摘要
为实现真正意义上的智能化、自主化无人作战,提出基于AI开发板构建作战任务计算机作为无人车的控制核心,模拟无人车作战机动态势图,并利用深度强化学习网络DQN建立角度及距离决策网络,实现无人车的智能机动决策。通过实验证实了无人车能够自主机动至目标区域,通过深度强化学习网络能实现平台自主、智能化决策,可为构建作战任务计算机实现真正意义上的智能化、自主化、无人化作战,提供可行技术途径及理论支撑。
With the appearance of the advantages of an unmanned combat platform in modern wars,research on unmanned combat platforms has become the focus of all circles.In order to realize intelligent and autonomous unmanned combats in a real sense,this paper constructs a combat mission computer as the control core of unmanned vehicles based on AI development board.It also simulates unmanned vehicle combat mobility situation diagram,and realizes intelligent maneuvering decisions of unmanned vehicles by using the deep reinforcement learning network DQN to establish the angle and the distance decision network.The experiment verifies that the unmanned vehicle can maneuver to the target area autonomously.Through the deep reinforcement learning network,the platform realizes autonomous and intelligent decision making,which provides a feasible technical approach and theoretical support for the construction of combat mission computers to realize intelligent,autonomous and unmanned combats.
作者
王瑶
马海强
李梓正
姜义
WANG Yao;MA Haiqiang;LI Zizheng;JIANG Yi(Naval Aviation University,Yantai 264001,China;Unit 32108 of the Chinese People’s Liberation Army,Manzhouli 021400,China)
出处
《兵器装备工程学报》
CAS
CSCD
北大核心
2023年第S01期198-205,共8页
Journal of Ordnance Equipment Engineering
基金
装备预研领域基金项目(6140247030216JB14004)。
关键词
无人作战
深度学习
强化学习
作战计算机
智能决策
unmanned combat
deep learning
reinforcement learning
combat computer
intelligent decision
