Blind equalization-combinations of Bussgang- and higher-order-statistics-based methods

A new generation of higher-order-statistics (HOS) based blind equalization techniques, like the eigenvector algorithm (EVA) published in 1997 by B. Jelonnek, D. Boss and K.-D. Kammeyer, leads to a much better performance. In this paper these techniques are used to initialize a conventionally Bussgang algorithm. The purpose is to show, through simulations, that combinations of both techniques have the ability to improve the equalization quality and to reduce the computational complexity as well. Also, we consider the modifications of these algorithms for the application of a fractionally-spaced (FTS) equalizer.