基于脱靶量预测的飞行器反拦截机动方法

doi:10.12305/j.issn.1001-506X.2023.09.32

摘要/Abstract

摘要：

随着反导拦截技术的迅猛发展, 高速滑翔式飞行器末段的生存能力受到日益严峻的考验, 针对此问题提出了一种基于长短期记忆网络(long short-term memory, LSTM)的飞行器反拦截机动方法, 事先规定好飞行器的机动方式, 将飞行器的规避机动问题简化为机动时机选择问题。随后, 以状态时间序列-拦截脱靶量为样本构建训练集, 利用LSTM网络对两者的非线性映射关系进行学习。最后, 利用该网络在飞行中对拦截脱靶量进行实时预测, 借此进行突防时机的选择。从不同状态时序长度, 不同LSTM神经元个数和不同传感器噪声水平三方面对该方法的性能进行了仿真验证和对比评价, 结果表明: 相比于传统的正弦、方波机动, 所提机动方法能使飞行器生存概率显著提高, 同时显著提高落速, 具有一定的工程应用价值。

关键词: 攻防对抗, 机动策略, 脱靶量, 长短期记忆网格

Abstract:

With the rapid development of anti-missile interception technology, the survivability of the end stage of high-speed gliding vehicle is increasingly severely tested. To address this issue, an anti-nterception maneuver method based on long short-term memory (LSTM) network is proposed. The maneuver mode of vehicle is specified in advance, and the evasion maneuver problem of aircraft is simplified to the choice of maneuver time. Then, taking the state time series interception and miss distance as the sample, the training set is constructed, and the LSTM network is used to learn the nonlinear mapping relationship between them. Finally, the network is used to predict the miss distance in flight, so as to select the penetration time. Simulation and comparison of the performance of the method are carried out from three aspects: simultaneous interpreting of different state time series, different LSTM neurons and different sensor noise levels. The results show that compared with traditional sinusoidal and square wave maneuvers, the proposed maneuvering method can significantly improve the survival probability of aircraft and significantly increase the speed of falling. It has certain engineering application value.

Key words: attack defense confrontation, maneuver strategy, miss distance, long short-term memory (LSTM) network

中图分类号:

陈劭博, 严佳民, 卜奎晨. 基于脱靶量预测的飞行器反拦截机动方法[J]. 系统工程与电子技术, 2023, 45(9): 2922-2930.

Shaobo CHEN, Jiamin YAN, Kuichen BU. Anti-intercept maneuver method of vehicle based on prediction of miss distance[J]. Systems Engineering and Electronics, 2023, 45(9): 2922-2930.

