系统仿真学报 ›› 2022, Vol. 34 ›› Issue (1): 145-152.doi: 10.16182/j.issn1004731x.joss.20-0669

• 仿真模型/系统置信度评估技术 • 上一篇    下一篇

基于PCE方法的无人机轨迹规划与分析

曾思劼, 晏良, 段晓君*   

  1. 国防科技大学 文理学院,湖南 长沙 410073
  • 收稿日期:2020-09-07 修回日期:2020-11-28 出版日期:2022-01-18 发布日期:2022-01-14
  • 通讯作者: 段晓君(1976-),女,博士,教授,研究方向为复杂系统理论。E-mail:xjduan@nudt.edu.cn
  • 作者简介:曾思劼(1997-),女,硕士生,研究方向为不确定性量化。E-mail:zeng99_97@163.com
  • 基金资助:
    国家自然科学基金(11771450); 国家数值风洞工程(NNW2019ZT7-B23)

Planning and Analysis on UAV Trajectory Based on PCE Method

Zeng Sijie, Yan Liang, Duan Xiaojun*   

  1. College of Liberal Arts and Science, National University of Defense Technology, Changsha 410073, China
  • Received:2020-09-07 Revised:2020-11-28 Online:2022-01-18 Published:2022-01-14

摘要: 针对无人机轨迹规划过程中不确定因素的特点,提出了一种利用多项式混沌展开(polynomial chaos expansion, PCE),并获得基于人工势能法的模型最优参数的无人机轨迹规划方法。通过建立PCE代理模型,利用随机配置点法进行快速求解,规避了对已有的最优控制模型直接优化带来的计算资源不足的问题;通过Sobol灵敏度分析,减少不确定性参数在轨迹规划模型中的计算量;通过无人机轨迹规划案例,证明了方法的有效性。仿真结果表明:该方法与经典的人工势能模型相比,增加了不确定性分析,提高了模型计算效率。

关键词: 多项式混沌展开模型, 无人机, 轨迹规划, 人工势能法, 代价函数

Abstract: Focusing on the uncertainty in the UAV trajectory planning, combined with the artificial potential energy method, a UAV trajectory planning method based on polynomial chaos expansion (PCE), which can also efficiently obtain the optimal parameters of the model based on artificial potential field method is proposed. The PCE proxy model is established, and the stochastic collocation method is used to quickly solve the problem, so as to avoid the insufficient computing resources. Through the Sobol sensitivity analysis, the calculation overhead of the uncertainty parameters in the trajectory planning model is reduced. Cases of UAV trajectory planning prove the effectiveness of the method. Compared with the classic artificial potential field model, uncertainty analysis is added and the calculation efficiency of the model is improved.

Key words: PCE model, UAV, trajectory planning, artificial potential energy method, cost function

中图分类号: