【作者】 罗晨； 陈韬； 崔明哲； 谢磊； 苏宏业；

【Author】 Chen Luo;Tao Chen;Mingzhe Cui;Lei Xie;Hongye Su;Zhejiang University,State Key Laboratory of Industrial Control Technology;

【机构】 浙江大学工业控制技术国家重点实验室；

【摘要】 稀疏随机模态分解(SRMD)是近年来新兴的信号分解方法,其通过具有随机性的基来分析信号,具有较少的模态混叠现象和Gibbs现象.然而,其模态分解精度以及噪声鲁棒性受到算法原理的限制.因此,本文提出了SRMD的改进方法,基于脊线检测技术获得信号瞬时频率的先验估计,并围绕估计频率生成随机特征,以获得更锐化的时频图.此外,利用循环STFT的时频分割技术对算法噪声参数的估计,以实现算法的参数自适应.仿真和真实信号的数值实验表明,所提出的分解方法相较于SRMD显著提升了分解精度和算法鲁棒性,具有优异的模态重构能力.更多还原

【Abstract】 Sparse Random Mode Decomposition(SRMD) is an emerging signal decomposition method in recent years,which analyzes the signal using bases with randomness,and has fewer modal aliasing and Gibbs phenomena.However,its decomposition accuracy and noise robustness are limited by the algorithm principle.Therefore,this paper proposes an improved method for SRMD,based on ridge detection technology to obtain prior estimation of the signal’s instantaneous frequency,and generates random features around the estimated frequency to obtain sharper time-frequency images.In addition,the algorithm’s noise parameters are estimated using the time-frequency segmentation technique of circular STFT to achieve parameter adaptation.Numerical experiments on simulation and real signals show that the proposed decomposition method significantly improves decomposition accuracy and algorithm robustness compared to SRMD,and has excellent modal reconstruction capability.更多还原