反应式集群体系结构在蜂群弹药中的应用研究

Research on Application of Reactive Architecture in Ammunition Swarm

随着无人武器集群式作战从想象加速变为现实,这种新型武器和战术势必将成为改变未来战争规则的重要推手。针对集群武器系统缺乏有效理论与系统技术支撑的问题,首先在多智能体理论和仿生集群智能技术的启发下,深入研究了基于集群行为驱动和个体行为规则的分布式决策框架,并结合视觉信息处理技术,构建了基于个体行为规则和集群行为原型的反应式集群框架;然后,建立了蜂群弹药任务和个体行为规则之间的映射关系;最后,将简单的区域搜索、饱和攻击等作战任务进行了嵌入式实现,对影响任务效率的关键因素进行了深入分析,为构建蜂群弹药等智能集群装备提供了技术储备和理论支持。初步仿真结果证明,对于50节点规模的集群,提出的反应式集群决策框架能够引导蜂群弹药高效完成避障、搜索和攻击等任务,且在40%、80%个体失效时搜索效率显著优于慎思式集群。

As the swarm of unmanned weapons turns from imagination to reality,this new type of weapons and tactics will certainly become an important driver to change the rules of future war.Aiming at the problem of lack of effective theory and system technology support for swarm weapon system,this paper inspired by multi-agent theory and biomimetic swarm intelligence technology,studies the distributed decision-making framework based on swarm behavior and individual behavior rules.Combined with visual information processing technology,a reactive swarm framework based on individual behavior rules and cluster behavior prototypes is constructed,and the mapping relationship between swarm ammo missions and individual rules of conduct is established.This paper attempts to explore a feasible way to realize the use of ammunition swarm,makes an embedded-level implementation of simple area search and saturation attack,and in-depth analysis of key factors affecting efficiency.The preliminary simulation results show that for the swarm with 50 nodes,the reactive swarm decision-making framework proposed can guide the hive munitions to efficiently complete tasks such as obstacle avoidance,search and attack.The search efficiency is significantly better than the deliberative swarm when 40%and 80%of individuals fail.