Frequency Domain Filter Implementation For Input Containing Outlier

This paper presents an overview of frequency domain LMS(Least Mean Square) algorithm for an input containing outlier.Signal processing have tendency to suppress the noise while leaving the signals unchanged by using system or filter.In signal processing the unwanted noise in the signal was removed or suppressed from the signal using adaptive filter. The filter coefficient of adaptive filter is capable of adjusting automatically to match the desired signal through an optimization algorithms . various algorithms have been implemented to recover the desired signal from the noise using different optimization technique such as Least mean square (LMS) and its variants and Recursive least square (RLS) etc.The convergence performance of LMS type adaptive filters is totally dependent on the correlation and on the eigenvalue of the input data.To enhance convergence peformance and reduce the complexity of a signal various transform domain adaptive filters have been developed . This proposed work presents an overview of the frequency domain LMS algorithm for system identification, for input containing outlier. Frequency domain LMS adaptive filter is investigated for input data using Discrete Fourier transformation (DFT) technique. Compared to the traditional adaptive filter the convergence speed and performance analysis of the proposed method shows better performance in terms of Mean Square Error (MSE), Excess Mean Square Error (EMSE) and Mean Square Deviation (MSD) with input containing 10 percent outlier.