无通信条件下基于视觉毁伤评估的弹群对地目标自主攻击决策

doi:10.12382/bgxb.2023.1078

摘要/Abstract

摘要：

面向利用巡飞弹集群对地面高防御移动目标进行作战的任务场景,提出基于高度划分巡飞弹不同作战功能的攻击决策流程方法。通过在任一巡飞弹部署YOLO X与VGG-16耦合的察打评估一体在线目标识别与毁伤评估网络(YOLO-VGGNet),能够使弹群仅基于视觉感知信息在无通信条件下进行自主作战。同时提出一种末端再评估与目标攻击点选取方法,使得巡飞弹既能在飞临目标的攻击末端对地面目标进行再评估,又能对目标未受打击的特征部位进行定位,从而弹群节点均能自主决定是否继续进行攻击并优先攻击目标的特征部位。仿真实验结果表明:在无通信条件下,通过应用所提的基于YOLO-VGGNet的在线毁伤评估网络,巡飞弹集群既能对任务区域中的运动目标进行完全毁伤,又能减少不必要的弹药损耗并优先攻击目标特征部位,研究有望提升弹群整体作战效能。

关键词: 大规模分布式弹群作战, 察打评估一体, 无通信条件, 末端攻击决策, 视觉毁伤评估

Abstract:

For the mission scenario of using a cluster of patrol missiles to fight against the ground high-defense mobile targets, an attack decision-making process method based on the height division of patrol missiles with different operational functions is proposed. By deploying the YOLO X and VGG-16 integrated online target recognition and damage assessment network (YOLO-VGGNet) on any patrol missile, the patrol missile can autonomously strike a target based on visual sensing information only under no communication conditions. At the same time, a method for terminal reassessment and target attack point selection is proposed so that the patrol missile can not only reassess the ground target at the terminal of attack when flying to the target, but also locate the feature parts that have not been struck. The nodes of missile swarm can autonomously decide whether to continuely attack and prioritize the attack on the feature parts of target. Simulated results show that under no communication conditions, the proposed YOLO-VGGNet online damage assessment network can not only destroy the targets in the mission area completely, but also reduce the unnecessary ammunition loss and prioritize the attack on the functional parts of the target, which is expected to improve the overall combat effectiveness of missile swarm.

Key words: massively distributed missile swarm operation, reconnaissance-strike-assessment, no communication condition, terminal attack decision-making, visual damage assessment

中图分类号:

TP249

徐艺博, 颜佳润, 曾志文, 吕云霄, 冯世如, 卢惠民. 无通信条件下基于视觉毁伤评估的弹群对地目标自主攻击决策[J]. 兵工学报, 2024, 45(12): 4435-4450.

XU Yibo, YAN Jiarun, ZENG Zhiwen, LÜ Yunxiao, FENG Shiru, LU Huimin. Autonomous Attack Decision-making of Missile Swarm on Ground Targets Based on Visual Damage Assessment without Communication[J]. Acta Armamentarii, 2024, 45(12): 4435-4450.

图/表 22

图1 弹群节点对地目标识别与毁伤评估流程

Fig.1 Schematic diagram of ground target recognition and damage assessment of swarm nodes

图2 YOLO-VGGNet毁伤等级评估

Fig.2 YOLO-VGGNet damage level assessment

图3 补给车不同毁伤状态

Fig.3 Different damage status of supply vehicle

表1 毁伤评估等级指标定义

Table 1 Definition of damage assessment level indexes

毁伤等级	雷达	军帐	坦克	补给车
未毁伤目标	车头和天线完好,具有移动能力	撑杆骨架结构完整,篷布纹理鲜明,没有火花	炮塔完好具有攻击能力,履带完好具有移动能力	车头、车箱完好,具有移动能力
中度毁伤目标	车头和天线受损/伴有浓烟/移动能力丧失	篷布纹理变暗/撑杆骨架结构塌陷/篷布分布有火花	履带受损,移动能力丧失/浓烟、火花分布在坦克周身	车头或车箱被摧毁/移动能力丧失/车身伴有浓烟
完全毁伤目标	车头和天线被完全摧毁,完全丧失移动能力	撑杆骨架结构塌陷并且篷布燃烧殆尽	炮塔完全脱离且移动能力丧失	车头和车箱被完全摧毁,移动能力完全丧失

表2 巡飞弹不同高度对地目标可见光成像与打击部位评估

Table 2 Visual imaging of patrol missile approaching a ground target and assessment of strike part at different altitudes

