Analysis of FXLMS algorithm for active noise control

Abstract

With the increased number of industrial equipment and automation, acoustic noise problem become more and more evident, which has many negative effects on human health. This work is concerned with analysis of filtered - XLMS (FXLMS) algorithm for active noise control. To enhance the convergency properties of the FXLMS algorithm, adaptive combination of two FXLMS filters with different step sizes and escalator type adaptive filtering with F-XLMS algorithms are proposed. Moreover, adaptive combination algorithm performance is enhanced by applying discrete cosine transformation(DCT) followed by power normalization. It is concluded in this study that two proposed algorithms; combination and escalator type for active noise control significantly improves the convergency properties of traditional FXLMS algorithm. This study not only gives a theoratical analysis for FXLMS algorithm, but also provides a brief experimental study and measurements realized in real time environment by using a Field Programmable Gate Array(FPGA).