An Increase of the Rate of Convergence for Blind Channel Estimation

Blind channel estimation, based on second order statistics, has recently been widely investigated in order to equalize the communication channel without resorting to training sequences. Unfortunately, blind estimation methods suffer from a slow rate of convergence. In this paper, we propose a novel approach which reduces the estimation error (and hence increases the rate of convergence) by suitable equalization of the channel's estimate. This method is based on an approximation of the solution of a linear prediction problem which requires an additional (blind) estimation of the autocorrelation function of the received signal. In order to demonstrate the performance of the proposed approach, blind channel estimation for OFDM systems is considered. Even if the necessary autocorrelation function of the received signal is estimated by means of simple time averaging, an increase of the rate of convergency can be achieved.