首页|CNN和GBDT结合的雷达辐射源个体识别

CNN和GBDT结合的雷达辐射源个体识别

Individual Recognition of Radar Radiation Sources Combined with CNN and GBDT

扫码查看
针对雷达辐射源个体识别准确率低的问题,本文提出一种将一维卷积神经网络与梯度提升决策树相结合的雷达辐射源个体识别算法.首先基于线性调频信号构建雷达辐射源个体信号模型,然后介绍梯度提升决策树算法和一维卷积神经网络,接着搭建网络模型,将雷达辐射源信号的中频数据作为网络的输入,实现基于深度学习的雷达辐射源个体识别.通过实验结果得出,本文提出的雷达辐射源个体识别算法具有良好的识别准确率.当SNR≥5dB时,识别正确率达到99.5%以上.
In order to solve the problem of low accuracy on individual recognition of radar radiation sources,a rec-ognition algorithm method combined with one-dimensional convolution neural network and gradient boosting decision tree was proposed.Firstly,an individual signal model of radar radiation source was constructed based on linear fre-quency modulation signal;then the gradient boosting decision tree and one-dimensional convolution neural network were introduced;a network model was built after that,while the intermediate frequency data of radar signal was taken as the input of the network to realize the individual recognition of radar emitter based on deep learning.The experi-mental results showed that the radar radiation source individual recognition algorithm proposed in this paper had a rel-atively high recognition accuracy rate.When SNR≥5dB,the recognition accuracy rate reached over 99.5%.

individual identificationgradient boosting decision treeone-dimensional convolution neural networkdeep learning

郭瑞鹏、李显鹏、余建宇、吴皓

展开 >

西安电子工程研究所 西安 710100

个体识别 梯度提升决策树 一维卷积神经网络 深度学习

2024

火控雷达技术
西安电子工程研究所

火控雷达技术

影响因子:0.234
ISSN:1008-8652
年,卷(期):2024.53(3)
  • 3