Blind signal separation in distributed space-time coding systems using the FastICA algorithm

Abstract

One of the main advantages of cooperative communication systems is the use of information at the surrounding nodes in order to create spatial diversity and so far obtaining higher throughput and reliability. We propose in this paper a blind detector that involves the formulation of the system as a Blind Source Separation BSS problem. In the BSS framework, we do not have to estimate the channel using training data, removing the necessity of pilot symbols and the prior estimation of the channel. We analyze two kinds of distributed space-time codes for the single relay system, showing that they can be stated in terms of BSS as a linear instantaneous mixture of complex-valued sources. The BSS method applied is the complex version of the FastICA algorithm since it is very flexible, robust and the convergence is very fast so we can estimate the symbols accurately with a low-complexity algorithm that can adapt to changes in the channel with relative simplicity.