Research On Interference Identification Technology Of Radar Digital Radio Frequency Memory

In modern electronic warfare,radar holds the initiative in war as the most important radio detection equipment in the military field.As a new type of radar interference technology,Digital radio frequency memory(DRFM)has the ability of sampling the radar signal at a high speed and storing the replication for a long time.With this ability,DRFM can modulate and generate an interference signal with a high correlation of the original source signal to deceive and interfere with the target radar.The effective identification of digital frequency storage jamming signals is the key to radar anti-jamming.This paper is based on fingerprint characteristics of the transmitted signal caused by hardware differences of the radar system,and mainly studies three aspects of preprocessing,fingerprint feature extraction and classification recognition of inner-pulse fine feature in radar.The main contents of this article are as follows:1.In this paper,we introduce the interference mechanism and characteristics of digital radio frequency memory,and discuss the identification principle and method of digital radio frequency memory based on fingerprint characteristics of the radar emitter signal detection.2.The wavelet threshold de-noising method is used to preprocess the fingerprint signal in the time domain,and to extract features which are based on the impulse envelope frontier of the transient signal by Hilbert transform.As for the fingerprint characteristics generated by phase noise,FFT is used to analyze the power spectrum in the pulse signal,and select the fingerprint data in the frequency domain which is obtained from power spectrum falling edge curve.3.Making use of fractional polynomial to obtain the fitting curve coefficients as the radiation source fingerprint characteristic data where are from curve fitting the data in the time and frequency domain of the radar signal.Based on the comparison of statistical residual squares,the differences in curve fitting performance between integer polynomial and fractional polynomial are discussed.4.Using the characteristic data of the radar signal in time and frequency domain as the input of the classifier of the radar radiation source to design a fingerprintidentification device for the radiation source based on the support vector machine,and the parameters of the classifier are optimized through cross-validation grid search.In this paper,simulation experiments show the method of Specific Emitter Identification can effectively recognize the radar echo from its own or digital radio frequency memory.