基于三阶累积量的稳健盲波束形成方法

Robust blind beamforming algorithm based on third-order cumulants

盲波束形成的目的就是在不知道阵列流形先验信息的情况下,从阵列输出数据中恢复出源信号。对于独立源信号的盲分离,Cardoso和Souloumiac提出了一种基于四阶累积量的盲波束形成方法(JointApproximate Diagonalisation of Eigen-matrices,JADE)。然而,巨大的计算量限制了该方法在工程中的应用。而在实际应用中很多信号并不服从对称分布,因而具有三阶累积量。为此,通过联合对角化由接收数据三阶累积量构造的一系列矩阵来估计混合矩阵,进而实现盲波束形成。与JADE相比,该方法具有较小的计算量,并且对于有限采样数据造成的估计误差更加稳健。

The goal of blind beamforming is to recover source signals only from the output of array without any a priori information about the array manifold.For blind separation of the independent sources,Cardoso and Souloumiac proposed an effectively blind beamforming method（JADE） using fourth-order cumulant.However,the high computation complexity of this method limits its applicability in practical problems.In many applications,the source signals have nonzero third-order cumulant duo to the fact that they obey asymmetric distribution.In this paper,a new method is proposed to separate independent sources with nonzero third-order cumulant,which estimates mixing matrix by jointly diagonalizing several matrices formed by the third-order cumulants of array data.Contrary to JADE method,the proposed method has much lower computation complexity and is more robust to the estimation errors resulting from finite sampling data.