Adaptive Artifact Elimination in Telecardiology Systems using Leaky LMS Variants
AsiyaSulthana1, Md. Zia Ur Rahman2

1Asiya Sulthana, Department of Electronics and Communication Engineering, K L University, Koneru Lakshmaiah Education Foundation, Green Fileds, Vaddeswaram, Guntur-522502, (Andhra Pradesh), India.
2Md Zia Ur Rahman, Department of Electronics and Communication Engineering, K L University, Koneru Lakshmaiah Education Foundation, Green Fileds, Vaddeswaram, Guntur-522502, (Andhra Pradesh), India.
Manuscript received on 02 June 2019 | Revised Manuscript received on 10 June 2019 | Manuscript published on 30 June 2019 | PP: 1478-1483 | Volume-8 Issue-8, June 2019 | Retrieval Number: H6629068819/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Evaluation of Electrocardiogram (ECG) facilitates the heart stroke volume in the sudden cardiac arrest. ECG is a noninvasive method for indirect analysis of stroke volume, monitoring the cardiac output and observing the hemodynamic parameters by changes in the blood volume of the body. Changes in the blood volume caused due to several physiological processes are extracted in the form of the impedance variations of the body segment. In the real time clinical environment ECG signals are contaminated with various artifacts. As these artifacts are not stationary in nature, we developed several hybrid adaptive filtering techniques to enhance the resolution of ECG signals. Least mean square (LMS) algorithm is the basic enhancement technique in the adaptive filtering. But, in the non-stationery environment the LMS algorithm suffers with low convergence rate and weight drift problems. In this paper we developed hybrid versions of LMS algorithm that is Normalized Leaky LMS (NLLMS) for ECG signal enhancement. More over to improve the rate of convergence, filtering capability and to minimize the computational complexity we also implement various sign versions of LLMS algorithms. The sign versions of NLLMS algorithms are sign regressor NLLMS (SRNLLMS), Sign NLLMS (SNLLMS), and Sign Sign NLLMS (SSNLLMS). Based on these adaptive algorithms, we developed several adaptive signal enhancement units (ASEUs) and performance is evaluated on the real ECG signal components obtained from MIT-BIT database. To ensure the ability of these algorithms, four experiments were performed to remove the various artifacts such as sinusoidal artifacts (SA), respiration artifacts (RA), muscle artifacts (MA) and electrode artifacts (EA). Among these techniques, the ASEU based on SRNLLMS performs better in the artifacts removing process. The signal to noise ratio improvement (SNRI) for this algorithm is calculated as 18.3165 dBs, 8.0964 dBs, 6.7025 dBs and 8.0825 dBs respectively for SA, RA, MA and EA. Hence, the SRLLMS based ASEUs are more suitable in ECG signal filtering in real time health care sensing systems.
Keyword: Adaptive filter, Artifacts, Electrocardiography, non-invasive, Signal enhancement.
Scope of the Article: Systems and Software Engineering.