样本不平衡下的空中群组意图识别方法

doi:10.12305/j.issn.1001-506X.2022.12.19

摘要/Abstract

摘要：

针对样本不平衡情况下空中群组意图识别问题, 提出了一种基于注意力机制的双向门控循环单元网络空中群组意图识别方法。将空中群组的状态信息编码成时序特征, 利用先验信息封装成样本标签, 提出一种改进的边界合成少数过采样方法对少数类样本边界进行处理, 以生成合适的样本集。通过双向循环机制提取空中群组目标时序特征的深层信息, 并利用注意力机制为深层信息分配权值, 以提高网络捕捉更具区分度特征的能力。实验仿真结果表明, 所提方法在样本不平衡情况下空中群组的意图识别问题上有更好的分类效果和更高的训练效率。

关键词: 意图识别, 双向门控循环单元, 注意力机制, 样本不平衡

Abstract:

To solve the problem of air group intention recognition under the condition of imbalance samples, a bidirectional gated recurrent unit network air group intention recognition method based on the attention mechanism is proposed. The state information of the air group is encoded into timing features. Prior information is used to encapsulate into sample labels. An improved borderline-synthetic minority oversampling method is proposed to generate a suitable sample set. The timing features of aerial group targets can be deeply extracted by bidirectional mechanism. Attention mechanism is introduced to improve the network's ability to capture more discriminative features by assigning weights to deep information. Experiment simulation results show that proposed method has better classification effect and higher training efficiency for the intent recognition problem of the air group in the case of imbalance samples.

Key words: intention recognition, bi-gated recurrent unit (BiGRU), attention mechanism, imbalance samples

中图分类号:

E824

马钰棠, 孙鹏, 张杰勇, 王鹏, 闫云飞, 赵亮. 样本不平衡下的空中群组意图识别方法[J]. 系统工程与电子技术, 2022, 44(12): 3747-3755.

Yutang MA, Peng SUN, Jieyong ZHANG, Peng WANG, Yunfei YAN, Liang ZHAO. Air group intention recognition method under imbalance samples[J]. Systems Engineering and Electronics, 2022, 44(12): 3747-3755.

