Separation theorem for independent subspace analysis and its consequences
Pattern Recognition
Hi-index | 0.00 |
This paper addresses the blind identification of multiple input multiple output (MIMO) systems with the number of inputs strictly less then the number of outputs. In contrast to the standard FIR modelling we assume that the overall channel has an arbitrary finite order rational transfer function. Certain quite reasonable technical hypotheses allow one to adapt the existing linear prediction and subspace based approach and to implement a finite order zero-forcing equalizer (ZFE) in the noise-free case. The noise-free condition also yields a simple performance analysis which is quite accurate at low noise levels and provides a meaningful comparison of the proposed estimators. The robustness to additive noise is studied by computer simulations for both techniques. We focus on the deconvolution performance, i.e the residual ISI at the output of ZFE.