Complexity reduced blind multichannel equalization

Abstract

Recently, an algorithm for blind (without training sequence) and direct (without prior estimation of the channel response) estimation of (all) Zero-Forcing (ZF) equalizers of a Single Input Multiple Output (SIMO) channel has been proposed. These span a linear subspace, making it necessary to introduce a criterion to select the best equalizer within the so-estimated subspace. This selection step increases considerably the algorithm complexity. We propose two modifications to the algorithm that achieve the same (asymptotic) equalization performances at a reduced complexity.