摘要
机载多功能雷达能根据作战需求灵活切换工作模式,在现代战场上执行空域扫描和目标跟踪等任务。其工作模式的动态变化不仅是评估威胁等级的关键依据,也为电子侦察带来了挑战。为提高对机载多功能雷达工作模式的识别效能,提出了一种基于集成深度学习的识别算法。该算法集成多个深度学习模型,并根据各模型的预测准确率动态调整权重,在确保整体性能最佳的前提下,采用“舍小取大”策略,提高集成模型的预测准确性和稳定性。仿真实验结果表明,基于舍小取大法的集成深度学习方法显著提高了识别效率,为机载多功能雷达工作模式的精准识别提供了有效解决方案。
In the modern battlefield,airborne multifunctional radars are capable of flexibly switching operating modes according to combat requirements,performing various tasks such as airspace scanning and target tracking.The dynamic switches of the operating modes not only serve as a critical basis for assessing threat levels but also pose challenges for electronic reconnaissance.To enhance the recognition efficiency of the operating modes of airborne multifunctional radars,this paper proposes a recognition algorithm based on integrated deep learning.The algorithm integrates multiple deep learning models and dynamically adjusts their weights based on the prediction accuracy of each model.On the premise of ensuring optimal overall performance,the strategy of"sacrificing the minor for the major"is adopted to enhance the predictive accuracy and stability of the integrated model.The simulation results demonstrate that this integrated deep learning method based on the strategy of"sacrificing the minor for the major"significantly enhances recognition efficiency,providing an effective solution for the precise identification of operating modes of airborne multifunctional radars.
作者
刘畅
王玮溪
王之腾
纪存孝
徐继伟
LIU Chang;WANG Weixi;WANG Zhiteng;JI Cunxiao;XU Jiwei(College of Communications Engineering,Army Engineering University of PLA,Nanjing 210007,China;Unit 75737 of PLA,Guangzhou 510800,China)
出处
《陆军工程大学学报》
2025年第3期57-64,共8页
Journal of Army Engineering University of PLA
基金
国家自然科学基金青年基金(62301604)。
关键词
机载多功能雷达
集成深度学习
舍小取大
工作模式识别
airborne multifunctional radars
integrated deep learning
sacrificing the minor for the major
recognition of radar operating modes
