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。
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.
作者
王宏伦
王延祥
刘一恒
WANG Honglun;WANG Yanxiang;LIU Yiheng(School of Automation Science and Electrical Engineering,Beihang University,Beijing 100191,China;The Science and Technology on Aircraft Control Laboratory,Beihang University,Beijing 100191,China;Shenyuan Honors College,Beihang University,Beijing 100191,China;Bejing Institute of Astronautical Systems Engineering,Beijing 100076,China)
出处
《航空学报》
EI
CAS
CSCD
北大核心
2023年第20期1-17,共17页
Acta Aeronautica et Astronautica Sinica
基金
国家自然科学基金(62173022,61673042)。
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
aerial recovery
trajectory optimization
trajectory mapping
cable-drogue-UAV assembly
Bidirectional Long Short-Term Memory(BiLSTM)
neural network
