采用相干积累矢量平滑实现小样本信号解相干

Small sample signals' decorrelation using the coherent accumulation vector' spatial smoothing

提出了一种基于相干积累矢量空间平滑的解相干算法,该算法基于相干积累的原理来构建矢量,其对阵列输出进行相干积累后得到特征矢量,由此特征矢量进行前后向空间平滑,构建矩阵,实现解相干.最后由该矩阵奇异值分解得到信号或噪声子空间,与现有解相干算法相比,运算量大大下降.仿真结果表明,该算法具有很好的抗噪性,其在信噪比高于-14 dB时保持了与基于特征分解的矢量奇异值法相似的性能.

When estimating the DOA of narrowband signals,if the snapshot number is small,decorrelation operation is necessary because for all the snapshots,their amplitude and phase are unchanged,which makes signals from different directions coherent.For this kind of signals,existing decorrelation algorithms always need too much computation or are sensitive to SNR.This paper proposes taking coherent accumulation operation for array outputs,and the resulting vector is used to construct a matrix by forward and backward smoothing.By taking SVD of this matrix,the signal and noise subspace is obtained then.Simulation results indicate that this method is of excellent noise-resistance performance with great computational complexity reduction in contract to existing decorrelation algorithms.It has the same performance as ESVD algorithms based on eigen-decomposition,when the SNR is higher than-14dB.