DCNN (Deep Convolutional Neural Network) based radar radiation source identification method

Abstract

The invention relates to a DCNN based radar radiation source identification method. A one-dimensional (1D) waveform signal generated by a radar radiation source model is converted into a spectrogram by means of short-time Fourier transform (STFT), and different network structures are designed for the 1D waveform signal and spectrogram. Waveform signals are generated according to data generated by8 types of radar radiation source signal models; the waveform signals are transformed into the spectrogram by STFT, and data enhancement as well as transformation from waveform to images is realized;the waveform signals and spectrogram are input to the DCNN, and convolution and pooling are carried out to obtain characteristic information thereof; and the extracted characteristic information is input to softmax for classification. Via the method, the radar radiation source signals are classified and identified more accurately, and a radar signal identification result is better.

Description

technical field [0001] The invention belongs to the field of computer technology, in particular to the field of radar radiation source signal identification, in particular to a radar radiation source identification method based on a deep convolutional neural network. Background technique [0002] One of the important functions of radar radiation source signal identification in today's electronic warfare is mainly to compare the parameters of the measured radar radiation source with the pre-accumulated parameters through the differences in the emission signals of each radiation source to compare the parameters of each radiation source. Identify, so as to realize the interception, positioning, analysis and identification of the radar signal, and then distinguish the received signal, determine which radar radiation source it is from, and finally complete the identification of the radar radiation source signal. So radar emitter identification algorithm is a very challenging prob...

AI Technical Summary

Problems solved by technology

The BP neural network is applied to the radiation source identification algorithm, but due to the complex and multivariable nonlinear combination of multiple neurons, it increases the difficulty of the neural network in the radar identification algorithm, prompting researchers to find another way

Then it is proposed that support vector machine (SVM) is applied to the signal recognition of radar radiation source. SVM has better recognition ability when dealing with small sample data, but when dealing with voice or image problems, SVM will have obvious deficiencies. The generalization ability of

Images