基于张量的互质面阵信号处理方法

Tensor-based approach to the co-prime planar array signal processing

针对由2个稀疏均匀矩形阵列(URA)构成的互质面阵(CPPA),提出了一种基于张量代数的阵列信号处理方法,以提高阵列自由度。首先,对CPPA中的2个URA进行拆分,将这2个URA的接收信号表示成2个张量;然后将其互相关结果处理成一个虚拟阵列的接收信号张量。分析表明,所提方法可将一个具有2L2-1个物理阵元的CPPA转换成一个具有(L+1)^4/16个阵元的虚拟稀疏非均匀面阵。针对该虚拟面阵,给出了利用张量分解从其接收信号张量中估计入射信号二维波达角的方法,以避免二维谱峰搜索。与文献报道的互质面阵信号处理方法相比,所提方法将阵列自由度从L^2提升至[(L+1)^4/16]+1,并具有更好的信号波达角估计性能及较低的计算复杂度,仿真结果证明了所提方法的有效性。

For the co-prime planar array(CPPA)consisting of two sparse uniform rectangular array(URA),a new processing method based on tensor algebra was proposed to enhance the degrees of freedom(DoF).By dividing each URA into some overlapping subarrays,the received signals of two URAs were expressed as two tensors.And then the cross-correlation between such two tensors was processed into a received signal tensor of the virtual array.Analysis show that by the new method,the CPPA with 2L2-1 physical elements can be transformed into a virtual sparse non-uniform planar array with(L+1)^4/16 elements.For the virtual array,the tensor decomposition-based approach for estimating the two-dimensional(2-D)direction of arrival(DoA)of the incident signal is also proposed,which means 2-D spectral peak searching is avoided.Compared with the co-prime planar signal processing methods reported in the literature,the proposed method can increase the DoF from L2 to[(L+1)^4/16]+1,and has the better performance of the 2-D DoA estimation and lower computational complexity.Simulation results demonstrate the efficiency of the proposed method.