Exploiting second order cyclostationarity or higher order statistics for the blind identification of mixed phase FIR systems
Dieter Boss, Michael Boe, Karl‐Dirk Kammeyer · 1995
This paper addresses the problem of blind identification of mixed phase time-invariant FIR systems. Two algorithms based on 2nd order statistics of the oversampled cyclostationary received signal are compared with an efficient approach exploiting 4th order statistics of the stationary received signal sampled at symbol rate. While the former suffer from systems which can not be identified from 2nd order cyclostationary statistics, the latter are said to require more samples of the received signal to obtain comparable levels of estimation variance. However, we show in this paper that the approach relying on 4th order statistics (EVI) delivers a superior estimation performance based on the same number of samples when it comes to the identification of realistic transmission channels in digital communication systems.