论文部分内容阅读
本文提出了一种利用矩阵奇异值分解来作空间谱估计的方法,即对由天线阵获取的数据所构成的数据矩阵作奇异值分解、删除来自噪声的贡献的诸最小奇异值以改善信噪比,并利用噪声奇异向量和天线阵的方向向量正交的性质来计算空间谱。除了奇异值分解算法本身给计算稳定性带来好处外,本方法的谱估计性能和计算量均优于近几年来国外广泛关注的一种谱估计算法——MUSIC算法。本方法可用于高分辨的测向系统中。
In this paper, we propose a method for estimating the spatial spectrum using matrix singular value decomposition (SVD) by singularly factorizing the data matrix formed by the data acquired by the antenna array and deleting the smallest singular values from the contributions of noise to improve the signal-to-noise Ratio, and uses the properties of the noise singular vector and the orthogonal direction vector of the antenna array to calculate the spatial spectrum. Except that the singular value decomposition algorithm itself brings the advantage of computational stability, the spectral estimation performance and the computational cost of this method are better than the MUSIC algorithm, which is widely concerned abroad in recent years. The method can be used in high-resolution direction finding system.