文档介绍：Blind Equalization of munication Channels
Using Recurrent Wavelet works*
Shichun He and Zhenya He, Felluw, IEEE
Department of Radio Engineering
Southeast University, Nanjing 2 10096, P. R. China
E-mail: sche@
Abstract This paper investigates the application of a have been developed for linear channels. The use of
Recurrent Wavelet work(RWN) to the these schemes with nonlinear unknown channels is
blind equalization of munication questionable.
channels We propose a RWNbased structure and a Blind equalization is however an inherently
novel training approach for blind equalization, and we nonlinear problem and it is desired to incorporate some
evaluate its performance puter simulations for nonlinearity in the equalizer structure. A Recurrent
munication channel model, It is shown Wavelet work(RWNN) being essentially an
that the RWN blind equalizer performs much better IIR nonlinear filter, can be trained to have desired
than the linear CMand the RRBF blind equalizers in dynamical behavior, using a stochastic gradient
nonlinear channel case. The small size and high approach via the Real Time Recurrent Learning")
performance ofthe RWN equalizer make it suitable algorithm. In paper we propose the use of a RWNN
for high speed channel blind equalkation. equalizer for the blind equalization of nonlinear
I. Introduction channels. A novel training ap