航空学报 > 2023, Vol. 44 Issue (20): 628775-628775   doi: 10.7527/S1000-6893.2023.28775

基于轨迹映射的无人机拖曳式空中回收轨迹优化

王宏伦1,2(), 王延祥1,2,3, 刘一恒1,2,4   

  1. 1.北京航空航天大学 自动化科学与电气工程学院,北京  100191
    2.北京航空航天大学 飞行器控制一体化技术重点实验室,北京  100191
    3.北京航空航天大学 沈元学院,北京  100191
    4.北京宇航系统工程研究所,北京  100076
  • 收稿日期:2023-03-30 修回日期:2023-04-27 接受日期:2023-05-30 出版日期:2023-06-07 发布日期:2023-06-05
  • 通讯作者: 王宏伦 E-mail:wang_hl_12@126.com
  • 基金资助:
    国家自然科学基金(62173022)

Recovery trajectory optimization for UAV towed aerial recovery based on trajectory mapping

Honglun WANG1,2(), Yanxiang WANG1,2,3, Yiheng LIU1,2,4   

  1. 1.School of Automation Science and Electrical Engineering,Beihang University,Beijing  100191,China
    2.The Science and Technology on Aircraft Control Laboratory,Beihang University,Beijing  100191,China
    3.Shenyuan Honors College,Beihang University,Beijing  100191,China
    4.Beijing Institute of Astronautical Systems Engineering,Beijing  100076,China
  • Received:2023-03-30 Revised:2023-04-27 Accepted:2023-05-30 Online:2023-06-07 Published:2023-06-05
  • Contact: Honglun WANG E-mail:wang_hl_12@126.com
  • Supported by:
    National Natural Science Foundation of China(62173022)

摘要:

针对无人机拖曳式空中回收过程中的轨迹优化问题,提出一种基于轨迹映射的无人机回收轨迹在线优化方法。首先,建立包括缆绳-浮标-无人机组合体的运动模型、机翼折叠模型在内的空中回收系统模型。随后,提出轨迹映射的思想,利用双向长短期记忆(BiLSTM)神经网络建立回收系统中回收指令和回收轨迹之间的精确映射关系。然后,利用轨迹映射网络实时预测不同指令下的回收轨迹,并根据计算的预测轨迹代价利用粒子群优化(PSO)算法优化得到最佳回收指令。最后,仿真结果表明:所提的轨迹映射网络具有较高的预测精度和计算速度,所提的方法可以优化出使无人机稳定快速回收的轨迹。

关键词: 空中回收, 轨迹优化, 轨迹映射, 缆绳-浮标-无人机组合体, 双向长短期记忆, 神经网络

Abstract:

For the problem of trajectory optimization in the process of Unmanned Aerial Vehicle (UAV) towed aerial recovery, an optimization method of UAV recovery trajectory is proposed based on trajectory mapping. First, an aerial recovery system model, including the cable-drogue-UAV assembly model and the wing fold model, is established. Second, the idea of trajectory mapping is put forward, and the accurate mapping relationship between the recovery instruction and the recovery trajectory in the recovery system is established by using the Bidirectional Long Short-Term Memory (BiLSTM) neural network. Third, the trajectory mapping network is utilized to predict the real recovery trajectory under different instructions in real time, and the Particle Swarm Optimization (PSO) algorithm is used to optimize the optimal recovery instruction according to the calculated predicted trajectory cost. Finally, the simulation results show that the proposed trajectory mapping network has high prediction accuracy and calculation speed, and the proposed optimization method can achieve stable and rapid recovery of UAV.

Key words: aerial recovery, trajectory optimization, trajectory mapping, cable-drogue-UAV assembly, Bidirectional Long Short-Term Memory (BiLSTM), neural network

中图分类号: