Coprime array exhibits many advantages over the uniform linear array (ULA) with the same number of physical sensors in resolution performance and interference suppression capability. In this study, the authors take the advantages of coprime array to improve the robustness of adaptive beamformer. In the coprime virtual ULA (CV-ULA), they prove that a constructed Toeplitz matrix can be taken as the sample covariance matrix from the perspective of virtual signal characteristics. The CV-ULA Capon spectrum estimator is modified to obtain the directions and powers of all impinging signals. Since the real directions of all impinging signals are located at different angular sectors, they form independent signal subspace for each impinging signal. They also assign independent steering vector mismatches for different impinging signals to obtain their real steering vectors. The steering vector mismatch of each impinging signal is independently obtained by solving its own convex optimisation problem. They reconstruct the interference-plus-noise covariance matrix (INCM) with precise steering vectors and powers of interference signals. The proposed weight vector is computed by combining the desired signal steering vector and the reconstructed INCM. Extensive simulations show that the proposed algorithm provides robustness against many types of model mismatches.

中文翻译：

---

具有转向矢量估计和协方差矩阵重构的鲁棒自适应共prime阵列波束形成

与具有相同数量的物理传感器的均匀线性阵列（ULA）相比，Coprime阵列在分辨率性能和干扰抑制能力方面显示出许多优势。在这项研究中，作者利用了互质矩阵的优势来提高自适应波束形成器的鲁棒性。在互质虚拟ULA（CV-ULA）中，他们证明从虚拟信号特性的角度来看，可以将构造的Toeplitz矩阵用作样本协方差矩阵。修改了CV-ULA Capon频谱估计器，以获得所有撞击信号的方向和功率。由于所有撞击信号的实际方向位于不同的角度扇区，因此它们为每个撞击信号形成了独立的信号子空间。它们还为不同的撞击信号分配独立的转向矢量失配，以获得其真实的转向矢量。通过解决其自身的凸优化问题，可以独立获得每个撞击信号的转向矢量失配。他们利用精确的引导向量和干扰信号的功率重建干扰加噪声协方差矩阵（INCM）。建议的权重向量是通过组合所需的信号控制向量和重构的INCM来计算的。大量的仿真表明，该算法针对多种类型的模型不匹配提供了鲁棒性。他们利用精确的引导向量和干扰信号的功率重建干扰加噪声协方差矩阵（INCM）。建议的权重向量是通过组合所需的信号控制向量和重构的INCM来计算的。大量的仿真表明，该算法针对多种类型的模型不匹配提供了鲁棒性。他们使用精确的引导向量和干扰信号的功率重建干扰加噪声协方差矩阵（INCM）。建议的权重向量是通过组合所需的信号控制向量和重构的INCM来计算的。大量的仿真表明，该算法针对多种类型的模型不匹配提供了鲁棒性。