基于子空间旋转变换的低复杂度波达角估计算法

多重信号分选(MUltiple SIgnal Classification, MUSIC)算法是波达方向(Direction-Of-Arrival, DOA)估计的最重要算法之一，但庞大的计算量使其工程实用性大打折扣。为降低MUSIC的计算量，该文基于子空间旋转(Subspace Rotation Technique, SRT)变换思想提出了一种高效改进算法，即SRT-MUSIC算法。SRT-MUSIC利用秩亏特性对噪声子空间矩阵按行分块并以旋转变换得到降维噪声子空间，进而基于该降维噪声子空间与导向矢量的正交性构造空间谱估计信号DOA。理论分析表明:SRT-MUSIC能有效避免空间谱搜索中的冗余运算，从而成倍降低算法的计算量。对于大阵元、少信号情况，所提算法计算效率优势更为明显。仿真实验证明了SRT-MUSIC的有效性和高效性。

Low-complexity DOA Estimation via Subspace Rotation Technique

The MUltiple SIgnal Classification (MUSIC) algorithm is one of the most important techniques for Direction-Of-Arrival (DOA) estimate. However, this method is found expensive in practical applications, due to the heavy computational cost involved. To reduce the complexity, a novel efficient estimator based on Subspace Rotation Technique (STR) is proposed. The key idea is to divide the noise subspace matrix along its row direction into two sub-matrices, and perform STR to get a new rotated sub-noise subspace with reduced dimensions. As this rotated sub-noise subspace is also orthogonal to the signal subspace, a new cost function is finally derived to estimate DOAs. Theoretical analysis indicates that redundancy computations in spectral search are efficiently avoided by the proposed method as compared to MUSIC, especially in scenarios where large numbers of sensors are applied to locate small numbers of signals. Simulation results verify the effectiveness and efficiency of the new technique.