冲击噪声环境下的快速实值算法

A Fast Real-Valued Algorithm in Impulsive Noise Environment

在冲击噪声环境中,包括MUSIC、ESPRIT和MNM在内的传统DOA算法估计性能急剧下降甚至失效,因此出现了一些基于分数低阶矩的稳健性算法,例如ROC-MUSIC和FLOM-MUSIC等。虽然在冲击噪声环境下,这些稳健性算法具有很好的工作性能,但是这些算法需要进行计算量较大的复数特征分解,从而不适宜实现DOA的快速有效估计。为了解决这一问题,提出了一种快速实值算法(FRA),由于该算法可在实数域工作并且不需要进行特征分解,所以它的计算量远小于现有算法,在同等条件下更有利于快速地估计DOA。计算机仿真证明该算法具有优良的DOA估计性能,适宜在冲击噪声环境下应用。

The conventional DOA algorithms, including MUSIC, ESPRIT, MNM etc, gave poor performance or even did wrong in the impulsive noise environment, so some robust ones are proposed based on fractional lower order moment（FLOM） in this paper, such as ROC-MUSIC, FLOM-MUSIC, and so on. The burden for the complex eigendecomposition prevents the algorithms mentioned above from estimating DOA quickly and efficiently. The robust ones show good performance under the impulsive noise environment. In the paper, a new fast real-valued algorithm（FRA） is proposed to smooth away the problem. Due to working in real domain without any eigendecomposition, tbe algorithm has smaller computation load than the former and is suitable for getting the DOA estimate more quickly than the former under the same condition. Computer simulations show that the new algorithm has good performance of DOA estimate and can be applied in the impulsive noise environment.