Blind separation methods based on Pearson system and its extensions

we introduce a mutual lnformation-based method for blind separation ot statistically independent source signals. The Pearson system is used as a parametric model. Starting from the definition of mutual information we show using the results by Pham (IEEE Trans. Signal Process. 44(11) (1996) 2768-2779) that the minimization of mutual information contrast leads to iterative use of score functions as estimation functions. The Pearson system allows adaptive modeling of the score functions. The characteristics of the Pearson system are studied and estimators for the parameters are derived using the method of moments. The flexibility of the Pearson system makes it possible to model wide range of source distributions including asymmetric distributions. Skewed source distributions are common in many key application areas, such as telecommunications and biomedical signal processing. We also introduce an extension of the Pearson system that can model multimodal distributions. The applicability of the Pearson system-based method is demonstrated in simulation examples, including blind equalization of GMSK signals.

[1]  J. Cardoso,et al.  Blind beamforming for non-gaussian signals , 1993 .

[2]  P ? ? ? ? ? ? ? % ? ? ? ? , 1991 .

[3]  L. Cobb,et al.  Estimation and Moment Recursion Relations for Multimodal Distributions of the Exponential Family , 1983 .

[4]  Edward J. Dudewicz,et al.  The extended generalized lambda distribution (EGLD) system for fitting distributions to data with moments, II: tables , 1996 .

[5]  Visa Koivunen,et al.  Blind Equalization for GSM Systems via Source Separation , 2001 .

[6]  J. K. Ord,et al.  Families of frequency distributions , 1973 .

[7]  Aapo Hyvärinen,et al.  Fast and robust fixed-point algorithms for independent component analysis , 1999, IEEE Trans. Neural Networks.

[8]  Shun-ichi Amari,et al.  Natural Gradient Works Efficiently in Learning , 1998, Neural Computation.

[9]  Hagai Attias,et al.  Independent Factor Analysis , 1999, Neural Computation.

[10]  M. Kendall,et al.  Kendall's Advanced Theory of Statistics: Volume 1 Distribution Theory , 1987 .

[11]  Visa Koivunen,et al.  Source Distribution Adaptive Maximum Likelihood Estimation of ICA Model , 2000 .

[12]  M. Kendall,et al.  Kendall's advanced theory of statistics , 1995 .

[13]  Aapo Hyvärinen,et al.  The Fixed-Point Algorithm and Maximum Likelihood Estimation for Independent Component Analysis , 1999, Neural Processing Letters.

[14]  Visa Koivunen,et al.  Pearson System based Method for Blind Separation , 2000 .

[15]  E. Dudewicz,et al.  The extended generalized lambda distribution system for fitting distributions to data: history, completion of theory, tables, applications, the “final word” on moment fits , 1996 .

[16]  Jean-François Cardoso,et al.  Equivariant adaptive source separation , 1996, IEEE Trans. Signal Process..

[17]  S. Amari,et al.  Flexible Independent Component Analysis , 1998, Neural Networks for Signal Processing VIII. Proceedings of the 1998 IEEE Signal Processing Society Workshop (Cat. No.98TH8378).

[18]  Shun-ichi Amari,et al.  Blind source separation-semiparametric statistical approach , 1997, IEEE Trans. Signal Process..

[19]  Peter H. Müller,et al.  Pearson's system of frequency curves whose left boundary and first three momentsare known , 1976 .

[20]  Jean-Francois Cardoso,et al.  Blind signal separation: statistical principles , 1998, Proc. IEEE.

[21]  Dinh-Tuan Pham,et al.  Blind separation of instantaneous mixture of sources via an independent component analysis , 1996, IEEE Trans. Signal Process..

[22]  Noboru Murata,et al.  A stable and robust ICA algorithm based on t-distribution and generalized Gaussian distribution models , 1999, Neural Networks for Signal Processing IX: Proceedings of the 1999 IEEE Signal Processing Society Workshop (Cat. No.98TH8468).

[23]  Philippe Garat,et al.  Blind separation of mixture of independent sources through a quasi-maximum likelihood approach , 1997, IEEE Trans. Signal Process..

[24]  J. Cardoso.  Infomax and maximum likelihood for blind source separation , 1997, IEEE Signal Processing Letters.