图4 巡飞弹对地目标识别评估与攻击决策框图

Fig.4 Block diagram of ground target recognition, assessment and attack of patrol missile

表3 不同类型目标卷积层Score-CAM可视化

Table 3 Score-CAM visualization for different type of target convolutional layers

表4 巡飞弹末端目标识别与攻击部位评估

Table 4 Terminal target recognition and strike part assessment by patrol missile

图5 标记末端攻击点

Fig.5 Mark terminal point of attack

表5 巡飞弹目标攻击点不合理样本

Table 5 Unreasonable samples of target attack point for patrol missiles

表6 摧毁目标所需弹的个数

Table 6 Number of missiles required to destroy a target

目标类型	坦克	补给车	雷达车	军帐
弹数	4	3	3	2

图6 基于高度划分巡飞弹不同功能

Fig.6 Different functions of patrol missiles divided based on height

图7 无通信条件下弹群任务中心点的转移

Fig.7 Shift of the mission center point of missile swarm without communication

表7 弹群有效攻击统计结果(未加末端决策)

Table 7 Statistical results of effective attacks by missile swarm (without terminal decision-making)

攻击轮数和总计	雷达		补给车		坦克		军帐		共计
攻击轮数和总计	最大攻击数	实际攻击数	最大攻击数	实际攻击数	最大攻击数	实际攻击数	最大攻击数	实际攻击数	有效攻击数	总攻击数	未攻击数
1	28	36	22	36	13	48	14	24	77	120	0
2	27	36	25	36	17	48	11	24	80	110	10
3	22	36	30	36	11	48	12	24	75	110	10
4	29	36	27	36	24	48	11	24	91	113	7
5	22	36	16	36	34	48	21	24	93	114	6
6	21	36	5	36	38	48	18	24	82	111	9
7	23	36	5	36	32	48	18	24	78	117	3
8	23	36	6	36	40	48	20	24	89	116	4
9	20	36	4	36	34	48	22	24	80	113	7
10	18	36	8	36	42	48	22	24	90	115	5
总计	233	360	148	360	285	480	169	240	835	1139	61

表8 弹群有效攻击统计结果(加入末端决策)

Table 8 Statistical results of effective attacks by missile swarm (with terminal decision-making)

攻击轮数和总计	雷达		补给车		坦克		军帐		共计
攻击轮数和总计	最大攻击数	实际攻击数	最大攻击数	实际攻击数	最大攻击数	实际攻击数	最大攻击数	实际攻击数	有效攻击数	总攻击数	未攻击数
1	31	36	34	36	21	48	17	24	103	118	2
2	33	36	34	36	20	48	18	24	105	120	0
3	24	36	28	36	35	48	16	24	103	114	6
4	25	36	25	36	36	48	17	24	103	115	5
5	27	36	15	36	43	48	17	24	102	111	9
6	23	36	18	36	39	48	20	24	100	116	4
7	17	36	19	36	45	48	21	24	102	108	12
8	27	36	10	36	43	48	21	24	101	109	11
9	19	36	9	36	47	48	22	24	97	108	12
10	24	36	14	36	45	48	18	24	101	103	17
总计	250	360	206	360	374	480	187	240	1017	1122	78

表9 末端攻击标记点合理率统计结果(添加偏离因子之前)

Table 9 Statistical results of terminal attack marker point reasonable rate (before adding the deviation factor)

雷达			补给车			坦克			军帐			总计
合理数量	总数	合理率/%	合理数量	总数	合理率/%	合理数量	总数	合理率/%	合理数量	总数	合理率/%	合理总数	测试集总数	合理率/%
297	300	99.00	261	300	87.00	236	300	78.67	291	300	97.00	1085	1200	90.42

表10 末端攻击标记点合理率统计结果(添加偏离因子之后)

Table 10 Statistical results of terminal attack marker point reasonable rate (after adding the deviation factor)

雷达			补给车			坦克			军帐			总计
合理数量	总数	合理率/%	合理数量	总数	合理率/%	合理数量	总数	合理率/%	合理数量	总数	合理率/%	合理总数	测试集总数	合理率/%
299	300	99.67	272	300	90.67	300	300	100.00	299	300	99.67	1170	1200	97.50

表11 标记末端攻击点(添加偏离因子之后)

Table 11 Mark terminal point of attack (after adding the deviation factor)

