基于互易点学习的LPI信号开集识别. (Chinese)
In: Systems Engineering & Electronics, Jg. 44 (2022-09-01), Heft 9, S. 2752-2759
academicJournal
Zugriff:
The existing method of low probability of intercept (LPI) radar signal recognition using timefrequency 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 d让ferently 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. [ABSTRACT FROM AUTHOR]
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, 提出一种基于亙易点学习(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。. [ABSTRACT FROM AUTHOR]
Copyright of Systems Engineering & Electronics is the property of Journal of Systems Engineering & Electronics Editorial Department and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Titel: |
基于互易点学习的LPI信号开集识别. (Chinese)
|
---|---|
Autor/in / Beteiligte Person: | 韩啸 ; 陈世文 ; 陈蒙 ; 杨锦程 |
Zeitschrift: | Systems Engineering & Electronics, Jg. 44 (2022-09-01), Heft 9, S. 2752-2759 |
Veröffentlichung: | 2022 |
Medientyp: | academicJournal |
ISSN: | 1001-506X (print) |
DOI: | 10.12305/j.issn.1001-506X.2022.09.07 |
Schlagwort: |
|
Sonstiges: |
|