强对抗条件下基于BO-BI-LSTM的目标轨迹预测方法

吴泽; 谭目来; 丁达理; 郭政委

doi:10.15892/j.cnki.djzdxb.2025.06.002

弹箭与制导学报 >

2025 , Vol. 45 >Issue 6: 978 - 985

DOI: https://doi.org/10.15892/j.cnki.djzdxb.2025.06.002

学术文章

强对抗条件下基于BO-BI-LSTM的目标轨迹预测方法

吴泽 ^,¹ ,
谭目来 ^,¹^,^* ,
丁达理 ¹ ,
郭政委 ²

展开

¹ 空军工程大学航空工程学院,陕西西安 710038
² 94865部队,浙江杭州 310000

谭目来(1998—),男,学员,博士研究生。E-mail:kgdtml@163.com。

吴泽(2002—),男,学员,硕士研究生。E-mail:824163756@qq.com。

收稿日期: 2024-07-17

网络出版日期: 2026-01-24

收起

Target Trajectory Prediction Method Based on BO-BI-LSTM under Strong Adversarial Conditions

WU Ze ^,¹ ,
TAN Mulai ^,¹^,^* ,
DING Dali ¹ ,
GUO Zhengwei ²

Expand

¹ Air Force Engineering University,Xi’an 710038,Shaanxi,China
² Unit 94865,Hangzhou 310000,Zhejiang, China

Received date: 2024-07-17

Online published: 2026-01-24

Fold

摘要

当前空战对抗条件下目标经常出现中大过载、强机动的现象,导致机动轨迹预测难度大。针对当前传统轨迹预测方法预测精度低、可预测时间短的问题,提出了一种基于贝叶斯优化的双向长短期记忆网络(BO-Bi-LSTM)多步轨迹预测方法。分析滑动预测方法,建立了在线滚动预测数学模型,解决了预测值的构造问题;利用贝叶斯优化方法自动对网络超参数进行寻优,迭代之后得到最优的超参数;对滚动优化滑动窗口长度进行分析,在经典滑动窗口长度中得出预测精度最高的滑动窗口长度;为检验该方法在机动轨迹上的预测性能,对一段经典机动飞行轨迹进行预测仿真,与其他3种神经网络预测模型进行对比,仿真结果证明贝基于叶斯优化的双向长短期记忆网络多步轨迹预测方法在预测精度上高于其他3种神经网络,在三维轨迹误差小于200m的精度能持续预测4.5s左右。

关键词： 轨迹预测; 双向长短期记忆网络; 在线滚动预测理论; 贝叶斯优化; 滚动优化滑动窗口长度

本文引用格式

吴泽 , 谭目来 , 丁达理 , 郭政委 . 强对抗条件下基于BO-BI-LSTM的目标轨迹预测方法[J]. 弹箭与制导学报, 2025 , 45(6) : 978 -985 . DOI: 10.15892/j.cnki.djzdxb.2025.06.002

Abstract

Under the current conditions of air combat confrontation,targets often behave medium and large overloads and strong maneuvers,and it is difficult to predict the trajectory of maneuver.In order to solve the problems of low prediction accuracy and short prediction time of the traditional trajectory prediction methods,a multi-step trajectory prediction method based on Bayesian optimization hyperparameters in bidirectional long short-term memory network (BO-Bi-LSTM)is proposed in this paper.The sliding prediction method is analyzed,and an online rolling prediction mathematical model is established to solve the problem of prediction value construction.The network hyperparameters are optimized automatically by using Bayesian optimization method,and the optimal hyperparameters are obtained after iteration several times.The length of the sliding window is analyzed,and the length of the sliding window with the highest prediction accuracy is obtained among the classical sliding window length.In order to test the prediction performance of this method on the maneuvering trajectory,a classical maneuvering flight trajectory is predicted and simulated in this paper,and compared with three other neural network prediction models,the simulation results prove that the bidirectional long and short time domain memory network multi-step prediction method with Bayesian optimization hyperparameters is higher in prediction accuracy than the other three neural networks.The accuracy of the 3D trajectory error is less than 200m,which can be predicted continuously for about 4.5s.

Key words： Trajectory prediction; Bi-LSTM; The online rolling recursive prediction theory; Bayesian optimization hyperparameters; Rolling optimized sliding window length

[an error occurred while processing this directive]

参考文献

原文顺序 | 文献年度倒序 | 文中引用次数倒序

[1]	SLATTERY R, ZHAO Y. Trajectory synthesis for air traffic automation[J]. Journal of Guidance,Control and Dynamics, 1997, 20(2):232-238. DOI

[2]	张宗腾, 张琳, 汪文峰. 一种基于双向GRU的UAV飞行轨迹预测方法[J]. 电光与控制, 2022, 29(3):11-15,26. ZHANG Z T, ZHANG L, WANG W F, et al. A UAV flight trajectory prediction method based on bidirectional GRU[J]. Electro Optical and Control, 2022, 29(3):11-15,26.

[3]	W.SCHUSTER. Trajectory prediction for future air traffic management complex manoeuvres and taxiing[J], Aeronaut.Journal, 2015, 199(1212):121-143.

[4]	戴礼灿, 刘欣, 张海瀛. 基于卡尔曼滤波算法展开的飞行目标轨迹预测[J]. 系统工程与电子技术, 2023, 45(6):1814-1820. DOI DAI L C, LIU X, ZHANG H Y, et al. Flight target trajectory prediction based on Kalman filter algorithm[J]. Systems Engineering and Electronic Technology, 2023, 45(06):1814-1820.

