【摘要】 为了提高求取EMD均值曲线的速度和精度,提出了一种基于分段高阶样条拟合的获取方法,它利用分段高阶样条曲线获得更精确的均值点,同时又克服了高阶样条拟合中运算量太大的缺点,对均值点序列直接进行拟合得到均值曲线,在保证均值曲线光滑的前提下减小了欠冲与过冲的影响,而且获得了较快的筛选本征模态函数的速度.使用该方法分别对平稳和非平稳信号进行了处理,结果表明该方法可准确实现模态提取,对2种信号都取得了良好的效果.更多还原

【Abstract】 This paper proposed a method of obtaining the mean curve for empirical mode decomposition(EMD).The method obtained the mean curve based on piecewise high order spline fitting.It obtained more accurate mean points by using a piecewise high order spline curve,but overcame the disadvantage of too much computation in high order spline fitting.The mean curve was attained by directly fitting the mean points sequence,and the effects of overshooting and undershooting were reduced on the premise of guaranteeing the smoothness of the mean curve;in the meantime a faster speed of sifting intrinsic mode functions(IMF) was achieved.Both stationary and non-stationary signals were analyzed by this method,and the result shows that the method realized accurate mode extraction and is useful for both kinds of signals.更多还原