The dynamic LMS for line echo cancellation

摘要

In this paper, the dynamic least mean square(DLMS) algorithm is firstly proposed to reduce the computation of LMS. In the DLMS architecture, the redundant coefficients of LMS can be grouped automatically. The coefficient number of filter and the computation are reduced. Moreover, two modes of DLMS algorithm are employed analyzed. The DLMS-based line echo cancellation (LEC) is employed to evaluate the performance of DLMS. The experimental results confirm the good performance in DLMS algorithm. Moreover, the performance of mode-II is superior to the mode-I. Moreover than 44% of coefficient saving of LMS is achieved.