Emitter Individual Recognition Based On Fine Feature Analysis

Emitter individual identification based on fine feature analysis originated from non-cooperative communication field. Fine feature refers to the identity of the individual differences embedded in the received signal which caused by the influence of signal individual, device individual or transmission channels.Traditional emitter recognition aims to obtain the transmitted information, while the purpose of emitter individual identification is extracting the subtle difference embedded in the received signal to recognize individual signal and get the informative information of enemy emitter sources. How to choose an effective signal processing method, realizing real-time,accurate recognition and extracting subtle difference is a hot research topic in recent years. To solve this problem, we deeply studied emitter signal and device individual identification method based on fine feature analysis. This paper includes the following parts :System model for emitter individual identification was established, the mechanism of fine feature was analyzed, typical emitter fine feature extraction methods were studied, which provided solid theoretical foundations for the study of new fine feature analysis methods.Entropy feature extraction based emitter identification method was studied. Since different information entropy describes signal differences differently, we proposed multi-dimensional information entropy recognition scheme. The performance of Euclidean distance classifier, artificial intelligence classifier and feature weighting method were compared.Parameter estimation based individual identification method was studied. First extracting individual features of different parameter modulated signal, then validating feature effectiveness under unstable SNR conditions. Here nonlinear fitting methods was adopted to recognize LFM emitter signal individuals.Since transmitted signal contains abundant individual information for the nonlinear differences of communication devices itself, nonlinear characteristics analysis of oscillator was performed. Traditional phase noise spectrum local feature extraction method is susceptible to noise interference, we presented a scattered difference feature extraction scheme and verified its recognition performance.