They may also be polluted by the sensing and numerical procedures. Thus, a direct processing of the data by the Fourier transform is not physically sound, since the latter is strictly valid to linear and stationary systems. The EMD method proposed by Huang et al. [5] was specially tailored for treating nonstationary and nonlinear data. In this paper, the EMD will be adopted to decompose the data collected (or computed) for the vehicle into IMFs that admit well-behaved Hilbert transforms [5]. The essence of the EMD is to identify the intrinsic oscillatory modes by their characteristic time scales in the data empirically, and then to decompose the data accordingly. Generally, the finest vibration mode or component of the shortest period at each instant will be identified and decomposed into the first IMF. And the components of longer periods will be identified and decomposed into the following IMFs in sequence. Therefore, the first bridge frequency, with a longer period, may not appear in the first few IMFs.