[5]	张凯, 熊家军, 李凡. 基于意图推断的高超声速滑翔目标贝叶斯轨迹预测[J]. 宇航学报, 2018, 39(11):1258-1265. ZHANG K, XIONG J J, LI F, et al. Bayesian trajectory prediction of hypersonic gliding targets based on intention inference[J]. Journal of Astronautics, 2018, 39(11):1258-1265.

[6]	SUN J, ELLERBROEK J, HOEKSTRA J M. WRAP:An open-source kinematic aircraft performance model[J]. Transportation Research Part C:Emerging Technologies, 2019,98:118-138.

[7]	XU J H, LIU Z. A centralized muoltisensory multi-target tracking algorithmbased on variable structure multiple-modle[J]. Ship Science and Technology, 2012, 34(1):67-72.

[8]	LYMPEROPOULOS I, LYGEROS J. Sequential Monte Carlo methods for multi-aircraft trajectory prediction in air traffic management[J]. International Journal of Adaptive Control and Signal Processing, 2010, 24(10):830-849. DOI

[9]	MAZOUZ L D, MECHE A, DAHMANI M. An efficient real Time implementationof a fast IMM for tracking a manoeuvring target[C]//2019 International Conference on Advanced Electrical Engineering (ICAEE).Piscataway,NJ,USA:IEEE,2019:1-6.

[10]	孟光磊, 陈振, 罗元强. 基于动态贝叶斯网络的机动动作识别方法[J]. 系统仿真学报, 2017, 29(1):140-145. MENG G L, CHEN Z, LUO Y Q. Mobile action recognition method based on dynamic Bayesian network[J]. Journal of System Simulation, 2017, 29 (1):140-145.

[11]	孟光磊, 张慧敏, 朴海音, 等. 自动化飞行训练评估中的战机机动动作识别[J]. 北京航空航天大学学报, 2020, 46(7):1267-1274. MENG G L, ZHANG H M, PIAO H Y, et al. Aircraft maneuver recognition in automated flight training evaluation[J]. Journal of Beihang University, 2020, 46(7):1267-1274.

[12]	WU Z J, TIAN S, MA L. A 4D trajectory prediction model based on the BP neural network[J]. Journal of Intelligent Systems, 2020, 29(1):1545-1557. DOI

[13]	谢磊, 丁达理, 魏政磊, 等. AdaBoost-PSO-LSTM网络实时预测机动轨迹[J]. 系统工程与电子技术, 2021, 43(6):1651-1658. DOI XIE L, DING D L, WEI Z L, et al. AdaBoost PSO LSTM network for real-time prediction of maneuvering trajectory[J]. Systems Engineering and Electronic Technology, 2021, 43 (6):1651-1658.

[14]	魏喜庆, 顾龙飞, 李瑞康, 等. 基于Singer模型的高超声速飞行器轨迹跟踪与预测[J]. 航天控制, 2017, 35(4):62-66,72. WEI X Q, GU L F, LI R K, et al. Trajectory tracking and prediction of hypersonic aircraft based on Singer model[J]. Aerospace Control, 2017, 35 (4):62-66,72.

[15]	SAHADEVAN D, PONNUSAMY P, GOPI V P, et al. Ground-based 4d trajectory prediction using bi-directional LSTM networks[J]. Applied Intelligence,2022:1-18.

[16]	SIGHENCEA B I, STANCIU R I, C LEANU C D. A review of deep learning-based methods for pedestrian trajectory prediction[J]. Sensors, 2021, 21(22):7543. DOI

[17]	王新, 杨任农, 左家亮. 基于HPSO-TPFENN的目标机轨迹预测[J]. 西北工业大学学报, 2019, 37(3):612-620. WANG X, YANG R N, ZUO J L, et al. Target Machine Trajectory Prediction Based on HPSO-TPFENN[J]. Journal of Northwestern Polytechnical University, 2019, 37 (3):612-620. DOI

[18]	刘迎军. 基于单步和多步模型的钱塘江南源流域水质预测[D]. 武汉大学, 2021. LIU Y J. Water quality prediction of the southern source basin of Qiantang River based on single step and multi-step models[D]. Wuhan University, 2021.

[19]	SRIVASTAVA N, HINTON G, Krizhevsky A, et al. Dropout:a simple way to prevent neural networks from overfitting[J]. The Journal of Machine Learning Research, 2014, 15(1):1929-1958.

[20]	SIAMI-NAMINI S, TAVAKOLI N, NAMIN A S. The performance of LSTM and BiLSTM in forecasting time series[C]//2019 IEEE International Conference on Big Data (Big Data).Los Alamitos,CA,USA:IEEE,2019:3285-3292.

[21]	MAKWE A, RATHORE A S. An empirical study of neural network hyperparameters[M]//Evolution in Computational Intelligence. Singapore:Springer,2021:371-383.

[22]	WU J, CHEN X Y, ZHANG H, et al. Hyperparameter optimization for machine learning models based on Bayesian optimization[J]. Journal of Electronic Science and Technology, 2019, 17(1):26-40.

[23]	VICTORIA A H, MARAGATHAM G. Automatic tuning of hyperparameters using Bayesian optimization[J]. Evolving Systems, 2021, 12(1):217-223. DOI

[24]	QIAO L, WANG Z, ZHU J. Application of improved GRNN model to predict interlamellar spacing and mechanical properties of hypereutectoid steel[J]. Materials Science and Engineering:A, 2020,792:139845.

Options

文章导航