基于多特征融合网络的空袭目标运动状态辨识

梁桐嘉; 郑欣磊; 刘雅梁; 范继

doi:10.12382/bgxb.2025.0133

您当前的位置：

首页 >

文章列表页 >

基于多特征融合网络的空袭目标运动状态辨识

更新时间：2026-04-24

- 基于多特征融合网络的空袭目标运动状态辨识
- Recognition of Airstrike Target Motion States based on Multi-feature Fusion Network
- 兵工学报 2026年47卷第2期页码：250133
- 作者机构：
  
  西北机电工程研究所，陕西咸阳 712099
- 作者简介：
  
  通信作者邮箱：893912165@qq.com
- 基金信息：
- DOI：10.12382/bgxb.2025.0133
  中图分类号： TJ01
- 收稿：2025-02-28，
  
  网络首发：2025-12-25，
  
  纸质出版：2026-02-28
- 稿件说明：
移动端阅览
梁桐嘉, 郑欣磊, 刘雅梁, 等. 基于多特征融合网络的空袭目标运动状态辨识[J]. 兵工学报, 2026,47(2):250133.

LIANG Tongjia, ZHENG Xinlei, LIU Yaliang, et al. Recognition of Airstrike Target Motion States based on Multi-feature Fusion Network[J]. Acta Armamentarii, 2026, 47(2): 250133.
梁桐嘉, 郑欣磊, 刘雅梁, 等. 基于多特征融合网络的空袭目标运动状态辨识[J]. 兵工学报, 2026,47(2):250133. DOI： 10.12382/bgxb.2025.0133.

LIANG Tongjia, ZHENG Xinlei, LIU Yaliang, et al. Recognition of Airstrike Target Motion States based on Multi-feature Fusion Network[J]. Acta Armamentarii, 2026, 47(2): 250133. DOI： 10.12382/bgxb.2025.0133.

摘要

针对防空火力控制系统对空袭目标运动状态辨识能力欠缺所导致的轨迹预测精度受限及毁伤效能低下问题，提出一种面向火控系统的分类范式以及基于多特征融合神经网络（Multi-feature Fusion Neural Network，MFF-Net）的运动状态智能辨识方法。通过构建三维近似熵特征描述子，系统量化空袭目标轨迹的时空复杂度特性；提出卷积注意力耦合机制，实现轨迹复杂度特征与卷积长短期记忆网络提取的时空关联特征之间的自适应融合；引入一维卷积模块强化时序动态特征的层次化提取，与注意力得分进行二次融合加强模型辨识能力。实验结果表明：基于雷达实测数据与典型运动模式仿真数据集构建的混合测试集上MFF-Net在稳态线性、盘旋等四类典型运动范式的辨识准确率达到96.56%，较长短期记忆网络或一维卷积网络等时序网络相比有着显著提升，验证了该方法对复杂轨迹运动状态辨识的有效性，为复杂战场环境下空袭目标运动模式在线辨识提供了融合特征量化与深度学习的复合框架。

Abstract

Aiming at the problems of limited trajectory prediction accuracy and low damage efficiency caused by the deficiency of recognition capability of air defense fire control system for the motion states of air strike targets

a motion state intelligent recognition method based on multi-feature fusion neural network(MFF-Net)is proposed. The spatiotemporal complexity characteristics of air strike target trajectory are systematically quantified by constructing a 3D approximate entropy feature descriptor. A convolutional attention coupling mechanism is proposed to achieve the adaptive fusion between trajectory complexity features and spatiotemporal correlation features extracted by convolutional long short-term memory network. An one-dimensional convolution module is introduced to enhance the hierarchical extraction of temporal dynamic features

and the secondary fusion with attention scores is performed to improve the identification ability of model. Experimental verification shows that

on a mixed test set constructed based on radar-measured data and typical motion mode simulation data sets

the recognition accuracy of MFF-Net in four typical motion modes such as steady-state linear motion and circling motion reaches 96.56%

which is significantly improved compared with temporal networks such as long short-term memory networks or one-dimensional convolutional networks. It is proved that the proposed method is effective in the recognition of complex trajectory motion states

and provides a composite framework that integrates the feature quantization and deep learning for online recognition of air strike target motion modes in complex battlefield environments.

关键词

Keywords

references

申程,张连超,张卓,等.轻量化弹道解算系统火控修正和瞄准线滤波预测[J].兵工学报,2024,45(2):429-442.