图/表 17

图1

图2

图3

图4

图5

图6

表1

表2

图7

表3

图8

表4

图9

表5

表6

表7

图10

参考文献 24

1	ZARCHAN P . Proportional navigation and weaving targets[J]. Journal of Guidance, Control, and Dynamics, 1995, 18 (5): 969- 974. doi: 10.2514/3.21492
2	雍恩米, 唐国金, 罗亚中. 弹道导弹终端机动突防制导问题仿真研究[J]. 导弹与航天运载技术, 2005, (4): 13- 18. doi: 10.3969/j.issn.1004-7182.2005.04.003
	YONG E M , TANG G J , LUO Y Z . Rearch on maneuvering penetration guidance law of ballistic missile in middle-course flight by simulation method[J]. Missile and Space Vehicles, 2005, (4): 13- 18. doi: 10.3969/j.issn.1004-7182.2005.04.003
3	SHINAR J , STEINBERG D . Analysis of optimal evasive maneuvers based on a linearized two-dimensional kinematic model[J]. Journal of Aircraft, 1977, 14 (8): 795- 802. doi: 10.2514/3.58855
4	王亚帆, 周韬, 陈万春, 等. 基于脱靶量级数解的最优机动突防策略[J]. 北京航空航天大学学报, 2020, 46 (1): 159- 169.
	WANG Y F , ZHOU T , CHEN W C , et al. Optimal maneuver penetration strategy based on power series solution of miss distance[J]. Journal of Beijing University of Aeronautics and Astornautics, 2020, 46 (1): 159- 169.
5	IMADO F . Some aspects of a realistic three-dimensional pursuit-evasion game[J]. Journal of Guidence, Control, and Dynamics, 1993, 16 (2): 125- 139.
6	IMADO F , UEHARA S . High-g barrel roll maneuvers against proportional navigation from optimal control, viewpoint[J]. Journal of Guidance, Control, and Dynamics, 2016, 29 (6): 876- 991.
7	李静琳, 陈万春, 闵昌万. 高超末段机动突防/精确打击弹道建模与优化[J]. 北京航空航天大学学报, 2018, 44 (3): 556- 567.
	LI J L , CHEN W C , MIN C W . Terminal hypersonic trajectory modeling and optimization for maneuvering penetration and precision strike[J]. Journal of Beijing University of Aeronautics and Astornautics, 2018, 44 (3): 556- 567.
8	ZHANG P , FANG Y W , ZHANG F M , et al. Differential game guidance law against high maneuver target with different information patterns[J]. Advanced Science Latters, 2012, (12): 143- 147.
9	程林, 蒋方华, 李俊峰. 深度学习在飞行器动力学与控制中的应用研究综述[J]. 力学与实践, 2020, 42 (3): 267- 276.
	CHENG L , JIANG F H , LI J F . A review on the applications of deep learning in aircraft dynamics and control[J]. Mechanics in Engineering, 2020, 42 (3): 267- 276.
10	SONG E J , TAHK M J . Real-time midcourse guidance with intercept point prediction[J]. Control Engineering Practice, 1998, 6 (8): 957- 967. doi: 10.1016/S0967-0661(98)00041-0
11	SONG E J , TAHK M J . Real-time neural network midcourse guidance[J]. Control Engineering Practice, 2001, 9 (10): 1145- 1154. doi: 10.1016/S0967-0661(01)00058-2
12	汪韧, 惠俊鹏, 俞启东, 等. 基于LSTM模型的飞行器智能制导技术研究[J]. 力学学报, 2021, 53 (7): 2047- 2057.
	WANGR, HUIJ P, YUQ D., 等. Research of LSTM model-based intelligent guidance of flight aircraft[J]. Chinese Journal of Theoretical and Applied Mechanics, 2021, 53 (7): 2047- 2057.
13	曾庆华, 董荣华, 皮术武. 基于最优制导模板的神经网络预测制导方法[J]. 国防科技大学学报, 2014, 36 (1): 137- 141.
	ZENG Q H , DONG R H , PI S W . Neural network predictive guidance method based on pattern of optimal guidance[J]. Journal of National University of Defence Technology, 2014, 36 (1): 137- 141.
14	SHI Y, WANG Z B. A deep learning-based approach to real-time trajectory optimization for hypersonic vehicle[C]//Proc. of the AIAA Scitech Forum, 2020.
15	CHENG L , JIANG F H , WANG Z B , et al. Multi-constrained real-time entry guidance using deep neural network[J]. IEEE Trans.on Aerospace and Electronic Systems, 2021, 57 (1): 325- 340.
16	GAUDET B , FURFARO R , LINARES R . Reinforcement learning for angle-only intercept guidance of maneuvering targets[J]. Aeropace Science and Technology, 2020, 99, 105746.
17	HOVELL K , UIRICH S . Deep reinforcement learning for spacecraft proximity operations guidance[J]. Journal of Spacecraft and Rockets, 2021, 58 (2): 254- 264.
18	HOVELL K, UIRICH S. On deep reinforcement learning for spacecraft guidance[C]//Proc. of the AIAA Scitech Forum, 2020.
19	GAO J S, SHI X M, CHENG Z T, et al. Reentry trajectory optimization based on deep reinforcement learning[C]//Proc. of the 31th Chinese Control and Decision Conference, 2019: 2588-2952.
20	HOCHREITER S , SCHMIDHUBER J . Long short-term memory[J]. Neural Computation, 1997, 9 (8): 1735- 80.
21	ROUDBARI A , SAGHAFI F . Intelligent modeling and identification of aircraft nonlinear flight[J]. Chinese Journal of Aeronautics, 2014, 27 (4): 759- 771.
22	CHEN J F , PI D C , WU Z Y , et al. Imbalanced satellite telemetry data anomaly detection model based on Bayesian LSTM[J]. Acta Astronautica, 2021, 180, 232- 242.
23	朱俊鹏, 赵洪利, 杜鑫, 等. 长短时记忆神经网络在卫星轨道预报中的研究[J]. 兵器装备工程学报, 2017, (10): 127- 132.
	ZHU J P , ZHAO H L , DU X , et al. Application of long short-term memory neural network to orbit prediction of satellite[J]. Journal of Ordnance Equipment Engineering, 2017, (10): 127- 132.
24	KINGMA D P, BA J. Adam: a method for stochastic optimization[C]//Proc. of the International Conference on Learning Representations, 2014.

初始条件参数	取值范围
x₀/m	[-80 000, -70 000]
y₀/m	[20 000, 22 000]
z₀/m	[-2 000, 2 000]
x_D0/m	[-10 000, 10 000]
y_D0/m	[0, 0]
z_D0/m	[-10 000, 10 000]
v₀/(m·s^－1)	[1 650, 1 750]
v_D0/(m·s^－1)	[450,500]
N	[3,6]
θ₀/rad	[-0.14, 0.1]
φ₀/rad	[-0.05, 0.05]
R_tf/m	[100, 12 000]

机动方式	${\bar R}$_min/m	${\bar R}$_final/m	P_surv/%	V_f/(m/s)
脱靶量预测机动	14.597 6	0.859 6	98.7	678.35
不机动	0.562 5	0.830 8	0	694.26
正弦机动	3.562 5	0.820 6	6.7	590.51
方波机动	6.133 2	0.844 9	13.6	509.01

输入序列长度	${\bar R}$_min/m	S/m	P_surv/%
u=5	14.899 8	2.753 2	98.2
u=10	14.895 6	1.480 6	98.6
u=20	14.959 8	1.004 0	100
u=40	14.848 9	0.989 4	100

神经元数量	${\bar R}$_min/m	S/m	P_surv/%
n=21	14.799 0	1.486 5	99.8
n=28	15.045 6	0.981 2	100
n=35	15.818 4	1.002 7	100
n=42	14.967 5	0.946 4	100

噪声水平	q₁^A, q₂^A/(°)	$\dot{q}_1^A, \dot{q}_2^A /\left({ }^{\circ} / \mathrm{s}\right)$	R/m	h/m	v/(m/s)
无	0	0	0	0	0
1	0.1	0.01	25	5	1
2	0.3	0.03	75	15	3