小波分析在AR模型重构路面不平度中的应用

摘要

针对线性滤波器AR模型重构路面不平度时低频处误差较大的问题,提出利用小波分析的时频特性来调整重构的路面随机序列.首先根据等级路面的功率谱利用AR模型构造C级路面随机序列,然后利用小波分析的小波分解方法和小波包分解方法对路面随机序列进行分解,根据信号能量与信号振幅平方成正比原理,调整信号分解分量各频段的振幅后再进行信号重构,并在时域和频域上与原信号进行对比.仿真结果表明,小波分解方法和小波包分解方法都能提高重构路面功率谱与目标功率谱的拟合精度,且小波包分解方法在处理高频分解分量时更具优势.%Aiming at the problem that the error in the low frequency band was large when reconstructing the road roughness in AR model,the wavelet analysis method was proposed to improve the reconstructed random road for its time-frequency characteristic.Firstly,the power spectrum of the grade road was used to reconstruct the C-level random road in the AR model.The random road was then decomposed by the wavelet decomposition method and the wavelet packet decomposition method.The amplitude of components at each frequency band was adjusted to approach the given power spectrum under the principle that the signal energy was proportional to the square of the signal amplitude.Finally,a random road was reconstructed by those adjusted components and compared with the standard road in time and frequency domain.The results show that both methods can improve the relative error between the reconstructed road roughness PSD and standard one.Furthermore,wavelet packet decomposition method has more advantages in fitting the high frequency components.