基于小波变换的变步长LMS自适应滤波算法

变步长LMS自适应滤波算法通过构造合适的步长因子有效的解决了传统LMS算法收敛速度和稳态误差相矛盾的问题.变换域LMS自适应滤波算法通过正交变换降低了输入信号矩阵的相关性,提高了算法的收敛速度.将这两种算法相结合,提出了一种新的基于小波变换的变步长LMS自适应滤波算法.仿真结果表明,该算法无论是收敛速度还是稳态误差都有了很大的提高.

Based on the Wavelet Transform of the Variable step LMS Adaptive Filtering Algorithm

The problem that it is difficult for the traditional LMS adaptive filtering algorithm to ease the conflict between convergence and steady-state error is solved by variable step-size LMS adaptive filtering algorithm solved. Transform domain LMS adaptive filtering algorithm can reduce the relativity of input sig- nals effectively through orthogonal transforms, so the convergence rate will be improved. Combining this two algorithms and a new variable step-size LMS algorithm based on wavelet transform can be presented. The computer simulation results indicate that the new algorithm improved both the convergence rate and the steady-state error.