摘要
针对经典波达方向(direction of arrival,Do A)估计算法复杂度高的问题,讨论了2种快速估计Do A的算法,即:传播算子求根多重信号分类(multiple signal classification,MUSIC)算法与多级维纳滤波器求根MUSIC算法.传播算子求根MUSIC算法是对协方差矩阵分块,得到传播算子构建噪声子空间,结合求根MUSIC算法估计出Do A.多级维纳滤波器不需要估计协方差矩阵,通过滤波器的前向递推,求解维纳-霍夫方程,得到信号子空间,根据正交投影原理,计算出噪声子空间与其共轭转置的乘积,结合求根MUSIC算法估计出Do A.这2种算法都不需对协方差矩阵奇异值分解和谱峰搜索,通过数学分析,复杂度明显降低.
Two fast Do A algorithms were discussed to reduce complexity by classical algorithm, rootMUSIC algorithm based on propagator method and root-MUSIC algorithm using multistage Wiener filter. By propagator method,we got propagator through dividing covariance matrix, built the noise subspace, and then estimate Do A by combining root-MUSIC algorithm. By using multistage Wiener filter, we obtained the signal subspace without estimating covariance matrix,and through forward recursion of the Wiener filter,we calculated Wiener-Hoff equation,which obtain the product of the noise subspace and its conjugate transpose,and finally,according to the orthogonal projection principle,we estimated Do A by combining root-MUSIC algorithm. Neither of these two algorithms requires eigen-values decomposition of the covariance matrix or searching the peak of the spectrum,and the complexity is reduced greatly by the mathematical analysis.
出处
《厦门理工学院学报》
2016年第3期57-60,共4页
Journal of Xiamen University of Technology
基金
国家高技术研究发展计划(863计划)重点项目(2009AA011302)
重庆邮电大学研究生教育创新计划重点项目(Y201019)
重庆市教委科研项目(KJ090513)
重庆市科委重点实验室专项项目
