基于解相干信号的空间谱估计算法的研究

【作者】 陈明；

【导师】 吕明； 蒋瑞炯；

【作者基本信息】 电子科技大学 ， 软件工程， 2011， 硕士

【摘要】 对信号波达方向(DOA)的估计是空间谱估计研究的主题。本文首先介绍了空间谱估计的研究背景、发展现状以及与本文相关的一些基础知识。同时介绍了波达方向估计的传统算法(延迟-相加法和Capon最小方差法)。然后,本文基于均匀线性阵列,考察了MUSIC算法、ESPRIT算法与GEESE算法在不同影响因素下的性能。论文的重点放在了对于相干信号源情况下三种算法的改进上。为了能够分辨相干信号,本文将前后向空间平滑技术与三种算法相结合,并从减小计算量的角度出发,加入了实值分解技术,通过构造实值变换矩阵,以达到减小计算量的目的。在改变影响DOA精度的不同因素的情况下考察改进后的算法对相干信号的估计性能。仿真结果表明,对于加入了实值分解技术与空间平滑技术的三种算法能够有效的分辨出相干信号的波达方向,实值分解技术的加入不仅没有牺牲算法的性能,还极大的减小了改进后算法的计算量。并且改进后的实值平滑算法对信号参数的估计精度随信噪比的增大,阵元数的增加而提高,均具有较高的估计精度,入射角较小时,DOA估计性能要好一些。通过对三种改进后的算法的性能比较可以看出,实值平滑MUSIC算法在性能上优于其他两种算法。更多还原

【Abstract】 The direction of arrival (DOA) estimation of multiple narrowband signals is a classic problem in array signal processing. Some research background of spatial spectrum is stated in detail in this paper, as well as its development and basic results. Two tradition algorithms are also presented in this paper, and then each performance of high-resolution eigencomposition methods (MUSIC, ESPRIT and GEESE algorithm) are investigated in different scenario. This paper focuses on the improvements of the above algorithms for the DOA estimation of coherent signals. In order to estimate the coherent signals, the forward/back spatial smoothing technique is introduced to those algorithms. Further, real-valued decomposition technique which transforms the complex matrix to a real-valued one, is employed to reduce the computational load. At the same time, this paper has carried on the detailed analysis to estimation performance of these algorithms. The simulation results illustrate that the estimation performance of these algorithms, combined with both spatial smoothing and real-valued decomposition technique, will be enhanced. It proves that the real-valued decomposition technique is a powerful method. The proposed algorithms outperform primeval algorithms, specifically in lower arrival direction, bigger SNR and more number of arrays. All the results demonstrate the effectiveness of these improved algorithms. Further, the comparison of these improved algorithms indicates that the performance of improved MUSIC algorithm is better than the others.更多还原