现有宽带子空间类方位估计方法需要对庞大的空时联合采样相关矩阵进行特征分解运算以获得好的估计性能,计算复杂度高。提出一种基于噪声子空间二阶锥规划求解、计算有效的宽带波达方向估计算法。该算法将宽带信号低秩模型的噪声子空间估计转换为凸最优理论中的二阶锥规划问题,实现宽带源方位估计,整个过程不需要估计样本协方差矩阵,不需要矩阵特征值分解或奇异值分解,运算量低。给出算法的理论分析、实现步骤。统计性能比较表明该算法分辨门限低、方位估计精度高。千岛湖湖试结果证明该算法在实际声纳阵系统中可有效实现宽带源波达方向估计。
Aim. At the beginning of the paper our brief review of existing literature reveals, in our opinion, that the propagator idea of Ref. 8 by S. Marcos et al and the convex optimization idea of Ref. 11 by Z.Q. Luo et al can be utilized in such a way that a fast broadband DOA estimation method can be brought forth. In the full paper, we explain our estimation method and its effectiveness in some detail; in this abstract, we just add some pertinent remarks to listing the two topics of explanation. The first topic is. the modified propagator method. In this topic, we solve for a signal direction matrix so as to estimate the array signal noise subspace and the signal spatial spectrum. The second topic is. broadband DOA estimation algorithm. Its four subtopics are. broadband frequency extension model (subtopic 2.1), solutions for second-order cone programming (SOCP) (subtopic 2.2), the broadband DOA estimation algorithm based on the SOCP (subtopic 2.3) and discussion on calculation complexity (subtopic 2.4). In subtopic 2.2, we convert the eq. (18) in the full paper into a form suitable for SOCP to obtain noise subspace. In subtopic 2.3, we adopt the spatial domain focusing matrix to work out the 8-step broadband DOA estimation algorithm. In subtopic 2.4, we conclude that the calculation complexity of the abovementioned estimation algorithm, in our opinion, is much less than other traditional broadband subspace methods. Finally, we do both computer simulation and lake experiments to verify the effectiveness of our algorithm. The simulation results, shown in Figs. 1 through 4 in the full paper, indicate preliminarily that. (1) our algorithm has better resolution; (2) the value ot DOA estimation with our algorlmm nas smaller standard deviation and error than those with other algorithms. We did experiments in the Qiandao Lake, Zhejiang Province, China. The experimental results, given in Fig. 5, demonstrate preliminarily that our algorithm has stable performance in an actual sonar system and ca