Radar emitter sorting and identification method and apparatus based on multiple synchronous compressive transformations

Abstract

The present invention belongs to the technical field of radar emitter identifications, in particular to a radar emitter sorting and identification method and apparatus based on multiple synchronous compressive transformations. The method comprises: obtaining a time-frequency image of a radar emitter signal through multiple synchronous compressive transformations; preprocessing the time-frequency image and extracting a texture characteristic and a moment characteristic of the time-frequency image and building a characteristic parameter set by combining a signal power spectrum parameter characteristic with a square spectrum complexity characteristic; and regarding the characteristic parameter set, sorting and identifying the signal with a support vector machine (SVM) sorter. The present invention solves the problems of a low rate in sorting and identifying the radar signal, high complexity and the like, under a condition of a currently low signal to noise ratio, such that different modulation types of radar signals can be identified accurately under the condition of the low signal to noise ratio. Further, the present invention has the advantages of good identification effect, high efficiency and good anti-noise property for complexly modulated types of radar signals as well as strong ability of adapting to a change in a parameter of the signal. Therefore, high identification performance can also be achieved under a small sample size, such that it has a certain value of being applied to a project.

Description

Technical field [0001] The invention belongs to the technical field of radiation source identification, and particularly relates to a radar radiation source sorting and identification method and device based on multiple synchronous compression transformations. Background technique [0002] With the rapid development of information technology, competition in the field of electronic countermeasures has become increasingly fierce, and the electromagnetic environment has become increasingly complex. Various new radar systems represented by phased array radars, ultra-wideband radars, and low probability of intercept radars have been continuously used. Radar radiator signals show the characteristics of time-frequency domain overlap, pulse parameter jumps, and diverse modulation methods. , Has brought great challenges to the subsequent sorting and identification. Traditional methods mainly rely on the carrier frequency (Radio Frequency, RF), pulse amplitude (Pulse Amplitude, PA), pulse...

AI Technical Summary

Problems solved by technology

Based on the Choi-Williams time-frequency distribution (Time Frequency Distribution, CWD), the time-frequency diagram of the signal is obtained, and the central moment and pseudo-Zernike moment features of the time-frequency image are extracted, and the identification of 8 types of radar signals is completed. However, the algorithm The recognition rate is not high at low signal-to-noise ratio

By performing singular value decomposition on the time-frequency image matrix of the signal, and then extracting statistical features such as singular value entropy, fractal dimension, box dimension and information dimension, it has strong recognition performance under low signal-to-noise ratio, but for Poor recognition rate of phase-encoded signals

Based on the improved local binary pattern operator (LBPV) to extract the texture features of the time-frequency image, a better recognition effect is achieved, but the complexity of the method is high

The above methods are all based on the time-frequency image of the signal obtained by CWD, and then extract the features of the time-frequency image for the sorting and recognition of radar signals; but CWD is a quadratic time-frequency tool, which cannot be used when dealing with nonlinear and non-stationary radar signals. Avoidance of cross-item interference will affect the accuracy of recognition

Images