一种脉冲噪声中的韧性子空间跟踪方法

A robust subspace tracking method in impulsive noise environment

一般地,基于二阶统计量的子空间跟踪方法对脉冲噪声敏感,性能出现退化.为此以α稳定分布作为脉冲噪声的模型,研究噪声环境中的韧性子空间跟踪方法.以分数低阶统计量理论为依据,把α稳定分布噪声中的子空间跟踪看做一个无约束的优化问题,提出了一个新的代价函数,并推导出一个韧性算法.同时还利用M估计对算法进行了简化.在数值模拟中把新算法应用于方向估计,结果表明了新算法和简化算法的有效性.

It is usually found that the subspace tracking algorithm based on the second-order statistics is sensitive to impulsive noises, which results in the degradation of performance. It is necessary to develop a robust subspace tracking method. The a-stable distribution as the impulsive noise model is introduced. Subspace tracking in impulsive noises as an unconstrained minimization problem is considered based on fractional lower-order statistics theory and a new recursive solution is proposed. Because of the computation complexity of the solution, the algorithm is simplified by M-estimation. Numerical simulation demonstrates the good performance of the new subspace tracking algorithm in direction-of-arrival （DOA） estimation.