Robust blind separation algorithms for heavy-tailed sources
M. Sahmoudi, Karim Abed‐Meraim · 2005
The impulsive or heavy-tailed characteristics of sources signals can degrade severely the performances of existing blind source separation (BSS) methods. In this paper, we focus on the use of normalized statistics (NS) of heavy-tailed sources for the BSS problem. In M. Sahmoudi et al., (2003), the NS have been introduced for alpha-stable sources to justify for the use of algebraic-based separation algorithms (JADE, SOBI, etc) to achieve the BSS in the heavy-tailed case. In this work, we propose to use the NS to robustify the class of equivariant adaptive source separation algorithms EASI. Algorithm derivation, discussion and simulation results are provided to illustrate the usefulness of NS in that context. The new method has been compared with two of the most popular BSS algorithms; EASI and quasi maximum-likelihood algorithm.