Performance analysis of LMS based algorithms used for impulsive noise cancellation

Abstract

In real world application noise can degrade the performance of the system. It is affected by impulsive noise, as this type of noise highly depends on physical environment. Though adaptive filters are widely used for noise cancellation, still it requires robust parameter for all type of noise. In this paper, an attempt is taken for reduction of impulsive noise. Initially, the popular LMS algorithm is used for comparison and extends towards Fx-LMS. It shows significant result, but becomes unstable the due to the second order moment does not exist for Gaussian process. Further it has been analyzed with NLMS and the norm is exploited. It is found that WLMS algorithm performs better than all the above algorithms and also robust for impulsive noise. The results show its comparison performance.