一种新的相干信源DOA估计算法: 加权空间平滑协方差矩阵的Toeplitz矩阵拟合

文献1]提出的最优加权空间平滑技术可以使相干源存在时的信源协方差矩阵恢复为对角阵.由于文献1]中导出的最优权矩阵是空间信源方位的函数矩阵,本文利用最优加权空间平滑后阵列协方差矩阵的Toeplitz性,构造了一个全新的优化拟合的代价函数,并基于此提出了一种相干源方位估计的新算法.与文献1]不同,算法的实现不需要方位估计的先验知识和协方差矩阵的去噪预处理.分辨性能的蒙特卡罗仿真实验表明,新算法对空间相干信源的分辨性能优于常规的空间平滑算法和最大似然算法,在小阵列和信源空间间隔较近时,算法的优越性尤为突出.

A Novel Genetic Approach to DOA Estimation of Coherent Sources Based on Weighted Spatial Smoothing and Toeplitz Matrix Fitting

The optimal weighted spatial smoothing(WSS)technique proposed in1]is capable of de correlating the coherent sources perfectly,with the equivalent sources covariance matrix diagonal.Based on the fact that the optimal weight matrix in1]is a function matrix of source directions and the array covariance matrix after optimal WSS is Toeplitz,a new cost function is constructed.The bearing estimates can then be achieved by virtue of a genetic algorithm based Toeplitz matrix fitting.Compared with1],the proposed algorithm can be realized without the a priori knowledge about source directions and de noising preprocessing.The Monte Carlo simulation results demonstrate that the new algorithm is highly superior to the conventional SS(spatial smoothing)technique in the resolving ability for spatially closed coherent sources,especially in the case of small array size.Besides,the comparison of new methods with maximum likelihood methods is also presented.