Research on Noise Reduction of Motor Ripple Based on Improved Wavelet Threshold Function

Abstract

The use of window motor ripple signals for window anti-pinching can effectively reduce costs compared to the use of Hall sensors. However, noise reduction is required since the ripple signal is small in amplitude and contains interference noise that is not conducive to feature detection. The traditional wavelet threshold de-noising method uses soft and hard threshold functions with constant deviation and additional oscillation problems. In this paper, an improved wavelet threshold function is proposed, and the noise reduction effect of the three threshold functions is examined in MATLAB using simulated signals. By analyzing the two data indexes of signal-to-noise ratio and root mean square error, the improved wavelet threshold function proposed in this paper can effectively improve the noise reduction effect of ripple signals.