EMD在路面不平度信号趋势项中的应用

为准确计算路面不平度功率谱估计,需要提取信号中的趋势项。提出一种基于相关系数矩阵判断准则的经验模态分解去除趋势项的方法。利用经验模态分解(Empirical Mode Decomposition,简称EMD)将信号分解为一系列固有模态分量及余项,通过分析余项和趋势项的差别,依据相关系数矩阵判断某一固有模态分量是否属于趋势项。仿真信号和实测路面不平度信号处理结果证明,基于相关系数矩阵判断准则的经验模态分解去除趋势项方法具有更高的精度和可靠性。

Application of EMD to Road Roughness Trend

With the purpose of calculating the road power spectra density accurately,it needs to extract road roughness trend.This paper proposes a method to remove trends based on correlation matrix by empirical mode decomposition(EMD).The signal is decomposed into a series of intrinsic mode functions and residue.Through the difference between residue and trends,the intrinsic mode functions can be decided whether it belongs to trend or not based on a correlation coefficient matrix.Through the analog and road roughness simulation analysis,it is concluded that the method has more accuracy and applicability to remove the trend based on the correlation coefficient matrix by EMD.