Research On Key Technology Of Radar Emitter Signal Recognition

Radar emitter signal recognition is one of the key technologies in radar reconnaissance and radar countermeasure system.The recognition level has been an important symbol to measure the technology level of the radar reconnaissance and radar countermeasures equipment.With the rapid development of radar technology and the wide use of the new modern radar,the density and the complexity of radar signals are greatly improved.It makes the conventional recognition method based on the traditional five parameters,which include Radio Frequency(RF),Time of Arrival(TOA),Pulse Width(PW),Pulse Amplitude(PA),and Angel of Arrival(AOA),can not meet the needs of signal recognition in current complex electromagnetic environment.Therefore,it is one of the urgent tasks to study the parameter measurement optimization methods and new feature extraction methods for radar emitter signal recognition in the complex electromagnetic environment.This paper makes a systematic and in-depth research on the improvement of existing parameter measurement methods,new feature parameter extraction methods and the comprehensive evaluation of feature parameters.The main contributions of this paper are as below.(1)The problem of accuracy degradation of signal detection,TOA and other parameters measurement under low signal-to-noise ratio(SNR)is studied.Firstly,a correlation processing method for signal detection is proposed to improve the detection sensitivity of the system.Then,on the advantages of Lifting Wavelet Transform in edge detection and locating signal mutation point,which is utilized to detect the start and the end time of signal arrival.Its purpose is to realize the high accurate estimation of TOA and PW.Meanwhile,aiming at the problem of parameter estimation error caused by the influence of multi-path signals,a new method for multi-path signals parameter estimation is proposed based on sparse decomposition theory.This method makes full use of the sparse characteristics of signal in time domain.It can achieve the accurate estimation of delay component and delay attenuation through sparse decomposition and signal reconstruction.Based on the recorded data,it is verified that the proposed method has higher estimation accuracy and time delay resolution.(2)The method of fingerprint feature extraction in radar emitter signal is studied.Firstly,the description parameters of fingerprint features such as the envelope edge feature,fine spectrum feature and fractal feature of radar emitter signals are studied.Then,the description parameters which can describe these three fingerprint features are measured based on the simulated data.From these measured data,the description parameters which can stably and uniquely describe these three fingerprint features are mined.The mathematical models and the extraction methods and the differences of reliability analysis among these description parameters are given.Finally,simulations are demonstrated to verify the reliability and robustness of the fingerprint feature parameters extracted in this paper based on the different signals transmitted from the same radar,and the application scenarios of each fingerprint feature are given.(3)The method of feature extraction based on the radar coherence is studied.On the fact that most active radars are coherent system,the feasibility of utilizing the radar coherence as the feature parameter for signal sorting has been proved,and the model of the coherent characteristics is given.Then,the instantaneous correlation function and Discrete Fourier Transform(DFT)are taken to analyze the coherent and non coherent signals respectively.The image processing technology is introduced to extract the feature.Then,sidelobe peak value to main peak value ratio(SMR)and normalized central moment(NCM)are defined to describe the difference between the coherent and non coherent signals and the decision thresholds of SMR and NCM are given based on the simulations.Finally,two features are used for simulations in emitter signal recognition to check the validity.Simulation results demonstrate that these two features proposed in this paper not only have good recognition performance in low SNR,but also they are insensitive to the change of the signal modulation types.(4)The method of feature evaluation and comparison in radar emitter signal recognition is studied.Currently,there are many new proposed feature parameters in radar emitter signal recognition.However,the comprehensive effectiveness of these features is not clear.How to choose these new features is an urgent problem to be solved in emitter recognition.Firstly,a comprehensive evaluation model of radar emitter signal features based on analytic hierarchy process(AHP)is proposed.Then,the set pair analysis(SPA)theory is introduced to solve problems in the construction of the decision matrix based on AHP,as it relies heavily on expert's subjective experience.The set pair decision matrix is constructed according to the identity of the measured data and the theoretical data,which can realize the features evaluation more objectively.Finally,the effectiveness of the proposed evaluation model based on SAHP has been checked according to the simulation data of radar emitter signals.(5)The method of radar working mode recognition based on radar emitter knowledge base is studied.As for the purpose of radar emitter signal recognition is to study the operational intention of radar.Meanwhile,the inversion of operational intention can be realized by the radar working mode recognition.Thus,the methods of working mode recognition for known and unknown radars are studied respectively in this paper.For the working modes recognition of known radars,the feature parameters that can be used for working modes recognition are mined by data mining from the radar knowledge base,and the priority of feature parameters are evaluated by AHP.Then,the convolutional neural network(CNN)classifier is introduced into radar working model recognition.Finally,experiments are carried out to check the validity and the robustness of the proposed method according to the data recorded from different active radars.For the working modes recognition of unknown radars,AHP is used to recognize the radar model based on the radar knowledge base.Then,the work mode recognition is evaluated according to the process of known radars and the feasibility of the proposed method is verified based on the simulated data.It provides a new idea for the work mode recognition of unknown radars.