西安电子科技大学学报 ›› 2021, Vol. 48 ›› Issue (2): 49-54.doi: 10.19665/j.issn1001-2400.2021.02.007

• 雷达技术进展专题 • 上一篇    下一篇

采用乘性RNN的雷达HRRP目标识别

徐彬1(),张永顺1(),张秦1(),王富平2(),郑桂妹1()   

  1. 1.空军工程大学 防空反导学院,陕西 西安 710051
    2.西安邮电大学 通信与信息工程学院,陕西 西安 710121
  • 收稿日期:2020-09-30 修回日期:2020-11-11 出版日期:2021-04-20 发布日期:2021-04-28
  • 作者简介:徐彬(1988—),男,讲师,博士,E-mail: xb3221@163.com|张永顺(1961—),男,教授,博士,E-mail: zysgcdx@sina.com|张秦(1974—),男,教授,博士,E-mail: kinzh@263.net|王富平(1987—),男,讲师,博士,E-mail: cyx0908715@163.com|郑桂妹(1987—),男,讲师,博士,E-mail: zhengguimei1987@163.com
  • 基金资助:
    国家自然科学基金(61802305);国家自然科学基金(61971438);国家自然科学基金(61631019);陕西省青年托举人才项目(20180109);陕西省自然科学基金(2019JM-155)

Radar HRRP target recognition based on the multiplicative RNN model

XU Bin1(),ZHANG Yongshun1(),ZHANG Qin1(),WANG Fuping2(),ZHENG Guimei1()   

  1. 1. Air and Missile Defence college,Air Force Engineering University,Xi’an 710051,China
    2. School of Communication and Information Engineering,Xi’an University of Posts & Telecommunications,Xi’an 710121,China
  • Received:2020-09-30 Revised:2020-11-11 Online:2021-04-20 Published:2021-04-28

摘要:

传统的高分辨距离像识别方法没有考虑时序相关性,且高分辨距离像的方位敏感性导致样本的时序性发生变化。因此,提出一种乘性循环神经网络模型。该算法首先将高分辨距离像样本转化为序列形式,用于考虑距离单元间的相关性;其次,为了缓解方位敏感性导致的高分辨距离像时序变化与参数固定模型不匹配的问题,模型根据输入数据自适应地选择对应的参数,并对高分辨距离像序列提取稳健的时序信息;最后,采用投票策略将所有时刻的信息进行融合,输出样本类别。采用实测数据的实验结果表明,当前的模型能够有效地提取可分性特征并识别目标。

关键词: 雷达自动目标识别, 乘性循环神经网络, 高分辨距离像, 方位敏感性, 时序相关性

Abstract:

The traditional HRRP recognition methods do not consider the temporal correlation,and the azimuth sensitivity of HRRP results in the temporal variation of the samples.This paper proposes a multiplicative recurrent neural network.In this paper,HRRP samples are converted into the sequence form first,which is used to consider the correlation between range cells.In order to alleviate the mismatch between the HRRP sequence variation caused by azimuth sensitivity and the model with fixed parameters,the model adaptively selects the corresponding parameters according to the input data,and extracts robust features from the HRRP sequence.Finally,the voting strategy is adopted to fuse the information at all time steps and predict the sample categories.Experimental results with measured data show that the current model can effectively extract discriminative features and identify targets.

Key words: radar automatic target recognition, multiplicative recurrent neural network, high resolution range profile, target-aspect sensitivity, temporal correlation

中图分类号: 

  • TN957.52
Baidu
map