联合特征值和特征子空间投影的信源数估计

Source Number Estimation Based on Eigenvalue and Eigen Subspace Projection

高分辨空间谱估计算法中信源数的准确估计是必要前提.文中结合矩阵重构和特征子空间投影方法,提出一种适用于弹载阵列系统的信源数估计算法.将阵列阵元分成相同的2组,求得这2组阵元接收数据的互协方差矩阵并重构信源数估计矩阵,对重构的矩阵特征分解,联合特征子空间投影和特征值加权的方法构造判决函数来估计信源数.理论分析与仿真结果表明:重构矩阵的信号子空间特征值呈平方倍增大,噪声功率得到抑制;算法有效提高了少量快拍数据和低信噪比条件下信源数估计的正确率.

Detection of the source number is an important step in the high resolution spatial spectrum estimation algorithm. A fast algorithm based on matrix reconstruction and eigen-subspaces projection was pro posed to solve source number estimation of the missile-borne array system. First, the array elements were divided into two groups. Then, a source number estimation matrix was reconstructed by covariance matri- ces of the data received from the array elements. Finally, the proposed algorithm utilized the eigen-subspace projection method to estimate the source number. It was proved that the signal energy was increased to the square times while the noise power was suppressed. Computer simulation confirmed that the algorithm accurately estimated the number of sources under scenarios of low SNR and deficient snapshots.