Research On Individual Identification Technology Of Communication Radiation Source

With the diversification of radio network attack forms,the identification of communication emitter equipment has become an important research topic in the field of communication security.At present,under the condition of low SNR,the individual identification rate of emitter is low.It cannot meet actual application requirements.At the same time,in practice,the receiving terminal may receive a large amount of unlabeled data.Due to the lack of training data,it is not possible to use supervised learning to identify and classify it.This article studies fingerprint extraction and recognition algorithms under low signal-to-noise ratio to improve the recognition rate of emitter equipment under low signal-to-noise ratio environments.At the same time,for the detection of unknown target categories,this article explores a scheme for detecting unknown target types of equipment.The main work of this article is as follows:This article designs time-frequency fusion fingerprint extraction scheme.It extracts the frequency domain characteristics of the Wigner distribution.Meanwhile it extracts the Time-domain statistical characteristics,and then it fuses them.For the extracted high-latitude fingerprint features,this article uses multiple discriminant analysis to reduce the dimensionality,and finally constitute the fingerprint feature of the radiation source.The results prove the effectiveness of the time-frequency fusion feature.Under different signal-to-noise ratios,the recognition rate using fingerprint features fused in the time domain and frequency domain is not much different.This article designs a fingerprint extraction scheme based on Variational mode decomposition(VMD).First,the VMD decomposition algorithm is implemented and simulated,and we analyzed the problems that occurred in the process of VMD decomposing the actual acquisition signal.Here,an optimized variational modal decomposition algorithm Reset variational mode decomposition(VMDR)is proposed to strengthen the suppression of the modal aliasing phenomenon of the decomposed signal,and reduce the influence of modal aliasing on the fingerprint extraction of the radiation source.Based on this,a box-dimensional VMDR fingerprint extraction scheme is designed.Experimental results illustrate that under different signal-to-noise ratios,noise has similar effects on the recognition rates of VMD,Empirical Mode Decomposition Empirical Mode Decomposition(EMD)and Local mean decomposition Local mean decomposition(LMD),but the recognition rate of fingerprint features extracted based on VMD is higher,and the fingerprint information extracted by VMD algorithm is more effective.This article Combines Teager energy operator to extract the fingerprint feature of the radiation source.Then,it through the energy analysis of Teager operator and the feature extraction of the box dimension of the decomposed modal signal,and the fingerprint feature of the radiation source device is formed.The results show that in a low signal-to-noise ratio environment,the proposed VMDR-Teager has a higher recognition rate than the box-dimensional VMDR.On this basis,the support vector domain is used to describe the hypersphere of the known target class.An unknown radiation source detection scheme based on Support Vector Data Description(SVDD)model is designed.Firstly,the SVDD models of known categories are trained by the extracted VMDR-Teager fingerprints,and then the unknown devices are detected.The experimental results show that the unknown radiation source equipment which does not belong to the target class can be detected by this scheme.