Blind channel equalization techniques using adaptive algorithms

Abstract

This paper presents a study of channel blind or autodidact equalization using adaptive algorithms that allow adequate
detection of signals that travel through the channel, mitigating the errors produced by intersymbol interference and noise. The
studied blind equalization method uses higher order statistics of the transmitted signal to calculate the error signal and thus
estimate the received data. The proposed algorithms used in blind equalization, generally known as self-taught algorithms that
offer better performance in real conditions, optimizing channel capacity, are shown in the development of the article. Among
them are the CMA (Constant modulus Algorithm) Godard (1980), Lucky’s direct decision algorithm (1966), Sato’s algorithm
(1975), the self-taught Least Mean Square algorithm, among others. From these algorithms the LMS is characterized, due to its
low computational cost, although it offers a relatively slow convergence speed and the CMA, which is computationally more
complex, nevertheless offers a data estimate relatively close to the transmitted symbols. Comparisons are also made with
respect to the speed of convergence and the degree of robustness in the estimation of errors for the simulated algorithms.