Sparse blind identification and separation by using adaptive K-orthodrome clustering

We propose a new algorithm for identifying a mixing (basis) matrix AA knowing only sensor (data) matrix XX for linear model X=AS+EX=AS+E, under some weak or relaxed conditions, expressed in terms of sparsity of latent (hidden) components represented by the unknown matrix SS. We present a simple and efficient adaptive algorithm for such identification and illustrate its performance by estimation of the unknown mixing matrix AA and source signals (sparse components) represented by rows of the matrix SS. The main feature of the proposed algorithm is its adaptivity to changing (non-stationary) environment and robustness with respect to outliers that do not necessarily satisfy sparseness conditions.