一种计算复杂度低收敛速度快的递推Capon谱估计算法

基于递推最小二乘方法实现对观测信号自相关矩阵的逆R-1L的估计,研究了一种计算复杂度低和收敛速度快的Capon谱估计算法。该算法避免了用Cholesky分解、奇异值分解、QR分解或目前文献中见到的其他方法所带来的Capon谱估计算法计算复杂的缺点。同时,在信号自相关矩阵条件数不好时,该算法仍然可以进行信号的谱估计。仿真结果表明,尽管该方法的谱估计分辨率不如修正协方差谱估计法,但是两者相差不多,该方法非常有利于谱估计算法的实时实现和实际应用,并且还可以实现Capon幅值估计。

Low computation complexity and fast convergence Caponspectral estimation algorithm

A low computation complexity and fast convergence Capon spectral estimation algorithm was presented by the recursive least mean square (RLS) method, the estimation of the inverse of covariance matrix of the observed signal R~(-1)_L was obtained instead of taking Cholesky decomposition, Singular-value decomposition, QR decomposition and other related matrix inverse methods that can bring the problems of high computation complexity. With bad matrix condition number, the proposed algorithm can also be used for the spectral estimation. The simulation results show that the resolution of the proposed algorithm is only a little inferior to modified covariance algorithm, In addition, this algorithm can be used for the estimation of the amplitude of spectrum Capon.