Research On Recognition Of Emitter Based On Analysis In-pulse Feature

In the increasingly complicated electromagnetism environment, how to extract useful feature parameters is a key to achieve emitter identification. Because of modern radar signals perform the characteristics of complicated types, diverse parameters and the large variety scope of dynamic of frequency and so on, the conventional feature parameters of radar do not meet the need of practicality. Thus, researchers at home and abroad who study the recognition of emitter identification, apply themselves to analysis and extract more new and effective features of radar signals to realize identification of modern complicated radar signals. Features of radar signals include purposive modulated features and individual features. So the paper is to utilize distinct method to extract the two features respectively. Last, support vector machine is combined to achieve the different classification and individual identification.Around the in-pulse modulation characteristics and individual features interested in feature extraction and identification issue, this paper mainly studies the following aspects:1,In order to solve the in-pulse modulation feature extraction, the S transform is used to extract signal instantaneous frequency. According to different modulated signals with different instantaneous frequency characteristics, The method uses S transform to extract the instantaneous frequency modulation signal, and further, the instantaneous frequency of the cascade features as identifying characteristics are extracted, the final combination of support vector machines have been exported to modulated signals classification.2,In order to analyze the instability characteristics of emitter signals, the method based on signal stability analysis of individual characteristics of emitter signals is proposed. According to the signal amplitude envelope is characterized by a random amount of instability, the proposed method uses Allan variance to calculate the signal stability and difference of signal stability, and joints these two parameters as the features of individual identification. Experimental results show that the method has a better ability to identify different types of radar signals.3,In order to further realize individual characteristics of radar signals extraction and identification, specific emitter identification based on cyclic spectrum is studied. Firstly, the cyclic spectrum of signal is calculated, and then the best feature of radar signals are extracted and reduced dimensional by using Fisher's discriminant ratio and principle component analysis respectively. In the last, support vector machine is utilized to achieve specific emitter identification. Experimental results show that the method based on cyclic spectrum is more effective to describe characterization of radar emitters, it can effectively achieve individual identification.