Adaptive algorithm for blind separation from noisy time-varying mixtures

This article addresses the problem of blind source separation from time-var ying noisy mixtures using a state variable model and recursive estimation. An estimate of each source signal is produced real time at the arrival of n ew observed mixture vector. The goal is to perform the separation and atten uate noise simultaneously, as well as to adapt to changes that occur in the mixing system. The observed data are projected along the eigenvectors in s ignal subspace. The subspace is tracked real time. Source signals are model ed using low-order AR (autoregressive) models, and noise is attenuated by t rading off between the model and the information provided by measurements. The type of zero-memory nonlinearity needed in separation is determined on- line. Predictor-corrector filter structures are proposed, and their perform ance is investigated in simulation using biomedical and communications sign als at different noise levels and a time-varying mixing system. In quantita tive comparison to other widely used methods, significant improvement in ou tput signal-to-noise ratio is achieved.