A Closed Form Solution for Whitenning Based Second Order Blind Identification

Abstract

This paper proposes specific analytical formulas that involve second order statistics for the blind identification of a two input two output system. Starting from a closed form solution of the blind identification of the system exploiting the temporal coherence properties of the input sources, we present a new algorithm CWSOBI (Closed Form Solution for Whitenning Based Second Order Blind Identification) to compute a zero forcing second order blind separator with low complexity. The new solution differs from an earlier proposed one ASOBI (Analytical Solution for Blind Identification) method in that it does not requires prior knowledge or estimation of the noise variance. Numerical experiments are provided to assess the performance of the proposed algorithm versus signal to noise ratio (SNR) and sample size. Theses Computer simulations demonstrate the efficiency of the proposed method.