A combined LMS-SOM algorithm for time-varying non-linear channel equalization

Author

Bouchired, Steven ; Ibnkahla, Mohamed ; Paquier, Williams

Author_Institution

Nat. Polytechnics Inst. of Toulouse, Toulouse, France

fYear

1998

fDate

8-11 Sept. 1998

Firstpage

1

Lastpage

4

Abstract

The paper proposes a self-organizing map (SOM) approach to equalize time-varying non-linear channels. The approach is applied to a satellite mobile communication channel which is composed of time invariant linear filters, a non-linear memoryless amplifier and a time-varying multi-path propagation channel. The SOM is combined to a Transversal Linear Equalizer (TLE) which is trained by the LMS algorithm. The paper studies the performance and convergence behavior of the SOM. The paper also illustrates the capability of the SOM to track the channel changes in particular for complicated modulation schemes.

Keywords

equalisers; least mean squares methods; mobile satellite communication; self-organising feature maps; time-varying channels; TLE; combined LMS-SOM algorithm; convergence behavior; least mean squares methods; modulation schemes; nonlinear memoryless amplifier; satellite mobile communication channel; self-organizing map approach; time invariant linear filters; time-varying multipath propagation channel; time-varying nonlinear channel equalization; transversal linear equalizer; Least squares approximations; Maximum likelihood detection; Mobile communication; Modulation; Neurons; Nonlinear distortion; Signal to noise ratio;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing Conference (EUSIPCO 1998), 9th European

Conference_Location

Rhodes

Print_ISBN

978-960-7620-06-4

Type

conf

Filename

7090125