数据分段输入的低秩自适应滤波方法

Segmented Data-Inputting Based Low-rank Adaptive Filtering

针对高采样率下的高阶自适应滤波器难以实时实现的问题,本文提出了一种数据分段输入的降秩自适应滤波方法,其实现机理与传统的LMS和RLS方法勾不相同,其收敛速度与RLS方法基本相当,运算量却比RLS方法要小,其显著优势是大大降低了对物理处理器的速度要求。

This paper proposes a low-rank adaptive filter of a different scheme from LMS and RLS algorithm mainly for high-order adaptive filtering under high sampling rate. The proposed filter converges almost as fast as RLS but the computational burden is less than the RLS. In most applications the demand for high capacity of the physical processor can be reached by adjusting the intervals between the segmented inputs.