Blind source separation in the adaptive reduction of inter-channel interference for OFDM

Abstract

In this paper, we propose a new application of blind source separation (BSS) in OFDM systems with zero padding (ZP). We first focus on the model of ZP-OFDM to reduce it to an instantaneous linear mixing matrix model. Then we apply blind separation source (BSS) techniques based on the natural gradient (NG) and particularly EASI and M-EASI algorithms. Hence, we endow the method with adaptive properties and extend the use of BSS to complex symbols. This novel method blindly cancels the inter-channel interference (ICI) introduced by the channel with no use of cyclic prefix, meaning a power efficiency improvement and avoiding the recovery difficulties when some of the subcarriers are hit by a channel frequency response null. We include some experiments to illustrate the excellent performance of the proposed application even when ill-conditioned channels. © Springer-Verlag 2004.