Title :
Robust versions of the PAST algorithm
Author :
Rupp, Markus ; Reyes, Carolina
Author_Institution :
Inst. of Telecommun., Vienna Univ. of Technol., Vienna, Austria
Abstract :
In this contribution we analyze the Projection Approximation Subspace Tracking (PAST) algorithm by a novel analysis method based on Singular Value Decomposition (SVD). Based on the gained new insights we propose several robust algorithmic modifications that allow to guarantee stability of the algorithm run for normalized step-sizes. Furthermore we investigate the tracking behavior for such step-size choices.
Keywords :
approximation theory; singular value decomposition; PAST; PAST algorithm; SVD; novel analysis method; projection approximation subspace tracking; robust versions; singular value decomposition; Algorithm design and analysis; Convergence; Eigenvalues and eigenfunctions; Equations; Robustness; Signal processing; Signal processing algorithms; PAST algorithm; blind source separation; subspace tracking;
Conference_Titel :
Signal Processing Conference (EUSIPCO), 2012 Proceedings of the 20th European
Conference_Location :
Bucharest
Print_ISBN :
978-1-4673-1068-0
