Study On Time-Frequency Analysis Algorithm Based On Hilbert-Huang Transform Technology

Time-frequency signal analysis as a powerful tool of non-stationary signal analysisis becoming a hotspot on modern signal disposal research. The Hilbert-Huang transform(HHT) is a new theoryof time-frequency signal analysis. Both the technologyof Hilbertspectrum and EMD are considered in it. This theory could decompose complex signalsinto numbered IMF so that instantaneous frequency is reasonable in physics. There is awide range of application in telecommunications, radar, seismic, speech recognition,and so on.This paper is mainly to do research on the principle of this new theory, improve theproblems of algorithms of this project and to attest the value of HHT as a powerful toolof non-stationary signal analysis. A new method of ending effect which is called endsextremum periods extending method is advanced based on former theories about endseffect disposal. Compared with many methods of this field, we have proved theadvantages of this method, amended the quite conditions on the basis of manyexperiments and optimized the algorithms amply. In addition, we have founded aphenomenon in the simulation test, which is the affect of sampling rate for Hilbertspectrum. Analyzing this phenomenon, we advanced the concept with the name of thewhole period sampling and named the phenomena the whole period samplingphenomena. And then we put forward the precondition of Hilbert-Huang transformwhich can be used in time-frequency signal analysis successfully. Finally we make aconclusion that the Hilbert-Huang transform can be used in time-frequency signalanalysis successfullyin engineering.