Variable Step Equivariant Adaptive Blind Source Separation Based ICA Algorithm

Blind signal separation (BSS) is a new powerful technique in modern signal processing. Nowadays, it has been wildly applied in diverse fields such as speech signal processing, image processing, array processing and biomedical signal processing. As main technology of solving BSS, independent component analysis (ICA) is used to extract the mutual statistical independent sources.In this paper, independent component analysis basic principle and several commonly used algorithms were introduced at first. On this basis, Equivariant adaptive blind source separation based ICA (EASI) algorithm was focused on. The main work is done on the two different angles are Equivariant adaptive blind source separation based ICA (EASI) algorithm. After detailed analyzing the local stability which EASI algorithm performance in the optimal solution . Further, the relationship between the steady error and the step factor was studied. As EASI algorithm used in a fixed step, so it is not achieves the best combination of the convergence rate and the separated performance. As to Variable step Equivariant adaptive blind source separation based ICA (VS-EASI) algorithm that has been existed, though the simulation tests found that this algorithm also cannot improve the balance between the steady error and the step factor so that the best unity. After the analysis above, a improved Variable step Equivariant adaptive blind source separation based ICA algorithm was proposed. The algorithm's step is based on separating matrix, it is chosen adaptively according to separating degree. Therefore it can improve convergence speed and reduce the maladjustment error in the steady state simultaneously. Computer simulations confirm the theoretical analysis and show the improved algorithm performance is superior to EASI algorithm and VS-EASI algorithm.