基于有效数据的经验模态分解快速算法研究  

Study on Valid-Data-Based EMD Fast Algorithm

作　　者：胡劲松[1] 杨世锡[2] 

机构地区：[1]宁波工程学院电信学院,宁波315010 [2]浙江大学机能学院,杭州310027

出　　处：《振动．测试与诊断》2006年第2期119-121,共3页Journal of Vibration,Measurement & Diagnosis

基　　金：国家自然科学基金资助项目(编号:50205025)

摘　　要：在介绍了经验模态分解(简称EMD)方法的理论和算法基础上,为了提高EMD算法的速度,提出了基于有效数据的EMD快速算法,即通过EMD分解中止的计算区域限定于有效数据段来实现算法的提速。通过对非线性信号的实验研究表明,基于有效数据的EMD快速算法不但能显著提高算法的速度,而且还可以提高算法的精度。该研究成果能广泛地用于信号时频分析领域。To reduce the calculation time while using the EMD method, this paper presented a valid-data-based EMD fast algorithm. The characteristic of the algorithm was that, the EMD's stop condition was limited to the valid data.A nonlinear simulation signal was introduced to study the algorithm. Applied the basic EMD algorithm and the valid data-based's to the simulation signal respectively, the result showed that, the precision of decomposition was improved and the calculation time was reduced which using the latter algorithm. The conclusion was obtained that the valid data-based EMD fast algorithm was better than the basic's. The result of the study can be generally used in time-frequency analysis field.

关 键 词：有效数据 经验模态分解 快速算法 时频分析 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置] TH113.1[自动化与计算机技术—控制科学与工程]