![]() | Only 14 pages are availabe for public view |
Abstract An Electrocardiogram (ECG) signal is a recording of the electrical activity of heart. It is considered as an important source of vital diagnostic information. ECG signal is exposed to different types of noise. These noises change the nature of the ECG signal and provide difficulties on its analysis. The one long Least Mean Squares (LMS) adaptive filter is an algorithm used to reduce the noise effect on the ECG signal. This algorithm is widely used in adaptive filter applications due to its simplicity and low computational complexity, but it suffers from low convergence speed. This thesis tries to improve the one long LMS adaptive filter convergence speed using the multiple sub-adaptive filters. In the suggested algorithm, Simulation showed a saving in the required number of iterations by about 4.3*104 times compared to the one long LMS adaptive filter at MSE of 0.04. Also, in terms of Signal to Noise Ratio (SNR) against the step size (µ) a comparison between them is performed. It is found that the suggested algorithm provides improvement in the SNR by 5 dB at µ=0.2. The ECG samples are recorded from MIT-BIH database and an additive white Gaussian noise (AWGN) is added to the signal to examine the proposed technique and 2011a Mat-lab platform is used to simulate these results. |