Discrimination of normal and studless tyres by wavelet sound analysis

This paper proposes a method for extracting tyre characteristics from the tyre sounds. The wavelet analysis is able to visualize the sounds in the time-frequency domain. From the vision, we can obtain much more vivid information on the tyre than just listening to them with our ears. We performed the experiments of recording some sounds of normal and studless tyres with a tyre-testing bench. Then we calculated the wavelet transforms of their sounds, and succeeded in discriminating them by their characteristic frequency bands in the wavelet domain. On the basis of the results, the more specific differences between normal and studless tyres are shown in their power spectrum densities.