摘要
提出了一种基于信号时频分布波达方向估计的新方法。该方法根据信号的时频特征来构造阵列的时频相关矩阵W_(xx)和W_(xy),再利用旋转不变技术对信号的波达方向(DOA)进行估计。由于充分利用了信号的时频域特征,此方法既适用于时不变的、平稳信号又适用于时变的、非平稳信号,与传统的一维ESPRIT方法比较,改善了DOA估计的性能,尤其是在低信噪比和空间近角度时,性能的改善是明显的。数值仿真的结果证明了上述结论。
In this paper, a novel method of the DOA estimation is proposed, which based on time-frequency distribution. This method firstly constructs time-frequency correlation matrix Wxx and Wxy, then estimates DOA of the signals via rotational invariance technique. Because of making the best of time-frequency domain characteristics of the signals, this method is suitable for not only time-invariance and stationary signals but also time-variance and nonstationary signals. Comparing with the 1-Dimension traditional method, this method improves performance of DOA estimation. The performance is obviously improved, especially in lower signal-to-noise ratio and close angle case. The numerical simulation results show the conclusion above is correct.
出处
《燕山大学学报》
CAS
2003年第4期305-309,共5页
Journal of Yanshan University
关键词
时频分布
波达方向
非平稳信号
TF-ESPRIT
信号估计
DOA (Direction-of-Arrival), nonstationary signals, time-frequency distribution, time-frequency ESPRIT.