Improved Multi-stage Clustering Based Blind Equalisation in Distributed Environments

Abstract

The recently proposed improved multi-stage clustering (IMSC) based blind equalisation algorithm in [1] gave significant performance improvement as compared to its state of the art counterparts. In that work, the performance was considered over a frequency-selective single input single output (SISO) additive white Gaussian noise (AWGN) channel. The practice of relaying is used in cooperative communications so as to give a variety of the independent signals to the receiver to choose from, the choice being dependent on the quality of the link. In other words, this results in a diversity gain at the receiver. In this paper, we propose a novel blind equalisation scheme which accepts inputs from relays, and finds a smart way of blindly fusing the incoming data, so as to reach a lower mean square deviation (MSD) from the Weiner solution. The simulations presented in this paper validate our algorithm. We also derive an expression for MSD from the Weiner solution of this algorithm as a function of step-size as in [2]. We find that it closely matches the experimentally obtained curves.