SHEN C, ZHANG L C, ZHANG Z, et al. Fire control correction and sighting line filtering prediction of lightweight ballistic calculation system[J]. Acta Armamentarii, 2024, 45 (2):429-442. (in Chinese)

郭瑞卿.无人机探测雷达数据处理算法研究与实现[D].西安：西安电子科技大学,2023.

GUO R Q. Research and implementation of data processing algorithms for UAV detection radar[D]. Xi'an: Xidian University, 2023. (in Chinese)

ZEINEB D, AMAL H B, CHOKRI A B. Fire object detection and tracking based on deep learning model and Kalman filter[J]. Arabian Journal for Science and Engineering, 2023, 49 (3):3651-3669.

陈清阳,辛宏博,鲁亚飞,等.高亚声速无人飞行器滑跑纠偏控制[J/OL].北京航空航天大学学报,2024(2024-01-19)[2024-09-12].

CHEN Q Y, XIN H B, LU Y F, et al. Runway deviation correction control for high-subsonic unmanned aerial vehicles[J/OL]. Journal of Beijing University of Aeronautics and Astronautics, 2024(2024-01-19)[2024-09-12]. (in Chinese)

SUN P, YANG G L, DING M L, et al. System design for detecting cannon launching times[C]∥Proceedings of 2017 1st International Conference on Electronics Instrumentation and Information Systems. Harbin, China: IEEE, 2017: 1-4.

杜青峰,荆武兴,高长生,等.考虑滑块动态特性的变质心四旋翼无人机双回路控制[J].北京航空航天大学学报,2024, 50(3):861-873.

DU Q F, JING W X, GAO C S, et al. Dual-loop control of variable center-of-mass quadrotor with consideration of slider dynamic characteristics[J]. Journal of Beijing University of Aeronautics and Astronautics, 2024, 50 (3): 861-873. (in Chinese)

BLOM H A P, BAR-SHALOM Y. The interacting multiple model algorithm for systems with Markovian switching coefficients[J]. IEEE transactions on Automatic Control, 1988, 33 (8): 780-783.

殷之平,黄勇胜,田珈玮,等.一种基于二维轨迹与时序重要点的飞机机动划分与识别方法[J].航空工程进展,2025,16(1):45-53.

YIN Z P, HUANG Y S, TIAN J W, et al. A method for aircraft maneuver segmentation and recognition based on two-dimensional trajectory and temporal key points[J]. Advances in Aeronautical Science and Engineering, 2025, 16(1): 45-53. (in Chinese)

李威,卢盈齐,范成礼.基于双层随机森林的空袭目标识别[J].兵器装备工程学报,2023,44(5):207-213.

LI W, LU Y Q, FAN C L. Air raid target recognition based on double-layer random forest[J]. Journal of Ordnance Equipment Engineering, 2023, 44(5): 207-213. (in Chinese)

蔡远利,邓逸凡,苏悦华.高超声速飞行器LSTM弹道分类与预报方法[C]∥第21届中国系统仿真技术及其应用学术年会论文集.中国,云南：中国自动化学会系统仿真专业委员会,中国仿真学会仿真技术应用专业委员会,2020:311-315.

CAI Y L, DENG Y F, SU Y H. LSTM-based trajectory classification and prediction for hypersonic vehicles[C]∥Proceedings of the 21st China Conference on System Simulation Technology and its Application. Yunnan, China: System Simulation Professional Committee of the Chinese Association of Automation, Simulation Technology Application Professional Committee of the Chinese Society of Simulation, 2020: 311-315. (in Chinese)

宋波涛,许广亮.基于LSTM与1DCNN的导弹轨迹预测方法[J].系统工程与电子技术,2023,45(2):504-512.

SONG B T, XU G L. Missile trajectory prediction method based on LSTM and 1DCNN[J]. Systems Engineering and Electronics, 2023, 45(2): 504-512. (in Chinese)

孙艺升.基于深度学习的飞机飞行轨迹目标识别[C]∥第一届空中交通管理系统技术学术年会论文集.中国,江苏：中国指挥与控制学会,2018:288-293.

SUN Y S. Aircraft flight trajectory target recognition based on deep learning[C]∥ Proceedings of the 1st Academic Conference on Air Traffic Management System Technology. Jiangsu, China:China Command and Control Society, 2018: 288-293. (in Chinese)

李琳,曾雅琴,朱惠民,等.基于LSTM-GBSVDD模型的飞行轨迹异常检测方法[J].兵工学报,2025,46(5):89-99.