图8 成像角度对攻击点的影响

Fig.8 Impact of imaging angle on point of attack

图9 真实图像部分测试样本

Fig.9 Partial test samples of real images

图10 真实图像特征部位评估样本

Fig.10 Assessment samples of feature parts of real images

表12 真实图像卷积层特征权重可视化

Table 12 Visualization of convolutional layer feature weights for real images

参考文献 38

[1]	程进. 弹群协同与自主决策[M]. 北京: 科学出版社, 2020.
	CHENG J. Missile-group coordination and autonomous decision-making[J]. Beijing: Science Press, 2020. (in Chinese)
[2]	张婷婷, 蓝羽石, 宋爱国. 无人集群系统自主协同技术综述[J]. 指挥与控制学报, 2021, 7(2): 127-136.
	ZHANG T T, LAN Y S, SONG A G. An overview of autonomous collaboration technologies for unmanned swarm systems[J]. Journal of Command and Control, 2021, 7(2): 127-136. (in Chinese)
[3]	陈立栋, 王原, 邸建勋, 等. 无人集群协同控制策略及军事应用[J]. 指挥与控制学报, 2023, 9(4): 380-392.
	CHEN L D, WANG Y, DI J X, et al. Unmanned cluster cooperative control strategy and military application[J]. Journal of Command and Control, 2023, 9(4): 380-392. (in Chinese)
[4]	孔国杰, 冯时, 于会龙, 等. 无人集群系统协同运动规划技术综述[J]. 兵工学报, 2023, 44(1): 11-26. doi: 10.12382/bgxb.2022.0930
	KONG G J, FENG S, YU H L, et al. A review on cooperative motion planning of unmanned vehicles[J]. Acta Armamentarii, 2023, 44(1): 11-26. (in Chinese) doi: 10.12382/bgxb.2022.0930
[5]	HOU J Y, LIU C W. Development situation and key technology analysis of cooperative combat for unmanned aerial vehicle swarm[J]. Modern Radar, 2020, 42(6): 30-40.
[6]	马征宇, 白阳. 机器人集群协同作战关键技术研究[J]. 中国电子科学研究院学报, 2022, 17(1): 98-104.
	MA Z Y, BAI Y. Research of robotic swarm’s core technology in cooperative operations[J]. Journal of China Academy of Electronics and Information Technology, 2022, 17(1): 98-104. (in Chinese)
[7]	WANG X K, SHEN L C, LIU Z H, et al. Coordinated flight control of miniature fixed-wing UAV swarms:methods and experiments[J]. Science China(Information Sciences), 2019, 62(11): 134-150.
[8]	周易. 蜂群+巡飞弹——首次亮相的蜂群1号无人作战车辆系统[J]. 兵工科技, 2021, 19: 15-18.
	ZHOU Y. Bee swarm + patrol missile: the debut of the swarm 1 unmanned combat vehicle system[J]. Ordnance Industry Science Technology, 2021, 19: 15-18. (in Chinese)
[9]	杜永浩, 邢立宁, 蔡昭权. 无人飞行器集群智能调度技术综述[J]. 自动化学报, 2020, 46(2): 222-241.
	DU Y H, XING L N, CAI Z Q. Survey on intelligent scheduling technologies for unmanned flying craft clusters[J]. Acta Autonatica Sinica, 2020, 46(2): 222-241. (in Chinese)
[10]	杨帆. 多巡飞弹协同作战关键技术研究[D]. 北京: 北京理工大学, 2018.
	YANG F. Research on key technologies for coordinated operation of multiple patrol missiles[D]. Beijing: Beijing Institute of Technology, 2018. (in Chinese)
[11]	郝峰, 王鹏飞, 张栋. 多巡飞弹侦察/打击/评估一体协同方案设计[J]. 火力与指挥控制, 2019, 44(12): 1-5.
	HAO F, WANG P F, ZHANG D. The collaborative strategy design for the integrated reconnaissance/strike/evaluation of multiple loitering munitions[J]. Fire Control & Command Control, 2019, 44(12): 1-5. (in Chinese)
[12]	SHAHID S, ZHEN Z Y, JAVAID U, et al. Offense-defense distributed decision making for swarm vs. swarm confrontation while attacking the aircraft carriers[J]. Drones, 2022, 6(10): 271.
[13]	CHI P, WEI J H, WU K, et al. A Bio-inspired decision-making method of UAV swarm for attack-defense confrontation via multi-agent reinforcement learning[J]. Biomimetics, 2023, 8(2): 222.
