感-通-物多目标融合应急无人机路径规划方法
Integrated perception-communication-logistics multi-objective oriented path planning for emergency UAVs
许云鹏 1谢雅琪 1于然 2侯鲁洋 1王凯亮 2徐连明3
作者信息
- 1. 北京邮电大学计算机学院(国家示范性软件学院),北京 100876
- 2. 国网冀北电力有限公司信息通信分公司,北京 100053
- 3. 北京邮电大学电子工程学院,北京 100876
- 折叠
摘要
为了完成多无人机应急救援场景下救灾点的需求感知(感)、数据收集(通)和物资投放(物)任务,提出了在考虑无人机能耗约束下,感-通-物多目标融合的两阶段的应急无人机路径规划求解框架.第一阶段提出基于时序图卷积网络的救灾点人数预测模型,并量化救灾点物资和通信需求;第二阶段提出基于贪心和禁忌搜索的多无人机路径规划算法,通过交替优化救灾点划分和单无人机路径规划来求解原优化问题.仿真结果表明,该算法在总服务收益上优于传统的无预测多无人机路径规划算法.
Abstract
In order to complete the tasks of demand perception(perception),data collection(communication),and mate-rial delivery(logistics)at disaster relief sites in multi-UAV emergency scenarios,a two-stage solution framework was proposed for multi-UAV path planning that integrated perception,communication,and logistics objective considering UAV energy consumption constraint.In the first stage,a temporal graph convolution networks-based model was intro-duced to predict the number of personnel at the relief sites to quantify its supply and communication needs.In the second stage,a multi-UAV path planning algorithm based on the greedy and tabu search was proposed to solve the optimization problem through iteratively optimizing the relief point clustering and the path planning of individual UAV.The simula-tion results demonstrate that the proposed algorithm is superior to the traditional prediction-free multi-UAV path plan-ning algorithm in terms of the total service revenue.
关键词
无人机/路径规划/时序图卷积网络/禁忌搜索Key words
UAV/path planning/temporal graph convolution network/tabu search引用本文复制引用
基金项目
国家自然科学基金(62171054)
国家自然科学基金(62101045)
中央高校基本科研业务费专项(24820232023YQTD01)
中央高校基本科研业务费专项(2023RC96)
"双一流"建设学科交叉团队基金(2023SYLTD06)
国家电网冀北电力有限公司科技项目(52018E230001)
北京市自然科学基金(L222041)
出版年
2024
国家科技期刊平台
NSTL
维普
万方数据