Second Generation Wavelet-based Denoising Technique for Track Inspection Signal

Abstract

Gap signal was utilized to inspect the track irregularities of high-speed maglev in track inspection system. A spike occurs to the gap signal when the gap sensor mounted on magnet unit passes by the beam joint. When a conventional filter was used for denoising the gap signal, the spike was also removed. However, wavelet-based denoising technique is available due to the multi-resolution feature of the wavelet transform. In this paper, second generation wavelet transform (SGWT) was employed to construct wavelet for denoising the gap signal. The corresponding predictor and update operator were designed, and the predefined soft-thresholding was used to modify the wavelet coefficients. Simulations were used to select the predictor and update operator for the measured gap signal denoising. The results show that when the gap signal is decomposed into three multi-resolution levels, SGWT can reduce the noise from the gap signal effectively, while the spike is also preserved. And this method is suitable to the real-time implementation for the track inspection system.