大规模UAV编队信息交互拓扑的分级分布式生成  

Hierarchical and distributed generation of information interaction topology for large scale UAV formation

作　　者：董文奇 何锋[1] DONG Wenqi;HE Feng(School of Electronics and Information Engineering,Beihang University,Beijing 100083,China)

机构地区：[1]北京航空航天大学电子信息工程学院,北京100083

出　　处：《航空学报》2021年第6期441-451,共11页Acta Aeronautica et Astronautica Sinica

基　　金：国家自然科学基金(62071023);装备预研领域基金(61403120404)。

摘　　要：UAV编队信息交互拓扑的优化设计是保证UAV编队安全性和任务执行高效性的重要基础。目前队形保持下UAV编队信息交互拓扑生成算法局限于小规模编队,且优化目标单一。针对这一问题,采用了分级分簇结构扩展信息交互拓扑层级,以满足大规模场景,同时以提高编队续航能力和减少编队总通信代价为组合优化目标,提出了基于最小树形图的分级分布式领航-跟随者编队信息交互拓扑生成算法,并通过OMNeT++进行了仿真验证。实验结果表明在考虑位置误差传递迭代时,分级分布式领航-跟随者编队的总通信代价明显低于传统领航-跟随者编队,且通过周期性更新簇首,网络能耗更加均衡,提高了编队续航能力;在80架UAV编队规模下,分级分布式领航-跟随者编队生成算法能够在0.3s内求解完成,相比于传统领航-跟随者编队算法,求解效率提高了约2.5倍;在100架UAV编队规模下,分级分布式领航-跟随者编队生成算法能够在0.4s内求解完成,而传统领航-跟随者编队由于位置误差的传递迭代已不能保持原有队形。The optimization design of the UAV formation information interaction topology is an important foundation to ensure safety and task execution efficiency of UAV formations.The generation algorithm of the UAV formation information interaction topology is currently limited to small-scale formation with a single optimization goal.To solve this problem,a hierarchical clustering structure is adopted to raise the information interaction topology level to meet the large-scale scene.A hierarchical distributed leader-follower formation information interaction topology generation algorithm based on the minimum cost arborescence is also proposed to improve the formation endurance and reduce the total communication cost of the formation.The simulation results are verified by OMNeT++.The experimental results show that the total communication cost of the hierarchical distributed leader-follower formation is significantly lower than that of the traditional leader-follower formation when considering the transfer iteration of position errors;the energy consumption of the network is more balanced and the endurance of the formation is improved through periodically updating the cluster head;at the formation scale of 80 UAVs,the hierarchical distributed leader-follower formation generation algorithm can be solved in 0.3 s,about 2.5 times that of traditional leader-follower formation algorithm;at the formation scale of 100 UAVs,the hierarchical distributed leader-follower formation generation algorithm can be solved within 0.4 s,while the traditional leader follower formation cannot maintain the original formation due to the transfer iteration of position errors.

关 键 词：无人机编队 信息交互拓扑 分级分簇 队形保持 领航-跟随者 

分 类 号：V243.1[航空宇航科学与技术—飞行器设计] N945.15[自然科学总论—系统科学]