图/表 16

图1

图2

图3

图4

图5

图6

表1

表2

表3

表4

图7

表5

图8

图9

图10

表6

参考文献 30

1	AZAREWICZ J, FALA G, HEITHECKER C. Template-based multi-agent plan recognition for tactical situation assessment[C]// Proc. of the 5th Conference on Artificial Intelligence for Applications, 1989.
2	夏曦. 基于模板匹配的目标意图识别方法研究[D]. 长沙: 国防科学技术大学, 2006.
	XIA X. The study of target intent assessment method based on the template-matching[D]. Changsha: School of National University of Defense Technology, 2006.
3	CHANG L L , ZHOU Z J , YOU Y , et al. Belief rule based expert system for classification problems with new rule activation and weight calculation procedures[J]. Information Sciences, 2016, 336 (1): 75- 91.
4	YIN X , ZHANG M , CHEN M Q . Combat intention recognition of the target in the air based on discri-minant analysis[J]. Journal of Projectiles, Rockets, Missiles and Guidance, 2018, 38 (3): 46- 50.
5	ZHOU T L , CHEN M , CHEN S , et al. Intention prediction of aerial target under incomplete information[J]. ICIC Express Letters An International Journal of Research and Surveys, 2017, 8, 623- 631.
6	戴革林, 陈伟, 刘志坚, 等. 基于区间灰关联度的飞机战术意图识别方法[J]. 数学的实践与认识, 2014, 44 (20): 198- 207.
	DAI G L , CHEN W , LIU Z J , et al. Method of target tactical intention recognition based on interval grey relational degree[J]. Mathematics in Practice and Theory, 2014, 44 (20): 198- 207.
7	CHEN Z G , WU X F . A novel multi-timescales layered intention recognition method[J]. Applied Mechanics and Materials, 2014, 644-650, 4607- 4611. doi: 10.4028/www.scientific.net/AMM.644-650.4607
8	XU Y H, CHENG S Y, ZHANG H B, et al. Air target combat intention identification based on IE-DSBN[C]//Proc. of the International Workshop on Electronic Communication and Artificial Intelligence, 2020: 36-40.
9	JIN Q, GOU X T, JIN W D, et al. Intention recognition of aerial targets based on Bayesian optimization algorithm[C]//Proc. of the IEEE International Conference on Intelligent Transportation Engineering, 2017: 356-359.
10	欧微, 柳少军, 贺筱媛, 等. 基于时序特征编码的目标战术意图识别算法[J]. 指挥控制与仿真, 2016, 38 (6): 36- 41. doi: 10.3969/j.issn.1673-3819.2016.06.008
	OU W , LIU S J , HE X Y , et al. Tactical intention recognition algorithm based on encoded temporal features[J]. Command Control & Simulation, 2016, 38 (6): 36- 41. doi: 10.3969/j.issn.1673-3819.2016.06.008
11	周旺旺, 姚佩阳, 张杰勇, 等. 基于深度神经网络的空中目标作战意图识别[J]. 航空学报, 2018, 39 (11): 200- 208.
	ZHOU W W , YAO P Y , ZHANG J Y , et al. Combat intention recognition for aerial targets based on deep neural network[J]. Acta Aeronautica et Astronautica Sinica, 2018, 39 (11): 200- 208.
12	陈浩, 任卿龙, 滑艺, 等. 基于模糊神经网络的海面目标战术意图识别[J]. 系统工程与电子技术, 2016, 38 (8): 1847- 1853.
	CHEN H , REN Q L , HUA Y , et al. Fuzzy neural network based tactical intention recognition for sea targets[J]. Systems Engineering and Electronics, 2016, 38 (8): 1847- 1853.
13	XUE J J , ZHU J , XIAO J Y , et al. Panoramic convolutional long short-term memory networks for combat intension recognition of aerial targets[J]. IEEE Access, 2020, 8, 183312- 183323.
14	ZHANG T W , ZHANG X L , LIU C , et al. Balance learning for ship detection from synthetic aperture radar remote sensing imagery[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 182, 190- 207.
15	CHAWLA N V , BOWYER K W , HALL L O , et al. SMOTE: synthetic minority over-sampling technique[J]. The Journal of Artificial Intelligence Research, 2002, 16, 321- 357.
16	YEN S J , LEE Y S . Cluster-based under-sampling approaches for imbalanced data distributions[J]. Expert Systems with Applications, 2009, 36 (3): 5718- 5727.
17	RAMENTOL E , CABALLERO Y , BELLO R , et al. Smote-RSB^*: a hybrid preprocessing approach based on oversampling and undersampling for high imbalanced data-sets using smote and rough sets theory[J]. Knowledge and Information Systems, 2012, 33 (2): 245- 265.
18	LEE Z J , LEE C Y , CHOU S T , et al. A hybrid system for imbalanced data mining[J]. Microsystem Technologies, 2020, 26, 3043- 3047.
19	GYOTEN D , OHKUBO M , NAGATA Y . Imbalanced data classification procedure based on SMOTE[J]. Total Quality Science, 2020, 5 (2): 64- 71.
20	BATISTA G , PRATI R C , MONARD M C . A study of the behavior of several methods for balancing machine learning training data[J]. ACM Sigkdd Explorations Newsletter, 2004, 6 (1): 20- 29.
21	GARCIA S , LUENGO J , HERRERA F . Tutorial on practical tips of the most influential data preprocessing algorithms in data mining[J]. Knowledge-Based Systems, 2016, 98, 1- 29.
22	FERNANDEZ A , GARCIA S , CHAWLA N V , et al. SMOTE for learning from imbalanced data: progress and challenges, marking the 15-year anniversary[J]. Journal of Artificial Intelligence Research, 2018, 61, 863- 905.
23	HUSSEIN A S , LI T , CHUBATO W Y , et al. A-SMOTE: a new preprocessing approach for highly imbalanced datasets by improving SMOTE[J]. International Journal of Computational Intelligence Systems, 2019, 12 (2): 1412- 1422.
24	CHEN B Y , XIA S Y , CHEN Z Z , et al. RSMOTE: a self-adaptive robust SMOTE for imbalanced problems with label noise[J]. Information Sciences, 2021, 553, 397- 428.
25	CHEN J X , JIANG D M , ZHANG Y N . A hierarchical bidirectional GRU model with attention for EEG-based emotion classification[J]. IEEE Access, 2019, 7, 118530- 118540.
26	HUI H, WANG W Y, MAO B H. Borderline-SMOTE: a new over-sampling method in imbalanced data sets learning[C]//Proc. of the in Intelligent Computing, 2005: 878-887.
27	MUKHERJEE M , KHUSHI M . SMOTE-ENC: a novel SMOTE-based method to generate synthetic data for nominal and continuous features[J]. Applied System Innovation, 2021, 4 (1): 18.
28	ZHOU T L , CHEN M , WANG Y H , et al. Information entropy-based intention prediction of aerial targets under uncertain and incomplete information[J]. Entropy, 2020, 22 (3): 279.
29	刘钻东, 陈谋, 吴庆宪, 等. 非完备信息下无人机空战目标意图预测[J]. 中国科学: 信息科学, 2020, 50 (5): 704- 717.
	LIU Z D , CHEN M , WU Q X , et al. Prediction of unmanned aerial vehicle target intention under incomplete information[J]. Scientia Sinica Informationis, 2020, 50 (5): 704- 717.
30	ZHAO X S, SHAO Y B, MAI J Y, et al. Respiratory sound classification based on BiGRU-Attention network with XGBoost[C]//Proc. of the IEEE International Conference on Bioinformatics and Biomedicine, 2020: 915-920.

参数	取值
损失函数	categorical_crossentroy
学习率	0.001
隐藏层节点数	128, 64, 64
批大小	128
训练周期	100
输入形状	12×11
优化器	Adam

样本类型	准确率/%
样本类型	原始样本	合成样本
攻击	60.00	62.33
监视	32.50	33.72
撤退	100	99.58

意图类型	攻击	佯攻	撤退	侦察	监视	电子干扰
攻击	9	6	0	0	0	0
佯攻	0	223	0	0	0	0
撤退	0	0	137	0	0	0
侦察	0	0	0	786	0	0
监视	0	0	0	27	13	0
电子干扰	0	0	0	0	0	399

意图类型	攻击	佯攻	撤退	侦察	监视	电子干扰
攻击	13	1	0	0	0	1
佯攻	2	221	0	0	0	0
撤退	0	0	137	0	0	0
侦察	0	0	0	774	12	0
监视	0	0	0	9	31	0
电子干扰	0	0	0	0	0	399

模型	隐藏层节点数	学习率	迭代轮次
BiGRU-Attention	128, 64, 64	0.001	100
SAE	256, 128, 128	0.001	200
DNN	50, 100, 100, 50	0.001	100
LSTM	128, 64, 64	0.001	200