基于SVM和EKF的高超声速滑翔飞行器轨迹预报

程云鹏; 孙成志; 闫晓东

doi:10.13700/j.bh.1001-5965.2020.0050

基于SVM和EKF的高超声速滑翔飞行器轨迹预报

doi: 10.13700/j.bh.1001-5965.2020.0050

程云鹏¹,
孙成志¹,
闫晓东^{1, 2, ,}

1.
西北工业大学航天学院, 西安 710072
2.
陕西省空天飞行器设计技术重点实验室, 西安 710072

详细信息

作者简介:
程云鹏  男, 博士研究生。主要研究方向:机动目标跟踪和预报

孙成志  男, 硕士研究生。主要研究方向:航天器故障诊断和识别

闫晓东  男, 博士, 副教授, 硕士生导师。主要研究方向:飞行器动力学与制导

通讯作者:
闫晓东, E-mail: yan804@nwpu.edu.cn

中图分类号: V19;TN953
计量
- 文章访问数: 642
- HTML全文浏览量: 120
- PDF下载量: 214
- 被引次数: 0
出版历程
- 收稿日期: 2020-02-25
- 录用日期: 2020-05-30
- 网络出版日期: 2020-11-20

Trajectory prediction of hypersonic glide vehicle based on SVM and EKF

1.
School of Astronautics, Northwestern Polytechnical University, Xi'an 710072, China
2.
Shaanxi Aerospace Flight Vehicle Design Key Laboratory, Xi'an 710072, China

More Information

Corresponding author: YAN Xiaodong, E-mail: yan804@nwpu.edu.cn

摘要

摘要:
高超声速滑翔飞行器（HGV）拦截问题中，轨迹预报是成功拦截的重要基础。针对HGV机动能力强、轨迹多变的特点，提出了一种基于支持向量机（SVM）和扩展卡尔曼滤波（EKF）的轨迹预报方法。在HGV的滑翔段机动模式分析的基础上，将HGV的机动运动分解为纵向运动模式和侧向运动模式，进而对运动模式的特征参数予以标定，形成SVM的训练集。建立地基单雷达轨迹跟踪模型，采用EKF对HGV滑翔段轨迹进行稳定跟踪并实现对运动模式特征参数的估计。基于SVM，建立了HGV运动识别框架，实现了对HGV滑翔段轨迹的预报。对平衡滑翔和跳跃机动2种典型机动模式进行数学仿真验证，结果表明，所提方法可以提高对该类目标的轨迹预报精度。
- 高超声速滑翔飞行器(HGV) /
- 机动模式 /
- 支持向量机(SVM) /
- 运动识别 /
- 轨迹估计和预报
Abstract:
In the scenario of intercepting a Hypersonic Glide Vehicle (HGV), the trajectory prediction is a key issue for successful interception. Considering HGV's strong maneuverability and variable trajectory, in this paper, a novel trajectory prediction method is proposed based on Support Vector Machine (SVM) and Extended Kalman Filter (EKF). First, the investigation on the maneuvering mode is performed. The maneuver motion of the HGV is divided into longitudinal mode and lateral mode, which are labeled and formulated into the training set of SVMs. Second, the tracking model of the trajectory for single ground-based radar is established, and EKF is applied to track the glide trajectory of HGV. Finally, the recognition framework of HGV motion is established based on SVM, and the prediction of the subsequent trajectory is accomplished. The results show that the proposed method can improve the trajectory prediction accuracy of HGV.
- Hypersonic Glide Vehicle (HGV) /
- maneuver mode /
- Support Vector Machine (SVM) /
- motion recognition /
- trajectory estimation and prediction

HTML全文

图 1 支持向量机训练流程图

Figure 1. Flowchart of training by support vector machine

下载: 全尺寸图片幻灯片

图 2 高超声速滑翔飞行器运动识别轨迹预报流程图

Figure 2. Flowchart of motion recognition based trajectory prediction for HGVs

下载: 全尺寸图片幻灯片

图 3 准平衡侧摆机动轨迹跟踪均方根误差

Figure 3. Trajectory tracking root-mean-square error for quasi-equilibrium pendulum maneuver

下载: 全尺寸图片幻灯片

图 4 AUC结果对比(算例1)

Figure 4. AUC results comparison (Case 1)

