Research on preprocessing method of tractor wheel speed signal based on wavelet transform

Abstract

As one of the most important parameters in the measurement of tractor slip ratio, wheel speed signal must ensure its accuracy in order to accurately measure the tractor slip ratio. Noises make tractor wheel speed signal a significant fluctuation, which may cause control system failure. Fault points elimination method suitable for tractor wheel speed signal was determined based on the characteristics of tractor wheel speed signal during working process. Meanwhile, soft-hard compromise threshold function wavelet denoising method is designed according to the characteristics of tractor wheel speed signal and trials. We use Carsim simulation software to get actual tractor wheel speed signal, and add white noise which SNR (signal to noise ratio) is 60 to the original signal. From the results of several wavelet denoising methods we can conclude that the soft-hard compromise threshold function wavelet denoising method is better than any other ordinary wavelet denoising methods. The SNR of denoised signal is 56.440 and the MSE (mean square error) is 0.0042. The wavelet transform denoising method is feasible to remove noise from tractor driving wheel speed signal.