Wavelet Transform Adaptive De-noising Algorithm and Application Based on a Novel Variable Step Function

Abstract

Research on a novel variable step function and use it in orthogonal wavelet transform least mean square (LMS) adaptive de-noising algorithm. The algorithmic principle was explained and the effect of orthogonal wavelet transform to arithmetic convergence speed was analyzed. A novel variable step function based on Sigmoid nonlinear functional relationship was proposed, and its characteristics were analyzed. It is applied to time domain LMS algorithm to analyze convergence speed and steady-state error of model identification. Then, the novel variable step function was used in orthogonal wavelet transform domain adaptive body vibration signal de-noising. The simulating results indicate that the novel variable step function gains well effect.