LI L, ZENG Y Q, ZHU H M, et al. Anomaly detection method for flight trajectories based onLSTM-GBSVDDmodel[J]. Acta Armamentarii, 2025, 46(5): 89-99. (in Chinese)

王晓博,朱靖,崔玉婕,等.一种目标轨迹聚类和分类方法[C]∥第十七届全国信号和智能信息处理与应用学术会议论文集.中国,重庆：中国高科技产业化研究会智能信息处理产业化分会,2023:43-45+90.

WANG X B, ZHU J, CUI Y J, et al. A method for target trajectory clustering and classification[C]∥ Proceedings of the 17th National Conference on Signal and Intelligent Information Processing and Application. Chongqing, China: Intelligent Information Processing Industrialization Branch of the China High-Tech Industrialization Research Association, 2023: 43-45+90. (in Chinese)

SHI X J, CHEN Z R, WANG H, et al. Convolutional LSTM network: A machine learning approach for precipitation nowcasting[J]. Advances in Neural Information Processing Systems, 2015, 28.

PINCUS S M. Approximate entropy as a measure of system complexity[J]. Proceedings of the National Academy of Sciences of the United States of America, 1991, 88(6):2297-2301.

王子江,张兆毅,樊友平,等.基于近似熵变化量判据的混合直流输电系统纵联保护方案[J].电力系统保护与控制, 2024,52(21):1-12.

WANG Z J, ZHANG Z Y, FAN Y P, et al. A pilot protection scheme for hybrid DC transmission systems based on the criterion of approximate entropy change[J]. Power System Protection and Control, 2024, 52(21): 1-12. (in Chinese)

夏英,陈航.融合Transformer和卷积LSTM的轨迹分类网络[J].重庆邮电大学学报(自然科学版),2024,36(1):29-38.

XIA Y, CHEN H. Trajectory classification network integrating Transformer and convolutional LSTM[J]. Journal of Chongqing University of Posts and Telecommunications (Natural Science Edition), 2024, 36(1) : 29-38. (in Chinese)

KURMAR A, ISLAM T, SCKIMOTO Y, et al. Convcast: An embedded convolutional LSTM based architecture for precipitation nowcasting using satellite data[J]. Plos one, 2020, 15 (3):e0230114.

CHEN W, SHI K. A deep learning framework for time series classification using relative position matrix and convolutional neural network[J]. Neurocomputing, 2019, 359: 384-394.

韩子鹏.弹箭外弹道学[M].北京：北京理工大学出版社, 2014:74-86.

HAN Z P. Exterior ballistics of projectiles and rockets[M]. Beijing: Beijing Institute of Technology Press, 2014: 74-86. (in Chinese)

任济寰,吴祥,薄煜明,等.基于增强上下文信息长短期记忆网络的弹道轨迹预测[J].兵工学报,2023,44(2) :462-471.

REN J H, WU X, BO Y M, et al. Ballistic trajectory prediction based on context-enhanced long short-term memory network[J]. Acta Armamentarii, 2023, 44(2): 462-471. (in Chinese)

杨书涵,韦楠楠,张兴敢.基于残差网络ResNet18-SVM的弹道中段目标识别[J].现代雷达,2024,46(4):8-14.

YANG S H, WEI N N, ZHANG X G. Midcourse target recognition based on ResNet18-SVM[J]. Modern Radar, 2024, 46(4): 8-14. (in Chinese)

李江,冯存前,王义哲,等.基于深度学习的弹道目标智能分类[J].系统工程与电子技术,2020,42(6):1226-1234.

LI J, FENG C Q, WANG Y Z, et al. Intelligent classification of ballistic targets based on deep learning[J]. Systems Engineering and Electronics, 2020, 42(6): 1226-1234. (in Chinese)

VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[J]. Advances in neural information processing systems, 2017, 30.

DOSOVITSKIY A. An image is worth 16x16 words: Transformers for image recognition at scale[C]//Proceedings of 2021 9th International Conference on Learning Representations. Vienna, Austria:[s. n.], 2021:611-631.

WANG Z, OATES T. Encoding time series as images for visual inspection and classification using tiled convolutional neural networks[C]∥Proceedings of 2015 29th American Association for Artificial Intelligence. Texas, US: Academic Press_Elsevier Science, 2015, 1: 91-96.

浏览量

下载量

CNKI被引量

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于神经网络的电静液起竖系统换级碰撞补偿积分鲁棒控制

基于双层多目标优化的翼身融合水下滑翔机设计

基于串联神经网络的雷达罩抗爆性能厚度等效模型

基于神经网络的密集城市建筑群爆炸载荷快速预测