摘要
阐述了未来战争中无人集群协同作战的概念内涵和发展现状,分析了基于强化学习的无人集群协同的特点,提出了基于多智能体强化学习的集中训练-分布执行的无人集群协同设计方案,完成了集群分散任务的场景建模和奖励设计,并在Player/Stage平台进行了仿真试验。仿真试验表明,使用多智能体强化学习的方法可实现无人集群协作。最后,展望了无人集群作战技术发展方向。
The concept and the development present situation of unmanned swarm are expounded,the characteristics of unmanned swarm cooperation based on reinforcement learning are analyzed,and the unmanned swarm collaborative design scheme of the centralized training-distributed execution based on multi-agent reinforcement learning is proposed,the scenario modeling and the reward design of the swarm dispersing tasks are achieved,and the simulation experiments are carried out in the Player/Stage platform.The simulation experimental results show the multi-agent reinforcement learning method can be used to achieve the unmanned swarm collaboration.Finally,the development direction of unmanned swarm warfare technology is prospected.
作者
郑健
陈建
朱琨
ZHENG Jian;CHEN Jian;ZHU Kun(College of Computer Science and Technology,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China)
出处
《指挥信息系统与技术》
2020年第6期26-31,共6页
Command Information System and Technology
基金
中国电科集团信息系统需求重点实验室开放基金(LHZZ2020-F01)资助项目。
关键词
无人集群
协同控制
强化学习
多智能体强化学习
多智能体系统
unmanned swarm
cooperative control
reinforcement learning
multi-agent reinforcement learning
multi-agent system