摘要
军用信息系统作为体系能力的倍增器,最近几十年取得了突出的成就.但也面临极大的挑战,尤其是以理解、推理、决策为代表的智能化认知技术成为当前信息系统智能化发展的瓶颈.本文在剖析当前军用信息系统智能化需求的基础上,深入分析了以"深绿"计划为代表的指挥信息系统智能化发展现状和不足,而以"深度学习"为代表的智能化认知技术发展为军用信息系统智能化建设带来了机遇和挑战;综合考虑体系作战的复杂性特点,提出需要重点突破的智能认知关键技术;最后,结合国防大学兵棋演习数据,采用深度学习等技术,初步实现了对作战体系威胁评估和作战态势优劣的智能化判断,展示了以深度学习为代表的智能认知技术在军事信息系统智能化建设中的潜在应用价值.
The military information system, as the amplifier of system of systems (SoS) effectiveness, have made dramatic achievements in recent decades. However, it still faced huge challenges. Especially, the intelligent cognitive technology to understand, reasoning and decision-making, has become the bottleneck of current military information systems. Based on the analysis of the current requirement for the future military information system, we analyzed current development status and disparity of the intelligent technologies in-depth compared to the requirements taking example of Deep Green Plan, which was proposed by Defense Advanced Research Projects Agency (DARPA) of USA. The great development of intelligent cognitive technology owing to “deep learning”recently brings opportunities and challenges to the intelligent construction of the military information system. Considering the complexity of SoS operations, we suggest the key technologies of intelligent cognition for the military information system which needs to break through at present. Finally,we introduced our works on evaluation of operation SoS and operational situation evaluation based on deep learning technologies, and illustrated their advantage and feasibility in methodology.
作者
郭圣明
贺筱媛
胡晓峰
吴琳
欧微
GUO Sheng-ming;HE Xiao-yuan;HU Xiao-feng;WU Lin;OU Wei(Department of Information Operation and Command Training, National Defense University of PLA, Beijing 100091, China)
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2016年第12期1562-1571,共10页
Control Theory & Applications
基金
国家自然科学基金项目(61273189
71401168
61174156
61403401
U1435218
61374179)资助~~
关键词
军用信息系统
深度学习
多层神经网络
威胁评估
态势判断
military information system
deep learning
multilayer neural networks
threat evaluation
operational situation evaluation