利用信源先验特征的混合测向算法

发布时间：2019-06-08 19:14

【摘要】：基于特征分解的子空间类测向算法均要知道信源个数,但在小快拍数、低信噪比,且信源间的信号强度差异明显的场合中,传统的AIC信息准则和MDL准则均不能准确判断信源个数。这直接恶化了基于特征分解类算法(如MUSIC法)的测向性能。针对该问题,提出了一种利用信源先验特征的混合测向算法。该算法既利用了信源在角度上呈稀疏分布的信息提高了信源数判决的准确性,也利用了信源的非圆特性改进了测向性能。计算机仿真证实了该方法的正确性。
[Abstract]:The subspace class direction finding algorithm based on feature decomposition should know the number of sources, but in the case of small fast beat number, low signal-to-noise ratio and obvious signal strength difference between the sources, The traditional AIC information criterion and MDL criterion can not accurately judge the number of sources. This directly degrades the direction finding performance based on feature decomposition class algorithms, such as MUSIC method. In order to solve this problem, a hybrid direction finding algorithm based on the prior characteristics of sources is proposed. The algorithm not only makes use of the information that the source is sparse in angle to improve the accuracy of the number of sources, but also makes use of the non-circular characteristics of the source to improve the direction finding performance. The correctness of the method is verified by computer simulation.
【作者单位】：电子科技大学电子工程学院;同方电子科技有限公司;
【基金】：国家自然科学基金(61172140)
【分类号】：TN911.7

【参考文献】