanism is combined with signum based algorithms to minimize number of multiplications during filtering process. Further, we proposed maximum value of data for normalizing the step size to decrease the number of multiplications in the denominator of the normalization function. These are suitable to operate at high data rate applications, and to test the working of these structures in real time conditions the MIT-BIH arrhythmia database was used. Here the signal to noise ratio, the miss adjustment error is used as performance measures and all the test data is tabulated. The structures have shown good performance over the standard LMS algorithm in terms of the filtering, complexity and convergence.
File Format	PDF HTM / HTML
Alternate Webpage(s)	http://www.jatit.org/volumes/Vol93No2/30Vol93No2.pdf
Language	English
Access Restriction	Open
Content Type	Text
Resource Type	Article