A new method of detecting the small-signal with uncertain frequency based on clustering analysis

Various identification methods have been applied in the field of signal detection, and satisfied results are obtained. However, there is no good method to detect the randomly occurring small-signal with uncertain frequency, amplitude and phase in broad frequency band. In this paper, Hierarchical clustering algorithms and fuzzy-clustering algorithm are investigated to determine the efficiency of recognition, utilizing feature values of signal. Hierarchical clustering algorithm clusters the sample information and the to-be detected information. A comparative analysis of classes between the sample information and the to-be-detected information has been conducted. The new classes are obtained which correspond to the feature values of randomly occurring small-signal. In the signal recognition process, the fuzzy-clustering algorithm is used to eliminate the effects of both short-time random noise and the frequency or intensity change of the noise. The membership grade determines the credibility of detected new signal. Experiment results show that randomly occurring small-signal with uncertain frequency can be recognized in a complicated environment, and the test result will be better if the signal is multi-frequency information.