Signal Detection and Identification Using the Wavelet Packet Decomposition and the Modified Wigner Distribution
| Autor: | Zi-Xian Yang, 楊子賢 |
|---|---|
| Rok vydání: | 2003 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 91 The power density of the non-stationary stochastic signal is a time-varying function, and it must be processed by appropriate time-frequency analysis methods for obtaining genuine information. The Wigner-Ville distribution (WVD) is one of the popular methods for these applications, but it has a serious problem for the cross-terms. Therefore, a revised WVD, namely the convolution type of Wigner-Ville distribution (CTWVD), is proposed. The essential variation of instantaneous spectrum of the signal can be shown clearly in the time-frequency plane, and the features of the signal can be easily identified. Furthermore, with huge data record in low signal-to—noise power ratio situation, the WVD is not an adequate processing method. In this thesis, the wavelet packet decomposition and CTWVD are combined to get a better time-frequency analysis method, and the Karhunen-Loeve transform (KLT) is used to simplify the computation for detection and identification of signals. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|