基于谱分解的降阶求根MUSIC算法

发布时间：2018-06-16 10:50

本文选题：波达方向估计 + 求根多重信号分类算法　；参考：《电子与信息学报》2017年10期

【摘要】：求根多重信号分类(Root-MUSIC)算法以多项式求根代替谱峰搜索,降低了波达方向(DOA)估计的计算量,但当阵元数较大时,其计算量依然很大。为进一步降低计算量,该文提出一种降阶Root-MUSIC(RD-Root-MUSIC)算法。该算法基于谱分解将Root-MUSIC多项式的阶次降低一半,再根据矩阵特征多项式与求根多项式的关系构造友阵,采用Arnoldi迭代计算得到友阵的L个大特征值(L为信号数)并估计DOA。仿真结果表明,RD-Root-MUSIC估计精度与Root-MUSIC相近,但其在大阵元下具有比Root-MUSIC更低的计算量。
[Abstract]:Root-MUSIC-based root-MUSIC-based algorithm uses polynomial roots instead of peak search to reduce the computational complexity of DOA estimation, but when the number of array elements is large, the computation is still very large. In order to further reduce the computational complexity, this paper presents a reduced order Root-MUSIC-RD-Root-MUSIC-based algorithm. Based on spectral decomposition, the order of Root-MUSIC polynomial is reduced by half, and then the friend matrix is constructed according to the relationship between eigenpolynomial and root-seeking polynomial. The L large eigenvalue L of the matrix is calculated by Arnoldi iteration and the DOA is estimated. The simulation results show that the estimation accuracy of RD-Root-MUSIC is similar to that of Root-MUSIC, but it has lower computational complexity than Root-MUSIC in large array elements.
【作者单位】：哈尔滨工业大学(威海);西安电子科技大学;中国人民解放军63891部队;
【基金】：国家自然科学基金(61501142) 中国博士后科学基金(2015M571414) 威海市科技攻关和哈尔滨工业大学(威海)学科建设引导基金(WH20160107) 中央高校基本科研业务费专项资金(HIT.NSRIF.201725)~~
【分类号】：TN911.7

【相似文献】