从空间距离的角度分析子空间跟踪问题,提出了一种新的子空间跟踪算法。提出的算法同时适用于信号和噪声子空间跟踪,而复杂度仅为3NL+0(N)。仿真表明算法对基的正交偏离稳健,对舍入误差不敏感,是一种快速稳健的子空间跟踪算法。
A novel subspace tracking algorithm based on subspace distance is proposed in this paper. The proposed algorithm provides robust orthonormal vector estimation adapting to signal subspace and noise subspace. Simulation results demonstrate that the proposed algorithm is numerically stable, because it does not accumulate roundoff errors. Moreover, the proposed method performs well in subspaee tracking and it has a low complexity of 3NL + O(N).