Estimating tire condition based on wavelet analysis and frequency modulation

Tire is one of the most important components of automobiles ,since tire is the only interface with roads. However, little is known about the interface. For example, although we can detect the slipping of wheels with the wheel angle sensors, we never know the grip margin for slipping. At this point, we have to design automobiles´ electro-controllers with much safety margin. This is undesirable in the meaning of automobile controllability as well as safety. Thus, it is important to get information of tire´s surface. In order to get information of tire surface, we made use of tire sound. We recorded the sound from rotating tires with a digital recorder, and got a general time-frequency trend by analyzing them with CWT (continuous wavelet transform). Then, to be more specific, we analyzed the particular frequency band of the CWT with its frequency modulation. We calculated the fluctuation of the phase of the CWT from the sound, and the tread pattern in its 250Hz band. And by evaluating the correlation between the two fluctuation, we can see the tread pattern which the sound is from.