首页>
外国专利>
Method for determining the step size Alpha for adjusting the convergence speed in the NLMS algorithm
Method for determining the step size Alpha for adjusting the convergence speed in the NLMS algorithm
展开▼
机译:NLMS算法中用于确定收敛速度的步长α确定方法
展开▼
页面导航
摘要
著录项
相似文献
摘要
The method involves forming a cross-correlation function of two microphone signals in each sampling period, determining an absolute maximum of the function, and evaluating, both an amplitude and a position of the maximum through fuzzy logic. A level comparison of the microphone signals, after a corresponding delay of each signal according to predetermined time constants, is performed, and a respective mean value is formed and filtered. The results of the cross-correlation function and the level comparison are linearly combined in a fuzzy logic, and the resulting signal is filtered in a non-linear low-pass filter. A step-width element is calculated from the resulting signal, and is normalized in dependence on a microphone sum signal, a residual echo, and the result of the level comparison.
展开▼