New criteria for blind source separation using second-order cyclic statistics (Q1971603)

The paper deals with the problem of blind separation of cyclo-stationary sources. The authors propose two algorithms for blind source separation which are based on second-order cyclic statistics. Both methods do not need any restrictions on the distributions of the signals. They are applicable to any stationary correlated or uncorrelated Gaussian or non-Gaussian noises. The effectiveness of the proposed approach is illustrated by a simulation example.