Abstract: In view of that vehicle road load spectra collected in engineering practices are easily disturbed by outer environment leading to noise and abnormal spike signal, while the traditional Fourier transform-based filtering method may remove true signals together with noise signals simultaneously, a denoising method for vehicle road load spectrum measured based on wavelet transform theory is studied. The Daubechies wavelet function is selected, combined with various wavelet threshold denoising criteria to conduct denoising processing on the road load spectra measured from a test vehicle and systematically study the denoising performance of different wavelet threshold denoising criteria and wavelet vanishing moments for noise-containing random load signals measured. Results show that the wavelet transform-based denoising method can effectively remove disturbing noise signals, in which Heuristic SURE criterion has the best result in retaining all load characteristics of original load spectra. In addition, an automatic inspection method for abnormal spike signals in load spectrum measured is studied based on wavelet transform theory, providing certain reference for optimized processing of vehicle road load spectra measured.

Key words: road load spectra, wavelet transform, wavelet threshold denoising