Blind methods for interference cancellation in array processing

Abstract

In this paper, we address several methods which permit to suppress the interference signals received at the antenna level. The classical method to suppress interference uses second order statistics and consists in forming a spatial filter or in calculating the opposition coefficients to apply on different channels. An original solution consists in using higher order statistics for the blind source separation. These algorithms do not use any a priori knowledge on the array manifold. One of the benefits of such blind separation is that source separation is essentially unaffected by errors in the propagation model or in array calibration. Only the statistical independence and the non-Gaussian source signals are important. An other method: the canonical correlation analysis is also used and compared to the others.