复杂对抗场景下的对空目标混合智能抗干扰研究  

Research on Hybrid Intelligent Anti-Interference against Air Targets in Complex Confrontation Scene

作　　者：张良 李少毅[3] 杨曦 田晓倩 Zhang Liang;Li Shaoyi;Yang Xi;Tian Xiaoqian(China Airborne Missile Academy,Luoyang 471009,China;Aviation Key Laboratory of Science and Technology on Airborne Guided Weapons,Luoyang 471009,China;School of Astronautics,Northwestern Polytechnical University,Xi’an 710072,China)

机构地区：[1]中国空空导弹研究院,河南洛阳471009 [2]航空制导武器航空科技重点实验室,河南洛阳471009 [3]西北工业大学航天学院,西安710072

出　　处：《航空兵器》2022年第1期22-28,共7页Aero Weaponry

基　　金：国家自然科学基金项目(61703337);航空科学基金项目(ASFC20191053002)。

摘　　要：目标识别与抗干扰技术已经成为决定精确制导武器性能优劣的关键技术。本文针对复杂对抗场景下红外空空导弹作战特点,分析了其目标识别与抗干扰发展需求,提出了融合传统算法与深度学习的混合智能抗干扰算法。该算法充分利用传统算法在确定场景下的高可靠性优势与深度学习算法在复杂场景下的高维特征提取能力,最大化挖掘了导弹探测的场景信息,对于提高系统抗干扰能力具有重要意义。在此基础上,构造了算法测试训练的空战数据集,覆盖了典型的空战作战场景。实验结果表明,相同特征融合条件下,典型场景混合智能抗干扰算法的全程抗干扰概率达到了71.56%,比传统算法提高了15.77%,验证了算法的有效性。Target recognition and anti-interference technology have become key technologies that determine the performance of precision guided weapons.Aiming at the combat characteristics of infrared air-to-air missiles in complex confrontation scene,this paper analyzes its target recognition and anti-interference development needs,and proposes a hybrid intelligent anti-interference algorithm that combines traditional algorithms and deep learning.This algorithm makes full use of the high reliability advantages of traditional algorithms in certain scene and the high-dimensional feature extraction capabilities of deep learning algorithms in complex scene,maximizing the mining of missile detection scene information,which is of great significance for improving the system’s anti-interference capability.On this basis,an air combat data set for algorithm testing and training is constructed,covering typical air combat scene.The experimental results show that under the same feature fusion conditions,the full anti-interference probability of the hybrid intelligent anti-jamming algorithm in typical scene reaches 71.56%,which is 15.77%higher than the traditional algorithm,which verifies the effectiveness of the algorithm.

关 键 词：红外空空导弹 智能抗干扰 空战数据集 特征提取 目标识别 深度学习 

分 类 号：TJ760.2[兵器科学与技术—武器系统与运用工程] V271.4[航空宇航科学与技术—飞行器设计]