基于实值信号子空间的虚拟阵列解相干算法

针对存在相干信源时,传统的DOA估计算法失效问题,提出一种基于实值特征子空间的虚拟阵列解相干算法.该算法根据虚拟阵列变换的思想,利用阵列接收数据构造虚拟子阵,实现对信号的解相干处理,并将协方差矩阵从复数域变换为实数域,获得一个实值信号子空间,最后利用实数域ESPRIT (Unitary ESPRIT)估计信号波达方向.该方法避免了阵列孔径损失,保持了阵列的空间分辨率,估计精度高,利用个阵元可估计个信源,且引入实数域处理和无需空间谱搜索,运算量小.计算机仿真验证了该方法的有效性和优越性.

A virtual array decorrelation algorithm based on thereal signal subspace

As for the problem of direction of arrival(DOA) estimation in a coherent environment,many existing methods do not work correctly when existing a coherent source.In view of this problem,a virtual array decorrelation algorithm based on the real-valued feature subspace method was proposed.With regard to this approach,a virtual sub-array can be constructed by the received data to de-correlate the coherent sources based on virtual array translation.Then a real signal subspace can be obtained by changing the complex number field into the real number field of the covariance matrix,and at last the direction of arrival(DOA) can be estimated by using the unitary ESPRIT method.With this method,the effective array aperture is not decreased,the spatial resolution of array is remained,and the estimation precision is high.The coherent signals are detected with sensors.The computation load is reduced by using the unitary ESPRIT method without a spectral peak search.This method has better performance in estimation accuracy;the results of simulation verified the effectiveness and superiority of the algorithm.