A sigmoid function based feedback Filtered-X-LmS algorithm with improved offline modelling

Abstract

As the world is a growing technological city, improvement on the current technology is given immense importance. This involves the improvement in the clarity of the signals used in an instrument; let it be in the form of text, audio or video. In short, the suppression of noise in any form is critical. Active Noise Control (ANC) has developed especially after the introduction of more powerful DSP chips in the 1980s and has become the most popular noise cancelling technique. Several structures such as the lattice ANC, feedforward ANC, feedback ANC and the Volterra filter based ANC have been developed. Based on the application and the type of noise signal, several algorithms have also been developed like the Least Mean Square (LMS), Normalized Least Mean Square (NLMS) and the Filtered-x- LMS (FXLMS) algorithm. In this paper, an improved sigmoid function based feedback FXLMS algorithm which has an improved offline modelling has been developed. A comparative study of the algorithm has been carried out with the conventional FXLMS algorithm. From the graphs obtained an analysis of the proposed algorithm has also been done which has an improved speed convergence over the conventional FXLMS algorithm.