The covariance matrix reconstruction based robust adaptive beamforming (RAB) methods overcome the performance degradation due to the imprecise knowledge of the steering vector and the covariance matrix. However, high complexity limits the application of them. In this paper, we proposed a new RAB method based on interference plus noise covariance (INC) matrix reconstruction and desired signal steering vector estimation. In this method, nominal interference steering vectors are estimated by the Capon spatial spectrum, as well as noise power. Subsequently, the iterative mismatch approximation algorithm based on maximizing the beamformer output power is proposed to estimate all the incident signal steering vectors and powers, and the INC matrix is reconstructed. Finally, the beamformer is determined by the estimated INC matrix and desired signal steering vector. Simulation results indicate that the proposed method obtains better performance than other existed methods at both the high signal to noise ratio (SNR) and the complexity.

基于协方差矩阵重建的鲁棒自适应波束成形 (RAB) 方法克服了由于导向向量和协方差矩阵的不精确知识而导致的性能下降。然而，高复杂度限制了它们的应用。在本文中，我们提出了一种基于干扰加噪声协方差（INC）矩阵重构和期望信号导向向量估计的新RAB方法。在该方法中，标称干扰导向矢量由 Capon 空间谱以及噪声功率估计。随后，提出了基于最大化波束形成器输出功率的迭代失配逼近算法来估计所有入射信号的导向向量和功率，并重建INC矩阵。最后，波束成形器由估计的 INC 矩阵和所需的信号导向向量确定。仿真结果表明，所提出的方法在高信噪比（SNR）和复杂度方面都比其他现有方法获得了更好的性能。