(The Processing Methods Of Analysis And Research Of The End Effects Of Hilbert-Huang Transform)

Hilbert-Huang transform (HHT) is a new signal analysis method on recent years. This new method consists of two parts:Empirical Mode Decomposition (EMD) and Hilbert transform. HHT can handle non-linear, non-stationary signal, has been widely used in engineering application. Hilbert-Huang transform is still in the development stage, The theory foundation is not enough to complete.so it has a lot of problems. For example, in most cases, the two endpoint of data sequence is not extreme points. When using interpolation technique to construct the envelope of the signal, Data sequence ends will appear " distortion" phenomenon, and this kind of distorted results continuous inward "pollution" with the shifting process. At the same time, Hilbert transform will bring the same phenomenon. Therefore the end effect of EMD process is the one of the unsolved difficult problems.In this paper, the main contents are as follows:1It has introduced the theoretical basis of Hilbert-Huang transform and its development history. It has displayed the advantages of HHT and expounded the research status at home and abroad and its development trend.2It has analyzed the causes on the end effects of HHT, Evaluation standard (observation of instantaneous frequency method and orthogonal index (IO) method).3It has researched two kinds of interpolation methods, different treatment methods to the different length data sequences. In the longer data sequence endpoint problem, this paper adopts continuously abandon ends data, to ensure that the final data sequence endpoint is extreme value for the resulting minimum distortion envelope. In the shorter data sequence endpoint problem, using the neural network and improved symmetric extreme method and the combination of two kinds of common interpolation methods to restrain end effects processing. At the same time, the neural network extension method, improved symmetric extreme method continuation method for dealing with a same signal, analysis of advantages and disadvantages.4Though analysis the experimental data, shows the Neural network and improved symmetric extreme combination method is effective.5The author summarized the work, and looked forward to the future research work.