高超声速滑翔飞行器智能轨迹识别与预报方法  

Intelligent trajectory recognition and prediction method for hypersonic glider vehicle

作　　者：王惟 王晓芳[1] 林海[1] WANG Wei;WANG Xiaofang;LIN Hai(School of Aerospace Engineering,Beijing Institute of Technology,Beijing 100081,China)

机构地区：[1]北京理工大学宇航学院,北京100081

出　　处：《飞行力学》2024年第4期64-71,87,共9页Flight Dynamics

摘　　要：针对具有强机动能力且在滑翔段长时间飞行的高超声速滑翔飞行器轨迹预测问题,提出一种基于深度学习神经网络的目标机动模式识别及轨迹预报方法。建立了高超声速滑翔飞行器运动模型,设定了由纵向准平衡滑翔、跳跃滑翔和侧向蛇形机动、转弯机动构成的目标机动模型轨迹库。设计基于长短记忆网络(LSTM)的目标机动模式识别网络,通过训练后能够有效地通过目标的运动参数输出机动模式对应标签序列。针对每种机动模式,构建了基于LSTM的回归拟合外推轨迹预报方法。仿真结果表明,提出的智能机动模式识别分类与回归拟合轨迹预报相结合的方法,适应性较好,能够在较短的时间内以较高的精度对高超声速滑翔目标的长时间机动轨迹进行预报。A method of target maneuver pattern recognition and trajectory prediction based on deep learning neural network is proposed for hypersonic gliding vehicle with strong maneuverability and long flight duration.The motion model of hypersonic glider is established,and the trajectory library of target maneuver model is set up,which is composed of longitudinal quasi-balance glide,jump glide,lateral serpentine maneuver and turn maneuver.A maneuvering pattern recognition network based on long short-term memory(LSTM)network is designed,which can effectively output the cor-responding tag sequence through the motion parameters of the target after training.For each maneu-ver mode,the regression fitting extrapolation trajectory prediction method based on LSTM is con-structed.The simulation results show that the method of combining intelligent maneuvering pattern recognition classification with regression fitting trajectory prediction is proposed in this paper,it can predict the long-time maneuvering trajectory of hypersonic gliding target with high accuracy in a short time and has a good adaptability.

关 键 词：高超声速滑翔飞行器 长短记忆网络 深度学习 机动模式识别 轨迹预报 

分 类 号：E927[军事—军事装备学]