Identification of Chaos with Wavelet Transform Technique

Abstract

This study is concerned with the identification of chaotic vibration by continuous wavelet transform technique. The time histories to be examined are from numerical simulation of dynamic response of a nonlinear seven degrees of freedom full-vehicle model. The results of continuous wavelet transform for the time histories with five mother wavelet, Harr, Mexican Hat, Meyer, Morlet and Daubechies are discussed. It is found that the selection of mother wavelet in the identification is necessary. It is shown that the chaotic motion and almost-periodic motion can be distinguished with presented method when power spectra analysis fails to do so. As a qualitative technique, the continuous wavelet transform could be used in searching for the boundaries of chaotic and non-chaotic responses.