Adaptive Blind Source Seperation Using Temporal Predictability

An adaptive blind source separation method is proposed in the paper. The method uses the temporal predictability as cost function which is not studied as much as other generic differences between the properties of signals and their mixtures. After adaptive pre-whitening the signal, step-adaptation on-line nature gradient algorithm is proposed to separate signals, which is robust and more effective. Simulations show a good separation performance of the algorithm is achieved with low signal to noise ratio