-
摘要:
战术活动识别是战场态势感知的重要研究内容。为提高战术活动识别的准确性与实时性,提出了一种基于上下文独立动态贝叶斯网络(CIDBN)的战术活动识别模型及在线精确推理。通过对战术活动机制的分析,采用动态贝叶斯网络(DBN)理论,建立了一个初始战术活动识别模型。该模型引入了威胁指数节点来影响战术活动的终止与选择,并采用模糊隶属度函数对连续变量进行离散化处理。依据上下文独立关系对该模型进行简化,获得了一个基于CIDBN的战术活动识别模型。将接口算法扩展于该模型上,提出了在线精确推理算法。仿真结果表明,所提出的战术活动识别方法,具有识别精度高、较低不确定性和实时性高的优点。
-
关键词:
- 动态贝叶斯网络(DBN) /
- 接口算法 /
- 上下文独立 /
- 威胁指数 /
- 精确推理
Abstract:Tactical activity recognition is an important research content of battlefield situational awareness. In order to improve the accuracy and real-time of tactical activity recognition, a tactical activity recognition model and online accurate reasoning based on Context-Independent Dynamic Bayesian Network (CIDBN) are put forward. Based on the analysis of tactical activity mechanism and Dynamic Bayesian Network (DBN) theory, an initial tactical activity recognition model is established. In this model, threat index nodes are introduced to influence the termination and selection of activities, and the fuzzy membership function is used to discretize the continuous variables. The model is simplified based on the relationship of context independence, and the new tactical activity recognition model based on CIDBN is obtained. The interface algorithm is extended to the model and an online accurate reasoning algorithm is proposed. The simulation results show that the proposed tactical activity recognition method has the advantages of high recognition accuracy, low uncertainty and high real-time performance.
-
表 1 战术活动识别的结果
Table 1. Results of tactical activity recognition
测试 HMM CRF CIDBN Acc U Acc U Acc U T-1 0.842 5 1.642 0 0.913 8 1.185 0 0.944 9 1.020 6 T-2 0.865 7 1.969 6 0.921 4 1.410 3 0.952 2 1.012 3 表 2 平均单步运行所需时间
Table 2. Average time of one-step running
算法 平均的单步推理时间/s 团结树算法(DBN) 0.007t+0.01 边界算法(DBN) 0.092 9 向前算法(DBN) 0.207 0 接口算法(DBN) 0.065 3 接口算法(CIDBN) 0.025 3 -
[1] 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. http://d.old.wanfangdata.com.cn/Periodical/xtgcydzjs201608020CHEN H, REN Q L, HUA Y, et al.Fuzzy neural network based tactical intention recognition for sea targets[J].System Engineering and Electronics, 2016, 38(8):1847-1853(in Chinese). http://d.old.wanfangdata.com.cn/Periodical/xtgcydzjs201608020 [2] 张天赫, 彭绍雄, 邹强, 等.无监督神经网络的潜艇对空战术意图识别[J].现代防御技术, 2018, 46(2):122-129. doi: 10.3969/j.issn.1009-086x.2018.02.020ZHANG T H, PENG S X, ZOU Q, et al.Unsupervised learning neural network based submarine recognize tactical intention for air target[J].Modern Defence Technology, 2018, 46(2):122-129(in Chinese). doi: 10.3969/j.issn.1009-086x.2018.02.020 [3] 欧微, 柳少军, 贺筱媛, 等.战场对敌目标战术意图智能识别模型研究[J].计算机仿真, 2017, 34(9):10-14. doi: 10.3969/j.issn.1006-9348.2017.09.003OU W, LIU S J, HE X Y, et al.Study on the intelligent recognition model enemy target's tactical intention on battlefield[J].Computer Simulation, 2017, 34(9):10-14(in Chinese). doi: 10.3969/j.issn.1006-9348.2017.09.003 [4] 尹光辉, 朱英贵, 汪洋.基于贝叶斯网络的装甲目标战术企图推理模型构建[J].火力与指挥控制, 2015, 40(7):122-125. doi: 10.3969/j.issn.1002-0640.2015.07.030YIN G H, ZHU Y G, WANG Y.Construction of models of armored target tactical intention reasoning based on Bayesian network[J].Fire Control & Command Control, 2015, 40(7):122-125(in Chinese). doi: 10.3969/j.issn.1002-0640.2015.07.030 [5] 余振翔, 胡笑旋, 夏维.基于模糊动态贝叶斯网空战敌方作战企图识别[J].合肥工业大学学报(自然科学版), 2013, 36(10):1210-1216. doi: 10.3969/j.issn.1003-5060.2013.10.013YU Z X, HU X X, XIA W.Foe intention inference in air combat based on fuzzy dynamic Bayesian network[J].Journal of Hefei University of Technology (Natural Science), 2013, 36(10):1210-1216(in Chinese). doi: 10.3969/j.issn.1003-5060.2013.10.013 [6] 葛顺, 夏学知.