Alpha稳定分布噪声中的韧性投影近似子空间跟踪算法

A Robust PAST Algorithm in Alpha Stable Noise Environment

为了克服投影近似子空间跟踪算法(PAST)在脉冲噪声环境下性能的退化,本文以Alpha稳定分布为脉冲噪声模型,依据韧性的M估计方法提出了一种新的代价函数,并推导出基于递归最小M估计的韧性投影近似自适应信号子空间跟踪算法(RLM-PAST).由于采用了适合噪声模型的M估计函数,新算法与采用递归最小二乘估计的子空间跟踪算法相比,在稳定分布脉冲噪声环境下具有更好的性能.把新方法应用于波达方向(DOA)估计,数值仿真结果表明了该算法的有效性.

In order to improve the performance of projection approximation subspace tracking （PAST） algorithm under impulsive noise environment, based on Alpha stable distribution as the impulsive noise model, a new cost function is proposed using the robust m-estimation method and then the robust PAST algorithm （RIM_ PAST） is deducted based on the recursive least m-estimate. Because the robust m-estimation function suitable for the Alpha stable noise model is used, the proposed algorithm offers good performance against impulsive noise over the PAST algorithm in direction of arrival （DOA） estimation. The simulation results show the efficiency of the proposed algorithm.