A Noise-Robust EASI Algorithm for Noisy Blind Interference-Signal Separation

This paper presents an improved Equivariant Adaptive Separation via Independence (EASI) algorithm to deal with the blind interference-signal separation in the noisy circumstance. This algorithm gets an unbiased estimate for separation matrix by adopting the noise bias removal technique, then, eliminates the noise in the estimate signals according to the probability density function (PDF) of the noise farther. The simulation result shows that the separation performance of the proposed algorithm is better and the error bit ratio (BER) is lower than those of the conventional EASI algorithm under the same signal noise ratio (SNR) case.