Method of efficiency increasing for independent components and convolurion analysis procedures that are used for signal form reconstruction on the base of mixture measuremens

The conception is considered that generate unified view-point on different algorithms for signal separation (including de-convolution of multi-channel mixtures) on the base of signal mixture measurements. Estimating functions method is considered that give the possibility to explain the structure of different adaptive algorithms that realize de-separating matrix estimation and signal form reconstruction using only mixture measurements. The problem of signal mixture separation is formulated on the base of semi-parametric statistical model and estimating functions family. In terms of Li group formulation there is filter manifold considered.