下载: 全尺寸图片幻灯片

图 5 准平衡侧摆机动轨迹预报结果

Figure 5. Trajectory prediction results for quasi-equilibrium pendulum maneuver

下载: 全尺寸图片幻灯片

图 6 准平衡侧摆机动轨迹100 s预报结果空间误差散布

Figure 6. Spatial error distribution of 100 s trajectory prediction results for quasi-equilibrium pendulum maneuver

下载: 全尺寸图片幻灯片

图 7 定迎角转弯机动轨迹跟踪均方根误差

Figure 7. Trajectory tracking root-mean-square error for fixed angle of attack turning maneuver

下载: 全尺寸图片幻灯片

图 8 AUC结果对比(算例2)

Figure 8. AUC results comparison (Case2)

下载: 全尺寸图片幻灯片

图 9 定迎角转弯机动轨迹预报结果

Figure 9. Trajectory prediction results for fixed angle of attack turning maneuver

下载: 全尺寸图片幻灯片

图 10 定迎角转弯机动轨迹100 s预报结果空间误差散布

Figure 10. Spatial error distribution of 100 s trajectory prediction results for fixed angle of attackturning maneuver

下载: 全尺寸图片幻灯片

表 1 高超声速滑翔飞行器纵向机动标签

Table 1. Longitudinal maneuvering label of HGV

迎角/(°)	Label_y
α_equi	0
[0, 3)	1
[3, 5)	2
[5, 7)	3
[7, 9)	4
[9, 11)	5
[11, 13)	6
[13, 15)	7
[15, 17)	8
[17, 20]	9

下载: 导出CSV

表 2 高超声速滑翔飞行器侧向机动标签(1)

Table 2. Lateral maneuvering label of HGV (1)

侧摆幅值/m	Label_z1
无机动	0
[750, 1 250)	1
[1 250, 1 750)	2
[1 750, 2 250)	3
[2 250, 2 750]	4

下载: 导出CSV

表 3 高超声速滑翔飞行器侧向机动标签(2)

Table 3. Lateral maneuvering label of HGV (2)

倾侧角/rad	Label_z2
[π/9, π/7)	5
[π/7, π/5)	6
[π/5, π/3]	7

下载: 导出CSV

表 4 准平衡侧摆机动位置预报误差统计平均值

Table 4. Statistical mean value of position prediction error for quasi-equilibrium pendulum maneuver

轨迹预报方法	位置误差统计平均值/km
轨迹预报方法	预报时长为30 s	预报时长为50 s	预报时长为100 s
典型控制规律	4.89	9.51	30.78
SVM-EKF	4.03	6.66	17.15

下载: 导出CSV

表 5 定迎角转弯机动位置预报误差统计平均值

Table 5. Statistical mean value of position prediction error for fixed angle of attack turning maneuver

轨迹预报方法	位置误差统计平均值/km
轨迹预报方法	预报时长为30 s	预报时长为50 s	预报时长为100 s
典型控制规律	4.07	4.34	4.64
SVM-EKF	3.52	3.60	3.71

下载: 导出CSV

参考文献(31)