[14]	李超, 王瑞星, 黄建忠, 等. 稀疏奖励下基于强化学习的无人集群自主决策与智能协同[J]. 兵工学报, 2023, 44(6): 1537-1546. doi: 10.12382/bgxb.2022.0177
	LI C, WANG R X, HUANG J Z, et al. Autonomous decision-making and intelligent collaboration of UAV swarms based on reinforcement learning with sparse rewards[J]. Acta Armamentarii, 2023, 44(6): 1537-1546. (in Chinese) doi: 10.12382/bgxb.2022.0177
[15]	YU H, DAI K R, LI H J, et al. Distributed cooperative guidance law for multiple missiles with input delay and topology switching[J]. Journal of the Franklin Institute, 2021, 358(17): 9061-9085.
[16]	ZHOU J L, LV Y Z, LI Z K, et al. Cooperative guidance law design for simultaneous attack with multiple missiles against a maneuvering target[J]. Journal of Systems Science & Complexity, 2018, 31(1): 287-301.
[17]	LI A, DENG Z H, HU X X. Robust cooperative attack of multiple missiles with external disturbances and communication delay[J]. IFAC-PapersOnLine, 2022, 55(3): 37-42.
[18]	刘大卫, 孙景亮, 龙腾, 等. 弹群分布式一致误差约束自适应最优协同拦截方法[J]. 兵工学报, 2023, 44(9): 2580-2590. doi: 10.12382/bgxb.2022.1160
	LIU D W, SUN J L, LONG T, et al. Distributed adaptive optimal cooperative interception method for missile swarm with synchronization error constraints[J]. Acta Armamentarii, 2023, 44(9): 2580-2590. (in Chinese) doi: 10.12382/bgxb.2022.1160
[19]	高昂, 董志明, 叶红兵, 等. 基于深度强化学习的巡飞弹突防控制决策[J]. 兵工学报, 2021, 42(5): 1101-1110. doi: 10.3969/j.issn.1000-1093.2021.05.023
	GAO A, DONG Z M, YE H B, et al. Loitering munition penetration control decision based on deep reinforcement learning[J]. Acta Armamentarii, 2021, 42(5): 1101-1110. (in Chinese) doi: 10.3969/j.issn.1000-1093.2021.05.023
[20]	陈中原, 韦文书, 陈万春. 基于强化学习的多发导弹协同攻击智能制导律[J]. 兵工学报, 2021, 42(8): 1638-1647.
	CHEN Z Y, WEI W S, CHEN W C. Reinforcement learning-based intelligent guidance law for cooperative attack of multiple missiles[J]. Acta Armamentarii, 2021, 42(8): 1638-1647. (in Chinese) doi: 10.3969/j.issn.1000-1093.2021.08.008
[21]	张婷婷, 杨学军. 基于强化学习的城市场景下巡飞弹自主协同饱和攻击方法[J]. 指挥与控制学报, 2023, 9(4): 457-468.
	ZHANG T T, YANG X J. Autonomous coordination saturation attacks method for loitering munitions in urban scenarios based on reinforcement learning[J]. Journal of Command and Control, 2023, 9(4): 457-468. (in Chinese)
[22]	YUE L F, YANG R N, ZUO J L, et al. Unmanned aerial vehicle swarm cooperative decision-making for SEAD mission: a hierarchical multi-agent reinforcement learning approach[J]. IEEE Access, 2022, 10: 92177-92191.
[23]	范祥瑞. 协同制导数据链认知抗干扰组网技术研究[D]. 北京: 中国航天科技集团公司第一研究院, 2018.
	FAN X R. The research of cognitive anti-jamming networking technology for collaborative guidance data link[D]. Beijing: The First Academy of China Aerospace Science and Technology Corporation, 2018. (in Chinese)
[24]	张曦, 水涌涛, 刘涛, 等. 具有高可靠性弹间通讯的多弹协同打击制导策略[J]. 战术导弹技术, 2019(1): 94-99.
	ZHANG X, SHUI Y T, LIU T, et al. Cooperative guidance strategy for multi-missile with high reliability communication[J]. Tactical Missile Technology, 2019(1): 94-99. (in Chinese)
[25]	金俊宇, 张婷婷. 通信带宽有限条件下无人机集群自主行为决策研究[J]. 指挥控制与仿真, 2022, 44(6): 7-15. doi: 10.3969/j.issn.1673-3819.2022.06.002
	JIN J Y, ZHANG T T. Research on autonomous behavior decision of UAV cluster with limited communication bandwidth[J]. Command Control & Simulation, 2022, 44(6): 7-15. (in Chinese)
