Signal Feature Extraction and Identification Based on Module Maximum Matrix Singular Value Decomposition
LI Yong-hua, Zi Fang · Journal of Air Force Engineering University · 2005
The singular value decomposition of wavelet transformation module maximum value matrix is proposed. By using the obtained singular characteristic vector as the signal's feature, the features' dimension can be compressed and the computer auto-identification becomes easier and faster. The method also has the merit of time translation invariability. The simulation shows that the method is effective.