[1]	张灿, 林旭斌, 刘都群, 等.2019年国外高超声速飞行器技术发展综述[J].飞航导弹, 2020(1):16-20. ZHANG C, LIN X B, LIU D Q, et al.Review of the development of hypersonic vehicle technology in 2019[J].Aerodynamic Missile Journal, 2020(1):16-20(in Chinese).
[2]	林旭斌, 李彦, 胡冬冬.美军2019财年高超声速科研预算简析[J].飞航导弹, 2018(6):1-3. http://www.cnki.com.cn/Article/CJFDTotal-FHDD201806004.htm LIN X B, LI Y, HU D D.Analysis on the budget of hypersonic scientific research of the US military in 2019[J].Aerodynamic Missile Journal, 2018(6):1-3(in Chinese). http://www.cnki.com.cn/Article/CJFDTotal-FHDD201806004.htm
[3]	LI G H, ZHANG H B, TANG G J.Maneuver characteristics analysis for hypersonic glide vehicles[J].Aerospace Science and Technology, 2015, 43:321-328. doi: 10.1016/j.ast.2015.03.016
[4]	ZARCHAN P.Tactical and strategic missile guidance[M].Reston:AIAA, 2012.
[5]	QIN L, LI J L, ZHOU D.Tracking filter and prediction for non-ballistic target HTV-2 in near space[C]//27th Chinese Control and Decision Conference.Piscataway: IEEE Press, 2015: 3556-3561.
[6]	KIM K, HWANG I.Intent-based detection and characterization of aircraft maneuvers in en route airspace[J].Journal of Aerospace Information Systems, 2018, 15(2):72-91. doi: 10.2514/1.I010566
[7]	张洪波, 黄景帅, 李广华, 等.典型控制规律滑翔飞行器的轨迹预测方法[J].现代防御技术, 2017, 45(4):112-118. http://www.cnki.com.cn/Article/CJFDTOTAL-XDFJ201704019.htm ZHANG H B, HUANG J S, LI G H, et al.Trajectory prediction of glide vehicle based on typical control law[J].Modern Defence Technology, 2017, 45(4):112-118(in Chinese). http://www.cnki.com.cn/Article/CJFDTOTAL-XDFJ201704019.htm
[8]	VAPNIK N.The nature of statistical learning theory[M].Berlin:Springer, 1995.
[9]	BERNHARD S.Advances in kernel methods:Support vector learning[M].Cambridge:MIT Press, 1999.
[10]	于震梁, 孙志礼, 曹汝男, 等.基于支持向量机和卡尔曼滤波的机械零件剩余寿命预测模型研究[J].兵工学报, 2018, 39(5):991-997. http://www.cnki.com.cn/Article/CJFDTotal-BIGO201805020.htm YU Z L, SUN Z L, CAO R N, et al.Research on remaining useful life predictive model of machine parts based on SVM and Kalman filter[J].Acta Armamentarii, 2018, 39(5):991-997(in Chinese). http://www.cnki.com.cn/Article/CJFDTotal-BIGO201805020.htm
[11]	陈旭梅, 龚辉波, 王景楠.基于SVM和Kalman滤波的BRT行程时间预测模型研究[J].交通运输系统工程与信息, 2012, 12(4):29-34. http://d.wanfangdata.com.cn/Periodical/jtysxtgcyxx201204005 CHEN X M, GONG H B, WANG J N.BRT vehicle travel time prediction based on SVM and Kalman filter[J].Journal of Transportation Systems Engineering and Information Technology, 2012, 12(4):29-34(in Chinese). http://d.wanfangdata.com.cn/Periodical/jtysxtgcyxx201204005
[12]	杨茂, 黄宾阳, 江博, 等.基于卡尔曼滤波和支持向量机的风电功率实时预测研究[J].东北电力大学学报, 2017, 37(2):45-51. http://www.cnki.com.cn/Article/CJFDTotal-DBDL201702008.htm YANG M, HUANG B Y, JIANG B, et al.Real-time prediction for wind power based on Kalman filter and support vector machines[J].Journal of Northeast Electric Power University, 2017, 37(2):45-51(in Chinese). http://www.cnki.com.cn/Article/CJFDTotal-DBDL201702008.htm
[13]	PENG H, BAI X L.Exploring capability of support vector machine for improving satellite orbit prediction accuracy[J].Journal of Aerospace Information Systems, 2018, 15(6):366-381. doi: 10.2514/1.I010616
[14]	VINH N X.Optimal trajectories in atmospheric flight[M].London:Pergamon Press Ltd., 1982:449-468.
[15]	赵汉元.飞行器再入动力学与制导[M].长沙:国防科技大学出版社, 1997. ZHAO H Y.Reentry dynamics and guidance of aircraft[M].Changsha:University of National Defense Science and Technology Press, 1997(in Chinese).
[16]	LI G H, ZHANG H B, TANG G J, et al.Maneuver modes analysis for hypersonic glide vehicles[C]//Proceedings of 2014 IEEE Chinese Guidance, Navigation and Control Conference.Piscataway: IEEE Press, 2014: 543-548.
