低信噪比条件下改进ESPRIT方法

The Modified ESPRIT Method Under Low SNR Condition

在低信噪比时,针对估计信源DOA实时性的问题,提出了一种新的适合于ESPRIT算法的多级维纳滤波器(MSWF)结构,找到了一种能判别信号子空间的方法。首先将多级维纳滤波器(MSWF)与ESPRIT算法相结合,采用多级维纳滤波器(MSWF)的前向递推,得到子空间,不需要通过特征值分解。低信噪比时,针对噪声子空间泄漏到信号子空间的现象,提出一种判别方法,找到了更精确的信号子空间,结合ESPRIT方法实现信号的DOA估计。由于该算法实现了真实的信号子空间的判断,因此,比传统基于MSWF算法具有更高地估计精度。特别是在低信噪比时,增强了算法的实用性,仿真结果证明了算法的有效性。

In low SNR,a new multistage wiener filtering（MSWF）suitable for ESPRIT is proposed for the instant problem among direction-of-arrival（DOA） estimation of signals. Firstly,the MSWF is combined with ESPRIT. Then subspace can be acquired through spatially smoothed forward recursion of the multistage wiener filtering,not through the matrix eigenvalue decomposition. In low SNR,the noise subspace leaked to the signal subspace,a discriminated method is proposed to find a more accurate signal subspace. The DOA of singles can be estimated combining with the subspace kind algorithms such as ESPRIT. Because this algorithm has realized the judgment of the real signal subspace,it has higher estimation precision than the traditional algorithm based on MSWF. Especially in low SNR,enhance the practicability of the algorithm. Simulation results verify that the proposed algorithm is effective.