摘要
针对强时敏场景下飞行器初始段飞行参数快速解算问题,提出了一种基于神经网络预测的参数快速解算方法。建立了飞行器动力学模型并设计了总体参数及程序段飞行策略,基于该飞行策略选取了适当的待设计参数,通过遍历仿真编制了参数表;通过参数表匹配建立数据集,并设计双隐藏层参数解算网络,该网络以航程和高度为输入,可以解算出初始段飞行参数。在典型场景下进行了飞行轨迹仿真。结果表明,该方法与传统方法相比解算速度大幅提升,且能够满足修正精度要求。
Aiming at the problem of fast calculation of flight parameters in the initial phase of aircraft under strong time-sensitive scenarios,a fast parameter calculation method based on neural network prediction is proposed.The dynamics model of the aircraft is established and the overall parameters and program segment flight strategy are designed.Based on the flight strategy,the appropriate parameters to be designed are selected and the parameter table is compiled by traversal simulation.The data set is established by matching the parameter table,and the double-hidden layer parameter calculation network is designed.The network can solve to calculate the initial flight parameters by taking the distance and altitude as the input.Flight trajectory simulation is carried out under typical scenarios.The results show that the proposed method is faster than the traditional method and can meet the requirement of correction accuracy.
作者
徐世杰
肖清
王淼
Xu Shijie;Xiao Qing;Wang Miao(Beijing Institute of Space Long March Vehicle,Beijing 100076,China;Beijing Institute of Space System Engineering,Beijing 100037,China;School of Astronautics,Harbin Institute of Technology,Harbin 150001,China)
出处
《战术导弹技术》
北大核心
2025年第2期104-111,共8页
Tactical Missile Technology
关键词
神经网络
轨迹设计
飞行参数
强时敏场景
飞行器初始段
快速解算
双隐藏层
neural network
trajectory design
flight parameters
strong time-sensitive scenario
the initial phase of aircraft
fast calculation
double hidden layer
