基于加权信号子空间投影的MUSIC改进算法

Improving MUSIC performance in snapshot deficient scenario via weighted signal-subspace projection

针对MUSIC算法在小快拍条件下DOA估计时性能降低的缺点,提出采用加权信号子空间投影的方法加以改善。首先对信号子空间进行主特征值倒数加权,并将结果与MUSIC空间谱进行叠加。新方法保持了噪声子空间处理的高分辨力,并通过信号子空间处理提高了对有限数据误差的稳健性。理论和统计性能分析表明其对多目标分辨力优于MUSIC方法,特别是在小快拍条件下表现了良好的性能。

Aiming at the performance of MUSIC algorithm decreases in DOA estimation with snapshot deficient scenario, a new method based on weighted signal subspace projection is proposed. The signal-subspace with reciprocal of principle eigenvalue is weighted in this method. Then the result is added over MUSIC spatial spectrum. The high-resolution of noise subspace processing is remained, and robustness via signal subspace processing is improved in this method. Theoretical and statistical analysis show that its performance is better than that of MUSIC, especially in snapshot deficient scenario.