基于深度学习的战场态势高级理解模拟方法

Simulation Method of Battlefields Situation Senior Comprehension Based on Deep Learning

当今军事领域问题的研究已步入了信息化主导的大数据时代,传统的智能辅助指挥员完成战场态势理解已遇到了瓶颈,亟需探索突破。认知智能中深度学习的提出可为该问题的突破提供契机。通常,指挥员进行战场态势理解是分层次的。其中的高级理解则需要深度学习来模拟。就此,展开了探索性研究。概述了战场态势评估的相关概念,分析了指挥员理解战场态势的思维模式,掌握了指挥员理解战场态势时的主要步骤,并结合深度学习运行原理,提出了一种基于深度学习的指挥员战场态势高级理解思维过程(以判断敌方对我方可能的主攻方向为例)模拟方法,该方法利用认知智能中的深度学习(以CNN为例)对指挥员战场态势高级理解过程进行非线性拟合处理,从而达到探索性模拟的目的,仿真实验结果验证了该方法的有效性。

Nowadays,the research of military field has entered into the era of Big Data dominated by information technology.The traditional intelligent assistant commander has been confronted with the bottleneck of the battlefield situation comprehension,needing to explore to breakthrough.Deep Learning within cognitive intelligence can provide an opportunity for the breakthrough.Usually,battlefield situation comprehension by commanders is hierarchical.However,the period of senior comprehension needs Deep Learning to simulate.Aimed at the issue of senior comprehension simulation,some exploring researches are carried out in the paper.The relative concepts of the battlefield situation assessment are generalized.The battlefield situation comprehension mode of thinking is analyzed,and the main steps of the battlefield situation comprehension by commanders are grasped.And combined with the operation principle of Deep Learning,an intelligent simulation method of the thinking process(example as judging the main offensive direction of enemy)of senior compre原hension of the commanders based on Deep Learning is put forward.In the method,Deep Learning(ex原ample as CNN)within cognitive intelligence is utilized to carry on the non-linear fitting for process of the commanders’battlefield situation senior comprehension,so as to achieve the purpose of simulation.The simulation results verify the effectiveness of the proposed method