SPSA for noisy non-stationary blind source separation
T.S. Durrani, Gordon Morison · 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03). · 2004
In this paper a novel application of the simultaneous perturbation stochastic approximation algorithm (SPSA) to the noisy non-stationary blind source separation problem is presented and described. The proposed approach demonstrates the algorithm with a second order cost function suitable for applications to non-stationary data. Some extensions to the algorithm that are currently being investigated are also described in the paper, and the algorithm performance is demonstrated via simulation.