基于压缩技术的子空间迭代法及在谱估计中的应用

提出了一种基于压缩技术和子空间迭代的特征向量迭代估计算法,由于该算法采用迭代形式,同目前的特征向量求解方法相比(如奇异值分解法),该算法计算量小、复杂度低、算法收敛速度快、易于实时实现,可对由信号构成的自相关矩阵的特征向量作出准确的估计。通过仿真实验可见该算法具有很高的估计精度。将该算法应用到MUSIC(multiple signal classification)谱估计中,通过计算机进行仿真对比可以看到,利用提出的算法进行谱估计精度要高于标准的MUSIC谱估计精度,且计算量大大减小,由此验证了算法的有效性和优越性。

Subspace iteration method based on practice of deflation technique and its application in spectral estimation

This paper presented a new iterative algorithm for eigenvector estimation based on subspace iteration method and utilization of deflation technique. Compared to current eigenvector calculation methods such as SVD, the proposed algorithm can effectively reduce calculation burden and complexity with high estimation accuracy, and it was easily implemented on time. The algorithm could attaint the higher accuracy eigenvector estimation solution of autocorrelation sequences. Its estimation effectiveness is illustrated by computer simulation. Spectrum estimation used the proposed algorithm is slightly higher accuracy and less calculation, than the MUSIC spectrum estimation algorithm applied by computer simulation, which proves the effectiveness and advantage of the algorithm.