The Realization of Adaptive Filter Based on LMS by Applying MATLAB
Abstract
Digital filter which is widely used is the important basement of the digital signal processing. Adaptive filter can adjust its coefficients automatically to minimize the mean-square error between its output and that of an unknown input, unnecessary to know the autocorrelation functions of signal and noise in advance, hence it can realize linearly optimum filtering.
Adaptive filter has many different algorithms such as RLS(Recursive Least Square) and LMS(Least Mean-Square) algorithms. Algorithm of adaptive LMS is a simple and very useful method which is used to estimate gradient. It has been widely applied in the signal processing areas.
This paper introduces the structure and the principle of the adaptive filter, and then introduces the LMS algorithm, at the same time, it has completed the design of adaptive LMS filter based on MATLAB platform and realized to decrease noise in signal processing.
Through the simulations, we realize the algorithm of adaptive LMS. Moreover, we can understand there exist two very important parameters, step-size and filter order M through the results. We also demonstrate that how the step-size affects the convergence rate toward the unknown system, and how the filter order M affects the performance of adaptive filter by changing these two parameters.
Keywords: Adaptive Filter MATLAB LMS Simulation
