采用前向空间平滑分组的混合信号波达方向估计算法

Direction of Arrival Estimation Method for Mixed Signals by Forward Spatial Smooth Grouping

针对混合信号在估计其波达方向时阵列孔径和信噪比损失严重的问题,提出了一种基于对称均匀线阵的波达方向估计算法。对于独立信号,利用旋转不变性子空间类算法估计出其波达方向;对于多径传播引起的多组相干信号,利用矩阵的差分以及斜投影分组估计出其波达方向。这种利用矩阵差分的算法不仅消除了独立信号对相干信源的影响,而且相比前向空间平滑算法减少了斜投影和空间平滑的次数,降低了阵列孔径和信噪比的损失。该算法还可以与多种平滑算法相结合,降低对阵元数的要求并提高仿真性能。在估计相干信号的波达方向时,利用旋转不变性子空间类算法思想代替多信号分类算法,避免了谱峰搜索,降低了运算时间。仿真结果表明:存在5个相干信源时,在-10dB的低信噪比的情况下,所提算法仅需7个阵元就可获得误差较低的估计结果,而且与多信号分类算法相比节约了大量运行时间。

Aiming at the problem of serious losses of array aperture and signal-to-noise ratio as estimating the direction of arrivals of mixed signals,a direction of arrivals estimation algorithm based on symmetric uniform linear array is proposed for hybrid independent and coherent sources.For independent sources,the estimating signal parameter via rotational invariance technique is exploited directly to estimate their direction of arrivals,while for multi-group coherent signals caused by multi-path propagation,the direction of arrivals is estimated by exploiting the difference of the matrices and the oblique projection.This algorithm using matrix difference eliminates the effect of independent sources on the coherent source,reduces the number of oblique projection and forward spatial smoothing,and lowers the losses of array aperture and signal-to-noise ratio.The proposed smoothing method can also be combined with various smoothing algorithms to reduce the number of array elements and improve the simulation performance.During estimating the direction of arrivals of coherent sources,the estimating signal parameter via rotational invariance technique is used instead of the multiple signals classification algorithm to estimate the direction of arrivals,which avoids the peak search and greatly reduces the computation time.Simulation shows that in the case of 5 coherent sources with low signal-to-noise ratio of-10 dB,this algorithm only needs 7 array elements to obtain the estimation result with very low error,and spends much shorter time compared with the multiple signals classification algorithm.