Cross-coupled signals separation and equalization by a mirror imaged adaptive decorrelation structure

An adaptive structure for cross-coupled signals blind separation and equalization is proposed. Cross-coupled signals might occur and observed in many MIMO applications. In this paper, the output observations of reverberant/non-reverberant simplified MIMO of 2-channel cross-coupled signals are considered. The structure is based on cross-coupled adaptive FIR cancellers for signals de-correlation and two other FIR equalizers that equalize the signals against the source signal cross-coupling channels distortion. Suitable de-correlation will result signals separation while equalization is devoted for signals reshaping. This structure is shown, under certain circumstances, to estimate the source signals, with no distortion even when the transmission coupling paths between the observation sensors have low attenuation and both signals are continuous. The main issues involved in this paper are the analysis of the structure topology, system constraint and convergence of the individual adaptive transfer function to solution(s) that ensures the signals separation and equalization for reverberant or non-reverberant channels. Also, suitable adopted LMS and least squares adaptive estimation algorithms for the structure are presented.