Subspace based blind identification and equalization of linear FIR MIMO systems and FIR SIMO volterra systems

In this paper a new algorithm based on subspace projections is developed for the blind equalization and kernel identification of linear FIR Multiple Input Multiple Output (MIMO) as well as nonlinear FIR SIMO Volterra systems. Simulations in the context of blind channel equalization show good performance in comparison to existing schemes.

1. Introduction

Blind methods are of great importance in digital signal communication systems as they allow channel identification/equalization at the receiver without the use of training signals. The topic of blind identification/equalization of linear time invariant (LTI) channels, both SIMO and MIMO, has drawn considerable attention over the past years and several algorithms have been developed (see [1]–[3] and references there in). Initially, focus on this area was centered in SOS methods, since compared to HOS methods, they require less data samples to obtain good statistical estimates. On the other hand, HOS methods provide system phase information, without requiring channel diversity, the ability to resolve matrix ambiguity to pure scaling and permutation indeterminacies. Finally, HOS are insensitive to additive Gaussian noise.