一种自适应波形匹配的EMD端点延拓算法

An EMD endpoint continuation algorithm of adaptive waveform matching

为了解决EMD分解微多普勒雷达信号的端点效应问题,提出了一种自适应波形匹配的EMD端点延拓算法。将信号划分为含一对极大极小值的多个子波,并对其进行相同点数的等间隔采样,通过设计匹配系数,选取与端点子波匹配度最高的子波作为延拓子波。匹配系数由匹配子波与端点子波采样点的幅度匹配度与频率匹配度加权求和而成,具有幅度和频率自适应性;将延拓后的信号做EMD分解,提取信号中的微多普勒频率;采用文中算法对微多普勒仿真信号与手势微多普勒雷达信号进行处理,实验结果表明,文中算法可以将上述信号分解为不同的频率分量,并且效果优于不延拓、镜像延拓以及极值平均延拓,验证了文中算法的有效性。

To solve the endpoint effect of EMD for micro-Doppler radar signals,an adaptive waveform matching EMD endpoint continuation algorithm is proposed.The signal is divided into multiple wavelets that contains a pair of maxima and minima.The wavelets are sampled at equal intervals with the same number of points.With the designed matching coefficient,the wavelet with the highest matching degree with the end point wavelet is selected as the continuation wavelet.The matching coefficient is weighted sum of the amplitude and frequency matching degree between the samplings of the matched wavelet and the endpoint wavelet,and is self-adapted on amplitude and frequency.The continued signal is decomposed by EMD method to extract the micro-Doppler frequency of the signal.The proposed algorithm is used to process a micro-Doppler simulation signal and a micro-Doppler radar signal of the human gesture.The experimental results show that the above signals are successfully decomposed into different frequency components by the proposed algorithm.And the results are better than non-continuation,the image continuation and the extremum average continuation,demonstrating the effectiveness of the proposed algorithm.