系统工程与电子技术 ›› 2022, Vol. 44 ›› Issue (9): 2752-2759.doi: 10.12305/j.issn.1001-506X.2022.09.07

• 电子技术 • 上一篇    下一篇

基于互易点学习的LPI信号开集识别

韩啸, 陈世文*, 陈蒙, 杨锦程   

  1. 信息工程大学数据与目标工程学院, 河南 郑州 450001
  • 收稿日期:2021-09-29 出版日期:2022-09-01 发布日期:2022-09-01
  • 通讯作者: 陈世文
  • 作者简介:韩啸(1998—), 男, 硕士研究生, 主要研究方向为电子信号分析与处理、机器学习|陈世文(1974—), 男, 教授, 博士, 主要研究方向为电子信号分析与处理|陈蒙(1997—), 男, 硕士研究生, 主要研究方向为电子信号分析与处理、机器学习|杨锦程(1998—), 男, 硕士研究生, 主要研究方向为电子信号分析与处理、机器学习

Open-set recognition of LPI radar signal based on reciprocal point learning

Xiao HAN, Shiwen CHEN*, Meng CHEN, Jincheng YANG   

  1. School of Data and Target Engineering, Information Engineering University, Zhengzhou 450001, China
  • Received:2021-09-29 Online:2022-09-01 Published:2022-09-01
  • Contact: Shiwen CHEN

摘要:

针对现有采用时频图结合深度学习模型对低截获概率(low probability of intercept, LPI)雷达信号识别的方法在开集场景下会失效的问题, 提出一种基于互易点学习(reciprocal point learning, RPL)和阈值判断的雷达信号开集识别方法。通过RPL对特征空间进行优化, 使已知类和未知类信号样本在特征空间中分布不同, 最后确定合适的阈值进行开集识别。根据时频图的特点, 在特征提取网络中加入注意力机制使网络更关注图像能量聚集的有效部分。实验结果表明, 该方法在开放的电磁环境条件下具有良好的适应性。

关键词: 低截获概率信号, 开集识别, 深度学习, 互易点学习, 注意力机制

Abstract:

The existing method of low probability of intercept (LPI) radar signal recognition using time-frequency image combined with deep learning model will fail in the open-set scenario. A method of radar signal open-set recognition based on reciprocal point learning (RPL) and threshold judgment is proposed to solve this problem. The embedding space is optimized by RPL, so the signals of known classes and unknown classes are distributed differently in it. Finally, an appropriate threshold is determined for open-set recognition. According to the characteristics of time-frequency image, the attention mechanism module is added in feature extraction network to make it pay more attention to the effective part of energy-intensive. Experimental results show that the proposed method has good adaptability in open electromagnetic environment.

Key words: low probability of intercept (LPI) signal, open-set recognition, deep learning, reciprocal point learning (RPL), attention mechanism

中图分类号: