摘要
基于交感网络的适应性能力涌现是航空集群的基本特征,其构型控制与演化是产生适应性能力涌现的关键.根据航空集群构型演化与控制的基本原则和需求,将航空集群构型控制组织结构分为任务层、决策层、执行层,提出基于事件触发-规则驱动的构型演化机制;针对航空集群以实现面向任务的适应性能力涌现为目标的构型设计准则,提出将复杂任务按照能力需求分解为原子任务;以航空集群协同反隐身、无源定位、空战攻击为例,分析航空集群典型能力-构型的映射关系,结果表明,不同的构型将涌现出不同类型和层次的能力;基于多智能体系统一致性理论设计了航空集群构型生成与保持控制协议.最后,通过仿真和搭建航空集群作战演示验证系统,对所提出的理论结果进行了验证.
The adaptive ability emerging based on the sympathetic network is a basic feature of aircraft swarm,and the critical points lie in the configuration control and evolutionary mechanism.According to the basic principles and requirements of configuration control and evolution,a three-level hybrid architecture with the mission managing level,decision making level and aircraft control level was proposed,and an event-triggerred & rules-driven evolutionary mechanism was also designed.In order to achieve the mission-oriented configuration design goal of "adaptive ability emerging",complex mission was decomposed into a sequence of atom tasks based on ability requirements.The cooperative anti-stealth,passive location and air-to-air attack were given as examples to show the mapping relationships between ability and configuration.The results indicate that different configurations can "emerge" different types and levels of ability.The configuration approaching and keeping control protocol was designed based on the consensus theory of multiagent system.Finally,theoretical results were verified by simulations and experiments with a demonstration system.
作者
梁晓龙
何吕龙
张佳强
柏鹏
LIANG XiaoLong;HE LyuLong;ZHANG JiaQiang;BAI Peng(National Key Laboratory of Air Traffic Collision Prevention, Air Force Engineering University, Xi'an 710051, China;Shaanxi Province Laboratory of Meta-synthesis for Electronic & Information System, Air Force Engineering University, Xi'an 710051, China)
出处
《中国科学:技术科学》
EI
CSCD
北大核心
2019年第3期277-287,共11页
Scientia Sinica(Technologica)
基金
国家重点基础研究发展计划
国家自然科学基金(批准号:61472443
61703427)
陕西省自然科学基础研究计划(编号:2013JQ8042)资助项目
关键词
航空集群
能力涌现
构型控制
构型演化
事件触发
规则驱动
aircraft swarm
ability emerging
configuration control
configuration evolution
event trigger
rules driven