Delayed adaptive filtering based on NLMS algorithm with an error adjustment procedure

Author

Somjai, Puttapong ; Chumchu, Prawit ; Demeechai, Tanee

Author_Institution

Dept. of Telecommun. Eng, Mahanakorn Univ. of Technol., Bangkok, Thailand

fYear

1998

fDate

24-27 Nov 1998

Firstpage

323

Lastpage

325

Abstract

In delayed adaptive filtering, the customary tool is the delayed least-mean-square (DLMS) algorithm. However, in the DLMS algorithm, selection of suitable step-size is always done by trial and error. This paper proposes an alternative to the DLMS algorithm which avoids this drawback. The proposed algorithm is formulated using a normalized least-mean-square (NLMS) algorithm. The customized step-size is simply unity and an error adjustment procedure is used. In addition, we modify the proposed algorithm using a quantizing method to reduce the number of multiplications and divisions per input sample (MADPS). The simulated behavior is shown to agree well with theoretical prediction

Keywords

adaptive filters; adaptive signal processing; least mean squares methods; quantisation (signal); MADPS; NLMS algorithm; delayed adaptive filtering; divisions; error adjustment procedure; multiplications; quantizing method; Adaptive filters; Computer errors; Convergence; Delay estimation; Equations; Error correction; Filtering algorithms; Least squares approximation; Signal processing algorithms; Transversal filters;

fLanguage

English

Publisher

ieee

Conference_Titel

Circuits and Systems, 1998. IEEE APCCAS 1998. The 1998 IEEE Asia-Pacific Conference on

Conference_Location

Chiangmai

Print_ISBN

0-7803-5146-0

Type

conf

DOI

10.1109/APCCAS.1998.743761

Filename

743761