On extending the FastISA algorithm to noncircular complex sources

Fast Independent subspace analysis (Fast ISA) algorithm is an efficient fixed-point type algorithm for solving ISA problem with real valued data. However, the applicable range of Fast ISA is limited due to the fact that it cannot work well for complex valued data. In this paper, we extend the contribution of Aapo Hyvarinen et al to a more general case of complex valued data. Furthermore, as the algorithm may convergence to local minima, we also propose a simple way to get rid of this problem. Simulation results demonstrate the effectiveness of the proposed method.