[26]	赵飞虎, 李哲, 梁晓龙, 等. 通信带宽受限条件下的无人机集群协同搜索[J]. 电光与控制, 2022, 29(10): 12-17, 23.
	ZHAO F H, LI Z, LIANG X L, et al. Cooperative search of UAV swarm under limited communication bandwidth[J]. Electronics Optics & Control, 2022, 29(10): 12-17, 23. (in Chinese)
[27]	王瑞东, 王世练, 张炜, 等. 多弹分布式协同智能抗干扰通信策略[J]. 战术导弹技术, 2022(4): 187-195.
	WANG R D, WANG S L, ZHANG W, et al. Distributed cooperative intelligent anti-jamming communication strategy for multi-missile system[J]. Tactical Missile Technology, 2022(4): 187-195. (in Chinese)
[28]	王浩同, 刘白林, 刘智平, 等. 基于区块链的无人机集群抗干扰通信模型[J]. 火力与指挥控制, 2022, 47(1): 72-79.
	WANG H T, LIU B L, LIU Z P, et al. Anti-interference communication model of drone cluster based on blockchain[J]. Fire Control & Command Control, 2022, 47(1): 72-79. (in Chinese)
[29]	罗栋. 无通信条件下基于视觉引导的无人机编队越障技术研究[D]. 成都: 电子科技大学, 2022.
	LUO D. Research on visual guided obstacle crossing technology for unmanned aerial vehicle formation under no communication conditions[D]. Chengdu: University of Electronic Science and Technology of China, 2022. (in Chinese)
[30]	杨凯达. 加入毁伤时间流的目标毁伤效果评估方法[J]. 舰船电子工程, 2022, 42(4): 129-134,170.
	YANG K D. Method of battle damage assessment with damage time stream[J]. Ship Electronic Engineering, 2022, 42(4):129-134, 170. (in Chinese)
[31]	WU C, CHEN H R X, DU B, et al. Unsupervised change detection in multitemporal VHR images based on deep kernel PCA convolutional mapping network[J]. IEEE transactions on cybernetics, 2022, 52(11): 12084-12098.
[32]	杨青青, 樊桂花. 基于改进模糊综合评判法的建筑物毁伤效果评估[J]. 系统工程与电子技术, 2018, 40(9): 2026-2031.
	YANG Q Q, FAN G H. Battle damage assessment of buildings based on improved fuzzy comprehensive evaluation method[J]. Systems Engineering and Electronics, 2018, 40(9): 2026-2031. (in Chinese) doi: 10.3969/j.issn.1001-506X.2018.09.19
[33]	张宗腾, 张琳, 谢春燕, 等. 基于改进GA-BP神经网络的目标毁伤效果评估[J]. 火力与指挥控制, 2021, 46(11): 43-48.
	ZHANG Z T, ZHANG L, XIE C Y, et al. Battle damage effect assessment based on improved GA-BP neural network[J]. Fire Control & Command Control, 2021, 46(11): 43-48. (in Chinese)
[34]	SHEN Y, ZHU S J, YANG T J A, et al. BDANet: multiscale convolutional neural network with cross-directional attention for building damage assessment from satellite images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 5402114.
[35]	ALQAHTANI H, RAY A. Neural network-based automated assessment of fatigue damage in mechanical structures[J]. Machines, 2020, 8(4): 85.
[36]	XU J L, ZENG F, LIU W, et al. Damage detection and level classification of roof damage after typhoon faxai based on aerial photos and deep learning[J]. Applied Sciences, 2022, 12(10): 4912.
[37]	徐艺博, 于清华, 王炎娟, 等. 基于多源信息融合的巡飞弹对地目标识别与毁伤评估[J]. 系统仿真学报, 2024, 36(2): 511-521. doi: 10.16182/j.issn1004731x.joss.22-1130
	XU Y B, YU Q H, WANG Y J, et al. Ground target recognition and damage assessment of patrol missiles based on multi-source information fusion[J]. Journal of System Simulation, 2024, 36(2): 511-521. (in Chinese) doi: 10.16182/j.issn1004731x.joss.22-1130
[38]	ZHOU B L, KHOSLA A, LAPEDRIZA A, et al. Learning deep features for discriminative localization[C]// Proceedings of the 2016 Conference on Computer Vision and Pattern Recognition. Las Vegas,NV, US: IEEE, 2016: 2921-2929.