Equalization of satellite UMTS channels using RBF networks

Author

Bouchired, Steven ; Ibnkahla, Mohamed ; Roviras, Daniel ; Castanié, Francis

Author_Institution

Nat. Polytech. Inst. of Toulouse, France

Volume

3

fYear

1998

fDate

8-11 Sep 1998

Firstpage

1250

Abstract

The paper proposes a radial basis function (RBF) neural network (NN) approach to satellite universal mobile telecommunication system (S-UMTS) channel equalization. Two main problems arise in S-UMTS communications: (i) non-linear distortions which are caused by the onboard nonlinear power amplifiers, and (ii) multipath propagation. This paper shows that RBF networks are particularly well suited to overcome these problems. The paper presents several simulation results which show that RBF networks outperform classical equalization techniques for different configurations of the mobile channel (mobile speed, noise level, number of paths, etc.)

Keywords

equalisers; error statistics; fading channels; mobile satellite communication; multipath channels; nonlinear distortion; power amplifiers; radial basis function networks; radiofrequency amplifiers; telecommunication computing; BER performance; MLP equalizer; RBF networks; S-UMTS channel equalization; mobile channel; mobile speed; multipath fading; multipath propagation; neural network; noise level; nonlinear distortions; onboard nonlinear power amplifiers; radial basis function; satellite UMTS channels equalization; satellite universal mobile telecommunication system; simulation results; 3G mobile communication; Artificial satellites; Bandwidth; Equalizers; Mobile communication; Neural networks; Neurons; Noise level; Nonlinear filters; Radial basis function networks;

fLanguage

English

Publisher

ieee

Conference_Titel

Personal, Indoor and Mobile Radio Communications, 1998. The Ninth IEEE International Symposium on

Conference_Location

Boston, MA

Print_ISBN

0-7803-4872-9

Type

conf

DOI

10.1109/PIMRC.1998.731386

Filename

731386