Letters Analysis and Online Realization of the CCA Approach for Blind Source Separation

Abstract—A critical analysis of the canonical correlation analysis (CCA) approach in blind source separation (BSS) is provided. It is proved that by maximizing the autocorrelation functions of the recovered signals we can separate the source signals successfully. It is further shown that the CCA approach represents the same class of generalized eigenvalue decomposi-tion (GEVD) problems as the matrix pencil method. Finally, online real-izations of the CCA approach are discussed with a linear-predictor-based algorithm studied as an example. Index Terms—Blind source separation (BSS), canonical correlation anal-ysis (CCA), linear predictor, matrix pencil, second-order statistics (SOS). I