双重视角下的无人集群智能性描述与等级划分

Unmanned Swarm Intelligence Description and Level Classification Based on Dual Perspectives

无人集群是群体智能的一个重要应用领域,也是无人系统的重要发展方向之一。如何描述和度量无人集群整体的智能性水平是当前无人系统领域一个十分具有挑战性的理论与技术研究课题。该文首先对已有无人系统智能性等级划分模型进行了述评;其次,结合对群体智能和系统智能概念的阐述,给出了对无人集群智能性的合理解释;接着,在无人集群系统内部与外部双重视角下全面剖析了影响其整体智能性水平评判的关键要素,基于任务和环境复杂性因素建立了无人集群的任务胜任力度量模型,在扩展经典OODA(Observe,Orient,Decide,Act)环的基础上建立了基于群体化OODA环的无人集群整体智能行为实现过程模型,并提出了一种基于两种模型关联映射的无人集群智能性描述框架;在此基础上,提出了一种比较详细的无人集群智能性等级划分;最后,对持续深化该研究的主要发展方向进行了总结与展望。

Unmanned swarm is an important application field of swarm intelligence and one of the important development trends of unmanned systems.How to describe and measure the overall intelligence of unmanned swarms is a very challenging theoretical and technical research topic in the field of unmanned systems at present.Firstly,we review the existing classification models for the intelligence level of unmanned systems.Secondly,we give a reasonable explanation for unmanned swarm intelligence by combining with the conceptual elaborations of swarm intelligence and system intelligence.Then,we analyze the key factors influencing the overall intelligence evaluation conclusions from the internal and external perspectives of unmanned swarm systems,establish the task competency evaluation model of unmanned swarms according to task and environment complexity as well as the overall intelligent behavior implementation process model of unmanned swarms based on the collective OODA(Observe,Orient,Decide,Act)loop by expanding the classical OODA loop,and propose the description framework for unmanned swarm intelligence based on the association mapping of the two models.Furthermore,we propose a more detailed level classification for unmanned swarm intelligence.Finally,the main directions worthy of continuously deepening the research are summarized and prospected.