Title :
On a new multilayer blind equalizer
Author :
Destro Filho, J.B. ; Favier, Génard ; Romano, João M Travassos
Author_Institution :
DECOM, Univ. of Campinas, Brazil
fDate :
6/21/1905 12:00:00 AM
Abstract :
This paper proposes a new multilayer blind equalizer, which consists of a multilayer perceptron (MLP) with an adjustable output activation function. Learning is based on a backpropagation algorithm. Parameters tuning is discussed in detail and simulations are provided in order to compare the convergence time of the new proposition with those of the classical MLP and RBF blind equalizers
Keywords :
backpropagation; blind equalisers; convergence of numerical methods; land mobile radio; multilayer perceptrons; telecommunication computing; MLP; RBF blind equalizers; adjustable output activation function; backpropagation algorithm; convergence time; mobile communication systems; multilayer blind equalizer; multilayer perceptron; parameters tuning; simulations; Blind equalizers; Cost function; Data communication; Filtering; Finite impulse response filter; Least squares approximation; Neurons; Nonhomogeneous media; Stochastic processes;
Conference_Titel :
Global Telecommunications Conference, 1999. GLOBECOM '99
Conference_Location :
Rio de Janeireo
Print_ISBN :
0-7803-5796-5
DOI :
10.1109/GLOCOM.1999.827582
