初 萍,张金凤.色噪声下的延时二阶预处理MUSIC算法[J].深圳大学学报理工版,2016,33(06):571-577.[doi:10.3724/SP.J.1249.2016.06571 ]
Chu Ping and Zhang Jinfeng.Delay second order preprocessing MUSIC algorithm in presence of colored noise[J].Journal of Shenzhen University Science and Engineering,2016,33(06):571-577.[doi:10.3724/SP.J.1249.2016.06571 ]
色噪声下的延时二阶预处理MUSIC算法

深圳大学信息工程学院,广东深圳 518060

雷达工程; 阵列信号处理; 波达方向估计; 色噪声; 滑动平均模型; 延时相关

Delay second order preprocessing MUSIC algorithm in presence of colored noise
Chu Ping and Zhang Jinfeng

Chu Ping and Zhang JinfengCollege of Information Engineering, Shenzhen University, Shenzhen 518060, Guangdong Province, P.R.China

radar engineering; array signal processing; direction of arrival estimation; color noise; moving average model; delay correlation

DOI: 10.3724/SP.J.1249.2016.06571

备注

为抑制色噪声对多信号分类(multiple signal classification, MUSIC)算法估计性能的影响,提出利用延时相关消除色噪声的二阶预处理方法.以L阶滑动平均(moving average, MA)模型为色噪声模型,利用大于L阶的延时相关函数为0的特点去除背景色噪声.为增加对多个入射信号成功分辨的概率,基于信号子空间估计稳健的特点,构造基于信号子空间加权的谱函数.提出的新算法无阵列孔径损失,不增加运算复杂度,因此,在色噪声背景下的实时波达方向(direction of arrival, DOA)估计中有较好应用价值.仿真实验表明,所提方法可有效抑制色噪声,在同样的估计条件下,该算法表现出较好的估计性能,且不增加计算量.

In order to avoid the effect of colored noise in multiple signal classification(MUSIC)algorithm, we propose a second order preprocessing method with delay correlation function. We eliminate colored noise by choosing a moving average(MA)model with L orders and constructing a delay correlated function that equals zero when its order is more than L. We construct a new weighted spectrum function based on the robust estimation performance of signal subspace in order to improve the resolving probability. The proposed algorithm improves estimation performance without loss of array aperture and complex computation. Hence, it has better application value in real direction of arrival(DOA)estimation with colored noise. Simulation results show that the proposed algorithm can eliminate colored noise effectively, and it has a better estimation performance without additional computation under the same estimation conditions.

·