Blind Separation of Mixtures of Piecewise AR(1) Processes and Model Mismatch

0447196 - ÚTIA 2016 RIV CZ eng C - Conference Paper (international conference)
Tichavský, Petr - Šembera, Ondřej - Koldovský, Zbyněk
Blind Separation of Mixtures of Piecewise AR(1) Processes and Model Mismatch.
Latent Variable Analysis and Signal Separation. Heidelberg: Springer, 2015 - (Vincent, E.; Yeredor, A.; Koldovský, Z.; Tichavský, P.), s. 304-311. Lecture Notes in Computer Science. ISBN 978-3-319-22482-4. ISSN 0302-9743.
[Latent Variable Analysis and Signal Separation 12th International Conference, LVA/ICA 2015. Liberec (CZ), 25.08.2015-28.08.2015]
R&D Projects: GA ČR(CZ) GA14-13713S
Institutional support: RVO:67985556
Keywords : Autoregressive processes * Cramer-Rao bound * Blind source separation
Subject RIV: BI - Acoustics
http://library.utia.cas.cz/separaty/2015/SI/tichavsky-0447196.pdf

Modeling real-world acoustic signals and namely speech signals as piecewise stationary random processes is a possible approach to blind separation of linear mixtures of such signals. In this paper, the piecewise AR(1) modeling is studied and is compared to the more common piecewise AR(0) modeling, which is known under the names Block Gaussian SEParation (BGSEP) and Block Gaussian Likelihood (BGL). The separation based on the AR(0) modeling uses an approximate joint diagonalization (AJD) of covariance matrices of the mixture with lag 0, computed at epochs (intervals) of stationarity of the separated signals. The separation based on the AR(1) modeling uses the covariances of lag 0 and covariances of lag 1 jointly. For this model, we derive an approximate Cram´er-Rao lower bound on the separation accuracy for estimation based on the full set of the statistics (covariance matrices of lag 0 and lag 1) and covariance matrices with lag 0 only. The bounds show the condition when AR(1) modeling leads to significantly improved separation accuracy.
Permanent Link: http://hdl.handle.net/11104/0249577