| 引用本文: | 张晓杰，郑纪彬，苏涛，刘宏伟，高琦. 无人机集群协同搜索跟踪任务规划方法[J]. 雷达科学与技术, 2022, 20(5): 480-491.[点击复制] |
| ZHANG Xiaojie, ZHENG Jibin, SU Tao, LIU Hongwei, GAO Qi. Mission Planning for Cooperative Search-Track of UAV Swarms[J]. Radar Science and Technology, 2022, 20(5): 480-491.[点击复制] |
|
| 摘要: |
| 无人机集群在目标搜索、定位和跟踪等方面具有巨大的应用潜力，有效的任务规划方案能极大提高无人机集群执行任务的效率。在不确定的动态环境中，任务规划方案需要适应环境的变化，对任务规划的求解效率提出了较高的要求。针对动态环境下的无人机集群协同搜索跟踪任务规划问题，本文将其建模为动态多约束多目标优化问题（DMCMOPs），并提出了基于动态自适应惩罚的动态约束双档案进化算法（DCTAEA），其在收敛性种群更新中引入自适应惩罚函数机制，整合不可行个体的目标函数值和违反约束的惩罚值获得修正的目标函数值，实现有价值不可行解的利用，促使种群进入可行区域并向帕累托前沿面收敛，极大促进了种群的收敛。仿真结果证明,与第二代非支配排序遗传算法（NSGA-II）的动态版本、基于分解的多目标进化算法（MOEA/D）、约束双档案进化算法（CTAEA）和动态双档案进化算法（DTAEA）相比，本文所提算法有效性较显著。 |
| 关键词: 目标函数改变 动态约束处理 多目标优化 无人机 任务规划 |
| DOI：DOI:10.3969/j.issn.1672-2337.2022.05.002 |
| 分类号:TN972;V279 |
| 基金项目:国家自然科学基金（No.61971336, 61601341, 61771367） |
|
| Mission Planning for Cooperative Search-Track of UAV Swarms |
|
ZHANG Xiaojie, ZHENG Jibin, SU Tao, LIU Hongwei, GAO Qi
|
|
1. National Key Lab of Radar Signal Processing, Xidian University, Xi’an 710071, China;2. Shaanxi Changling Electronic Technology Co Ltd, Baoji 721006, China
|
| Abstract: |
| Target search, localization and tracking by unmanned aerial vehicle (UAV) swarms have attracted much attention recently in research and industrial applications, and effective mission planning method can greatly improve the efficiency of executing missions by UAV swarms. The optimal mission planning scheme requires being solved immediately once the environment changes. Consequently, the demand for the efficiency of working out a mission planning solution has increased dramatically. The mission planning problem in the dyna-mic environment is modeled as a dynamic multi-constraint and multi-objective optimization problem (DMCMOP), and we propose the dynamic constrained two-archive evolutionary algorithm (DCTAEA). The convergence archive (CA) and the diversity archive (DA) are adaptively reconstructed when the environment changes, and we introduce the dynamic adaptive penalty mechanism into the CA updating mechanism, which utilizes valuable infeasible solutions and promotes population convergence. Simulation results demonstrate that the proposed algorithm has effectiveness and superiority compared with the non-dominated sorting genetic algorithm II（NSGA-II), multi-objective evolutionary algorithm based on decomposition (MOEA/D), constrained two-archive evolutionary algorithm (CTAEA) and the dynamic two-archive evolutionary algorithm (DTAEA). |
| Key words: changing objectives dynamic constraint handling multi-objective optimization unmanned aerial vehicle mission planning |