[17]	ZHU J W, LIU L H, TANG G J, et al.Highly constrained optimal gliding guidance[J].Proceedings of the Institution of Mechanical Engineers, Part G:Journal of Aerospace Engineering, 2015, 229(12):2321-2335. doi: 10.1177/0954410015573973
[18]	李凡, 熊家军, 张凯, 等.临近空间高超声速目标跟踪动力学模型[J].宇航学报, 2019, 40(3):266-276. http://www.cnki.com.cn/Article/CJFDTotal-YHXB201903004.htm LI F, XIONG J J, ZHANG K, et al.Near space hypersonic target dynamics tracking model[J].Journal of Astronautics, 2019, 40(3):266-276(in Chinese). http://www.cnki.com.cn/Article/CJFDTotal-YHXB201903004.htm
[19]	谢愈, 刘鲁华, 汤国建, 等.高超声速滑翔飞行器摆动式机动突防弹道设计[J].航空学报, 2011, 32(12):2174-2181. http://www.cqvip.com/QK/91925X/201112/40162349.html XIE Y, LIU L H, TANG G J, et al.Weaving maneuver trajectory design for hypersonic glide vehicle[J].Acta Aeronautica et Astronautica Sinica, 2011, 32(12):2174-2181(in Chinese). http://www.cqvip.com/QK/91925X/201112/40162349.html
[20]	王丹.基于运动行为识别的高超声速目标弹道估计[D].哈尔滨: 哈尔滨工业大学, 2017. http://cdmd.cnki.com.cn/Article/CDMD-10213-1017863950.htm WANG D.Trajectory estimation of hypersonic target based on motion behavior recognition[D].Harbin: Harbin Institute of Technology, 2017(in Chinese). http://cdmd.cnki.com.cn/Article/CDMD-10213-1017863950.htm
[21]	HOUGH M E.Reentry maneuver estimation using nonlinear Markov acceleration models[J].Journal of Guidance, Control, and Dynamics, 2017, 40(7):1693-1710. doi: 10.2514/1.G002577
[22]	张凯, 熊家军, 韩春耀, 等.一种基于气动力模型的高超声速滑翔目标跟踪算法[J].宇航学报, 2017, 38(2):123-130. http://www.cnki.com.cn/Article/CJFDTotal-YHXB201702003.htm ZHANG K, XIONG J J, HAN C Y, et al.A tracking algorithm of hypersonic glide reentry vehicle via aerodynamic model[J].Journal of Astronautics, 2017, 38(2):123-130(in Chinese). http://www.cnki.com.cn/Article/CJFDTotal-YHXB201702003.htm
[23]	SINGER R A.Estimating optimal tracking filter performance for manned maneuvering targets[J].IEEE Transactions on Aerospace and Electronic Systems, 1970, AES-6(4):473-483. doi: 10.1109/TAES.1970.310128
[24]	ZHU J W, HE R Z, TANG G J, et al.Pendulum maneuvering strategy for hypersonic glide vehicles[J].Aerospace Science and Technology, 2018, 78:62-70. doi: 10.1016/j.ast.2018.03.038
[25]	HOU Q L, ZHANG J X, LIU L M, et al.Discriminative information-based nonparallel support vector machine[J].Signal Processing, 2019, 162:169-179. doi: 10.1016/j.sigpro.2019.03.026
[26]	CHANG M, LIN C.Leave-one-out bounds for support vector regression model selection[J].Neural Computation, 2005, 17(5):1188-1222. doi: 10.1162/0899766053491869
[27]	CHANG C, LIN C.LIBSVM:A library for support vector machines[J].ACM Transactions on Intelligent Systems and Technology, 2011, 2(3):1-27.
[28]	KJELDSEN H T.A contextualized historical analysis of the Kuhn-Tucker theorem in nonlinear programming:The impact of World War Ⅱ[J].Historia Mathematica, 2000, 27(4):331-361. doi: 10.1006/hmat.2000.2289
[29]	CHEN K L, YU J.Short-term wind speed prediction using an unscented Kalman filter based state-space support vector regression approach[J].Applied Energy, 2014, 113:690-705. doi: 10.1016/j.apenergy.2013.08.025
[30]	张强, 肖刚, 蓝屹群.基于BP神经网络多类分类的湍流目标探测[J].系统工程与电子技术, 2018, 40(7):1486-1490. http://www.cnki.com.cn/Article/CJFDTotal-XTYD201807011.htm ZHANG Q, XIAO G, LAN Y Q.Turbulence target detection based on BP neural network multi-level classification[J].Systems Engineering and Elctronics, 2018, 40(7):1486-1490(in Chinese). http://www.cnki.com.cn/Article/CJFDTotal-XTYD201807011.htm
[31]	何伟.基于朴素贝叶斯的文本分类算法研究[D].南京: 南京邮电大学, 2018. http://cdmd.cnki.com.cn/Article/CDMD-10293-1019690701.htm HE W.Text classification algorithm research based on naive Bayes[D].Nanjing: Nanjing University of Posts and Telecommunications, 2018(in Chinese). http://cdmd.cnki.com.cn/Article/CDMD-10293-1019690701.htm