Blind Source Extraction for Ill-Conditioned Dynamic Multi-input Multi-output Channels

摘要

This paper presents a new approach to blind source extraction from convolutive mixtures in multi-input multi-output (MIMO) channels in two ill-conditioned cases: the number is less than that of sensors and the number is larger than or equal to that of sensors but the inverse system does not exist. First, two blind extraction models are presented. Second, a necessary and sufficient condition is derived for the solvability of blind source extraction. A sequential blind extraction approach is then proposed to extract all potentially separable filtered sources. Third, based on a general principle of blind extraction, a cost function and an algorithm are presented. Finally, simulation results are discussed to show the validity of the approach.