Research On HHT Time-Frequency Resolution Algorithm And Its Performance Evaluation

With the developing of signal processing techniques, the higher resolution in time and frequency domain is required in the non-linear and non-stationary signal processing methods. Hence, how to choose an efficient time-frequency analysis method to gain a higher resolution is a valuable issue.HHT as a new adaptive time-frequency analysis method, it bases on the analysis and introduction of traditional time-frequency analysis. The HHT has a specific ability to recognize characteristics of the local instantaneous, completely avoid the drawback in traditional method which has fake harmonics when synthetizing original signals, while the time domain and frequency domain obtain a higher resolution in both directions at the same time, profoundly reveals the internal features of non-linear and non-stationary signals. However, the algorithm theory and the model are not strong enough. This paper deals with the study of the key difficulties in the theoretical basis and algorithm models and the performance of the algorithm is evaluated comprehensively. The followings are the main research points and features:1) The paper researches the basis function and intrinsic mode functions of HHT, and the condition of instantaneous frequenting physical model and physical meaning of it, building the physical model and mathematical model of intrinsic mode functions. Then the theory of empirical model decomposition is deeply discussed.2) A semi-quantitative modeling method SQME is proposed based on the qualitative of phase space to deal with the problem of serious end effect in HHT algorithm. According to the basic character of the problem, the paper tactfully infers the data flow qualitative methods to construct the volume space and symbolically expressing it which accurately describe the evolution processing of the system and effectively achieved the extension at both end point which provides a new ken in end effects. Then the paper gives some other ideas in mean evaluation of end point and terminal criteria.3) The thesis evaluates the performance with traditional time-frequency analysis technology. And the theory and experiment results prove that the HHT is an efficient time-frequency analysis technology. The provided theoretical proofs and experimental data results in the thesis sufficiently verify the performance of HHT method which is the fundamental of high time-frequency system when contracts with traditional methods in time-frequency resolution.