有控弹箭气动参数辨识技术
Research on aerodynamic parameter identification technology of controlled projectile
康其庄 1王康健 2易文俊 1段耀泽 2夏悠然1
作者信息
- 1. 南京理工大学 瞬态物理重点实验室,南京 210000
- 2. 空军工程大学 航空工程学院,西安 710000
- 折叠
摘要
快速准确获取气动参数是精确制导的必要前提.针对受限于模型构建精度,传统气动参数辨识方法对受力复杂的有控弹箭气动参数辨识困难、精度不足等问题,引入Elman递归神经网络,利用Elman神经网络强大的延时记忆和非线性拟合能力辨识气动参数,探究Elman神经网络应用于有控弹箭气动参数辨识的可行性,并与BP神经网络辨识结果进行了对比.仿真结果表明,Elman神经网络能较好地辨识出滑翔飞行阶段的气动参数,且辨识精度要高于BP神经网络.
Abstract
Fast and accurate acquisition of aerodynamic parameters is a necessary prerequisite for precision guidance.Limited by the accuracy of model construction,the traditional aerodynamic parameter identification method is not accurate enough to identify the controlled projectile with complex forces.Aiming at the difficulty of aerodynamic parameter identification of controlled projectile,this paper introduces Elman recurrent neural network,uses Elman neural network's powerful delay memory and nonlinear fitting ability to identify aerodynamic parameters,explores the feasibility of applying Elman neural network to aerodynamic parameter identification of controlled projectile,and compares the identification results with BP neural network.The simulation results show that Elman neural network can identify the aerodynamic parameters of gliding flight stage well,and the identification accuracy is higher than BP neural network.
关键词
有控弹箭/参数辨识/Elman神经网络/数据插值Key words
controlled projectile arrow/parameter identification/Elman neural network/data interpolation/Matlab引用本文复制引用
出版年
2024
NETL
NSTL
万方数据