空战场多分支态势仿真生成方法
|
摘要
针对空战场态势要素多、影响因素复杂和不确定性强的特点,提出了空战场环境下多分支态势仿真生成方法。该方法将空战场态势预测分为战场实时态势情报接收、敌方目标分群和多分支态势生成3个模块。首先,对态势信息进行收集、格式化及分类存储,为后续多分支态势生成奠定基础;然后,对敌方目标多维属性信息进行深层次抽象,从而将众多敌方目标分解成多个战术编队,达到降解态势分支生成复杂度的目的;最后,结合深度神经网络和贝叶斯网络等方法生成主分支及旁路分支,并对多分支态势进行规范化表征。通过案例分析可见,该方法可提前预测战场态势信息,为指挥员快速、科学地决策提供支撑。
Aimed at the air battlefield situation features of many factors, complex influencing factors and strong uncertainty, a multi-branch situation simulation generation method is proposed. The method is divided the air battlefield situation prediction into three modulars, including the battlefield real-time situation intelligence collection modular, the hostile target clustering modular and multi-branch situation generation modular. Firstly, the situation information is collected, formatted and stored in classifications, and the foundation is established for the future multi-branch situation generation. Then, the multi-dimensional attribute information of hostile targets is deeply abstracted. Thus, the hostile targets are classified into multiple tactical formations to achieve the goal of degrading the complexity of the situation branch generation. Finally, combined with the methods of the deep neural network and the Bayesian network, the main branch and bypass branches are generated, and the standardizing multi-branch situation is represented. The case analysis shows that the method can predict the battlefield situation information in advance and provide support for commanders′ quick and scientific decision-makings.
Aimed at the air battlefield situation features of many factors, complex influencing factors and strong uncertainty, a multi-branch situation simulation generation method is proposed. The method is divided the air battlefield situation prediction into three modulars, including the battlefield real-time situation intelligence collection modular, the hostile target clustering modular and multi-branch situation generation modular. Firstly, the situation information is collected, formatted and stored in classifications, and the foundation is established for the future multi-branch situation generation. Then, the multi-dimensional attribute information of hostile targets is deeply abstracted. Thus, the hostile targets are classified into multiple tactical formations to achieve the goal of degrading the complexity of the situation branch generation. Finally, combined with the methods of the deep neural network and the Bayesian network, the main branch and bypass branches are generated, and the standardizing multi-branch situation is represented. The case analysis shows that the method can predict the battlefield situation information in advance and provide support for commanders′ quick and scientific decision-makings.
引文
[1] 王杨.战场态势目标识别与态势意图预测的算法研究[D].无锡:江南大学,2015.
[2] 柳玉,文家焱,陈建华.计算机兵棋系统发展及应用研究[J].兵工自动化,2015,34(8):20-26.
[3] 朱丰,胡晓峰.基于深度学习的战场态势高级理解模拟方法[J].火力与指挥控制,2018,43(8):25-30.
[4] ENDSLEY M R,GARLAND D J.Situation awareness analysis and measurement[M].Mahwah:Lavorence Erlbaum Associates,2000.
[5] 方冰,张翠侠.基于时空维度分析的战场态势预测方法[J].指挥信息系统与技术,2017,8(1):59-64.
[6] 周志强,钱建刚,尹康银,等.一种基于DS证据理论的目标意图预测方法[J].空军预警学院学报,2014,28(2):116-118.
[7] 易卓,廖鹰,胡晓峰,等.基于深度时空循环神经网络的协同作战行动识别[J].系统仿真学报,2018,30(3):793-800.
[8] 郭圣明,贺筱媛,吴琳,等.基于强制稀疏自编码神经网络的作战态势评估方法研究[J].系统仿真学报,2018,30(3):772-784.
[9] 黄柯棣,赵鑫业,杨山亮,等.军事分析仿真评估系统关键技术综述[J].系统仿真学报,2012,24(12):2439-2447.
[10] 胡晓峰.军事指挥信息系统中的机器智能:现状与趋势[J].人民论坛·学术前沿,2016(15):22-34.
[11] 胡晓峰,荣明.作战决策辅助向何处去:“深绿”计划的启示与思考[J].指挥与控制学报,2016,2(1):22-25.
[12] 周云,黄教民,黄柯棣.深绿计划关键技术研究综述[J].系统仿真学报,2013,25(7):1633-1638.
[13] 郭圣明,贺筱媛,胡晓峰,等.军用信息系统智能化的挑战与趋势[J].控制理论与应用,2016,33(12):1562-1571.
[14] 金欣.“深绿”及AlphaGo对指挥与控制智能化的启示[J].指挥与控制学报,2016,2(3):202-207.
[15] 周云,黄教民,黄柯棣.美国“深绿”计划对指挥控制的影响[J].火力与指挥控制,2013,38(6):1-5.
[16] 郑昌艳,梅卫.基于卷积神经网络的空中目标战术机动模式分类器设计[J].微型机与应用,2015,34(22):50-52.
[17] 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.
[18] 郑昌艳,梅卫,冯小雨.基于多时间窗口的空中目标机动模式提取及识别方法[J].探测与控制学报,2016,38(5):81-87.
[2] 柳玉,文家焱,陈建华.计算机兵棋系统发展及应用研究[J].兵工自动化,2015,34(8):20-26.
[3] 朱丰,胡晓峰.基于深度学习的战场态势高级理解模拟方法[J].火力与指挥控制,2018,43(8):25-30.
[4] ENDSLEY M R,GARLAND D J.Situation awareness analysis and measurement[M].Mahwah:Lavorence Erlbaum Associates,2000.
[5] 方冰,张翠侠.基于时空维度分析的战场态势预测方法[J].指挥信息系统与技术,2017,8(1):59-64.
[6] 周志强,钱建刚,尹康银,等.一种基于DS证据理论的目标意图预测方法[J].空军预警学院学报,2014,28(2):116-118.
[7] 易卓,廖鹰,胡晓峰,等.基于深度时空循环神经网络的协同作战行动识别[J].系统仿真学报,2018,30(3):793-800.
[8] 郭圣明,贺筱媛,吴琳,等.基于强制稀疏自编码神经网络的作战态势评估方法研究[J].系统仿真学报,2018,30(3):772-784.
[9] 黄柯棣,赵鑫业,杨山亮,等.军事分析仿真评估系统关键技术综述[J].系统仿真学报,2012,24(12):2439-2447.
[10] 胡晓峰.军事指挥信息系统中的机器智能:现状与趋势[J].人民论坛·学术前沿,2016(15):22-34.
[11] 胡晓峰,荣明.作战决策辅助向何处去:“深绿”计划的启示与思考[J].指挥与控制学报,2016,2(1):22-25.
[12] 周云,黄教民,黄柯棣.深绿计划关键技术研究综述[J].系统仿真学报,2013,25(7):1633-1638.
[13] 郭圣明,贺筱媛,胡晓峰,等.军用信息系统智能化的挑战与趋势[J].控制理论与应用,2016,33(12):1562-1571.
[14] 金欣.“深绿”及AlphaGo对指挥与控制智能化的启示[J].指挥与控制学报,2016,2(3):202-207.
[15] 周云,黄教民,黄柯棣.美国“深绿”计划对指挥控制的影响[J].火力与指挥控制,2013,38(6):1-5.
[16] 郑昌艳,梅卫.基于卷积神经网络的空中目标战术机动模式分类器设计[J].微型机与应用,2015,34(22):50-52.
[17] 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.
[18] 郑昌艳,梅卫,冯小雨.基于多时间窗口的空中目标机动模式提取及识别方法[J].探测与控制学报,2016,38(5):81-87.