摘要
针对一类变体飞行器自主变形决策问题,提出了一种基于深度确定性策略梯度(DDPG)算法的智能变形决策方法。首先,针对一种后掠角可连续变化的飞行器,通过计算流体力学方法获得飞行器的气动参数并分析其气动特性;然后,联合制导过程与DDPG算法,以获得最优气动特性和制导性能为目标,提出了一种变体飞行器智能变形决策算法;最后,仿真结果表明所提算法收敛效果好,相比于固定外形,可通过合适的变形决策指令在得到最优气动外形的同时获得更好的制导性能。
An intelligent deformation decision-making approach for morphing aircraft based on deep deterministic policy gradient(DDPG)is put forward to address the autonomous deformation decision-making issue of morphing aircraft.Firstly,for the sweptback aircraft,the aerodynamic parameters are acquired through the computational fluid dynamics method and its aerodynamic characteristics are analyzed.Subsequently,by integrating the optimal prediction midguidance and DDPG algorithms,an intelligent deformation decision-making method is proposed to obtain optimal aerodynamic characteristics and guidance performance.The simulation outcomes demonstrate that the proposed algorithm can converge effectively,and in comparison with the fixed shape,it can achieve superior guidance performance while obtaining the optimal aerodynamic shape via appropriate deformation decision commands.
作者
王青
刘华华
屈东扬
WANG Qing;LIU Huahua;QU Dongyang(International Innovation Institute of Beihang University,Hangzhou 311115,China;School of Automation Science and Electrical Engineering,Beihang University,Beijing 100191,China)
出处
《宇航学报》
EI
CAS
CSCD
北大核心
2024年第10期1560-1567,共8页
Journal of Astronautics
基金
国家自然科学基金(61873295)。
关键词
变体飞行器
自主变形决策
深度强化学习
深度确定性策略梯度算法
Morphing aircraft
Autonomous morphing decision
Deep reinforcement learning
Deep deterministic policy gradient