用于战术意图识别的动态序列贝叶斯网络[J].系统工程与电子技术, 2014, 36(1):76-83. http://d.old.wanfangdata.com.cn/Periodical/xtgcydzjs201401012GE S, XIA X Z.DSBN used for recognition of tactical intention[J].Systems Engineering and Electronics, 2014, 36(1):76-83(in Chinese). http://d.old.wanfangdata.com.cn/Periodical/xtgcydzjs201401012 [7] 葛顺.基于规则发现和贝叶斯推理的战术意图识别技术[D].哈尔滨: 哈尔滨工程大学, 2015. http://cdmd.cnki.com.cn/Article/CDMD-10217-1018048235.htmGE S.Tactical intention recognition technology based on rule discovery and Bayesian reasoning[D].Harbin: Harbin Engineering University, 2015(in Chinese). http://cdmd.cnki.com.cn/Article/CDMD-10217-1018048235.htm [8] DUONG T V, BUI H H, PHUNG D Q, et al.Activity recognition and abnormality detection with the switching Hidden semi-Markov model[C]//Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Piscataway: IEEE Press, 2005: 838-845. [9] LIAO L, FOX D, KAUTZ H.Learning and inferring transportation routines[J].Artificial Intelligence, 2007, 171:311-331. doi: 10.1016/j.artint.2007.01.006 [10] ZHAO L, WANG X, SUKTHANKAR G, et al.Motif discovery and feature selection for CRF-based activity recognition[C]//2010 20th International Conference on Pattern Recognition (ICPR).Piscataway: IEEE Press, 2010: 3826-3829. [11] SÁNCHEZ D, TENTORI M, FAVELA J.Activity recognition for the smart hospital[J].IEEE Intelligent Systems, 2008, 23(2):50-57. http://cn.bing.com/academic/profile?id=96c8cb014f0eaba4910aba3340bc99d8&encoded=0&v=paper_preview&mkt=zh-cn [12] OGRIS G, STIEFMEIER T, LUKOWICZ P, et al.Using a complex multi-modal on-body sensor system for activity spotting[C]//2008 12th IEEE International Symposium on Wearable Computers.Piscataway: IEEE Press, 2008: 55-62. [13] LORINCZ K, MALAN D J, FULFORD-JONES T R F, et al.Sensor networks for emergency response:Challenges and opportunities[J].IEEE Pervasive Computing, 2004, 3(4):16-23. doi: 10.1109-MPRV.2004.18/ [14] MURPHY K.The Bayes net toolbox for MATLAB[J].Computing Science and Statistics, 2001, 33(2):1024-1034. http://cn.bing.com/academic/profile?id=ad8a4cacd9afbd5f005e4bbf44871481&encoded=0&v=paper_preview&mkt=zh-cn [15] POOLE D, ZHANG N L.Exploiting contextual independence in probabilistic inference[J].Journal of Artificial Intelligence Research, 2003, 18:263-313. doi: 10.1613/jair.1122 [16] 杨春, 郭健, 张磊, 等.采用卡方检验的模糊自适应无迹卡尔曼滤波组合导航算法[J].控制与决策, 2018, 33(1):81-87. http://d.old.wanfangdata.com.cn/Periodical/kzyjc201801010YANG C, GUO J, ZHANG L, et al.Fuzzy adaptive unscented Kalman filter integrated navigation algorithm using Chi-square test[J].Control and Decision, 2018, 33(1):81-87(in Chinese). http://d.old.wanfangdata.com.cn/Periodical/kzyjc201801010 [17] 文飞, 吕艳, 段刚, 等.空空威胁评估在综合辅助决策系统中的应用研究[J].系统仿真学报, 2009, 21(12):3734-3737. http://d.old.wanfangdata.com.cn/Periodical/xtfzxb200912055WEN F, LV D, DUAN G, et al.Application study of air-air assessment in synthetic decision aiding system[J].Journal of System Simulation, 2009, 21(12):3734-3737(in Chinese). http://d.old.wanfangdata.com.cn/Periodical/xtfzxb200912055 [18] 张乐, 曹爽, 李士雪, 等.层次分析法的改进及在权重确定中的应用[J].中国卫生统计, 2016, 33(1):154-158. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgwstj201601051ZHANG L, CAO S, LI S X, et al.The improvement of AHP and its application in weight determination[J].Chinese Journal of Health Statistics, 2016, 33(1):154-158(in Chinese). http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=zgwstj201601051 [19] PATRICK M K.Dynamic Bayesian networks: Representation, inference and learning[D].Berkeley: University of California, 2002.