Title :
Normalised natural gradient algorithm for the separation of cyclostationary sources
Author :
Jafari, M.G. ; Chambers, J.A.
Author_Institution :
Centre for Digital Signal Process. Res., King´´s Coll., London, UK
Abstract :
A normalised natural gradient algorithm (NGA) for the separation of cyclostationary source signals is proposed in this paper. It improves the convergence properties of the cyclostationary natural gradient algorithm (CSNGA) by employing a gradient adaptive learning rate whose value changes in response to some change in the filter parameters. Experimental results demonstrate the improved behaviour of the approach.
Keywords :
adaptive filters; adaptive signal processing; convergence of numerical methods; gradient methods; learning (artificial intelligence); source separation; convergence properties; cyclostationary source signal separation; filter parameters; gradient adaptive learning rate; normalised natural gradient algorithm; Adaptive filters; Autocorrelation; Blind source separation; Convergence; Digital signal processing; Educational institutions; Performance analysis; Signal processing; Signal processing algorithms; Source separation;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03). 2003 IEEE International Conference on
Print_ISBN :
0-7803-7663-3
DOI :
10.1109/ICASSP.2003.1199931
