Unified Parametric ICA Algorithm for Hybrid Sources and its Stability Analysis

摘要

Independent component analysis (ICA) refers to extract independent signals from their linear mixtures without assuming prior knowledge of their mixing coefficients. The purpose of this paper is to develop a novel unified parametric ICA algorithm, which enable to separate hybrid source signals including symmetric and asymmetric sources with a self-adaptive score functions. It is derived from the parameterized asymmetric generalized Gaussian density (AGGD) model. The parameters of the score function in the algorithm can be chosen adaptively by estimating the high order statistics of the observed signals online. Stability analysis of the proposed AGGD-ICA learning algorithm is also discussed. Compared with conventional ICA algorithm, the method can separate a wide range of source signals using only one unified density model. Simulations confirm the effectiveness and performance of the proposed algorithm.