现代防御技术 ›› 2025, Vol. 53 ›› Issue (1): 52-62.DOI: 10.3969/j.issn.1009-086x.2025.01.006

• 空天防御体系与武器 • 上一篇    下一篇

基于度量学习的半监督空中目标作战意图识别

张晨浩1, 周焰1, 梁复台1,2, 周通3, 宋子豪1, 袁凯1   

  1. 1.空军预警学院,湖北 武汉 430014
    2.中国人民解放军31121部队,江苏 南京 210000
    3.中国人民解放军93950部队,青海 海西 816000
  • 收稿日期:2023-11-26 修回日期:2024-01-12 出版日期:2025-02-28 发布日期:2025-02-27
  • 作者简介:张晨浩(1996-),男,湖北襄阳人。博士生,研究方向为态势评估,深度学习。

Semi-supervised Air Targets Combat Intention Recognition Based on Metric Learning

Chenhao ZHANG1, Yan ZHOU1, Futai LIANG1,2, Tong ZHOU3, Zihao SONG1, Kai YUAN1   

  1. 1.Air Force Early Warning Academy,Wuhan 430014 China
    2.PLA 31121 Troops,Nanjing 210000,China
    3.PLA 93950 Troops,Haixi 816000,China
  • Received:2023-11-26 Revised:2024-01-12 Online:2025-02-28 Published:2025-02-27

摘要:

空中战场态势是对空中战场中所有参与方行动和状态的总体描述,而目标作战意图识别则为空中战场态势评估提供重要依据。为了解决在激烈对抗、快速演化的空中战场态势背景下,大量已标记的空中目标战场态势数据获取难度大的问题,提出了一种基于度量学习的半监督空中目标作战意图识别模型。该模型提供了一种从无标签样本中发掘潜在模式的方法,缓解了对大量标记数据的需求。模型通过目标时序数据编码器对目标序列数据进行降维并得到其嵌入表示。在此基础上,通过分别度量已标记的目标序列与意图类型、未标记的目标序列之间的相似度,计算对应的损失值。实验结果表明,在有标签样本不同占比为30%、40%和50%的情况下,该模型识别空中目标作战意图的准确率分别为86%、89%和91%。

关键词: 空中目标, 战场态势, 作战意图, 意图识别, 度量学习, 半监督学习

Abstract:

The air battlefield situation provides a general description of the actions and states of all participants, while target combat intention recognition serves as a critical foundation for air battlefield posture assessment. To address the challenge of acquiring a large amount of labeled air target battlefield data amidst intense confrontation and rapidly evolving scenarios, a semi-supervised air targets combat intention recognition model is proposed based on metric learning. The model offers a method for uncovering potential patterns from unlabeled samples, thus reducing the reliance on extensive labeled data. In this model, the target time-series data encoder reduces the dimensionality of target data and produces an embedded representation. Based on this, loss values are calculated by measuring the similarity between labeled target sequences and intention types, as well as between labeled and unlabeled target sequences. The experimental results demonstrate that the model achieves combat intention recognition accuracy rates of 86%, 89%, and 91% with labeled sample rates of 30%, 40%, and 50%, respectively.

Key words: air targets, battlefield situation, combat intention, intention recognition, metric learning, semi-supervised learning

中图分类号: