弹道学报2025，Vol.37Issue(2)：60-68,9.DOI:10.12115/ddxb.2024.04008
基于机器学习的制导炮弹控制方法研究
Research on Guided Projectile Control Method Based on Machine Learning
曹文洁 1周文涛 1常思江1
作者信息
- 1. 南京理工大学 能源与动力工程学院,江苏 南京 210094
- 折叠
摘要
Abstract
The traditional impact point prediction guidance methods require a lot of iterations in the solution process,and the real-time processing ability of information is poor.The impact point prediction model based on machine learning method can reduce the prediction time and improve the prediction accuracy.However,the existing impact point prediction model is mostly trained by the uncontrolled flight data of the projectile,without considering the influence of control force and torque,resulting in a remarkable difference from the actual flight state of the guided projectile.In this paper,a guidance and control method based on machine learning was proposed.The flight state parameters of the projectile in the controlled state were used as the training set,and the controlled landing point prediction model was established.Combined with the proportional-differential control law,the guidance and control process based on this landing point prediction model was constructed.The simulation results show that the impact point prediction model trained by BP neural network can accurately predict the impact point,and the proposed guidance and control method can achieve high-precision guidance and control.The relative error of the range direction of the hit point is about 0.01%,and the relative error of the sideslip direction is about 3.1%.The feasibility and effectiveness of the designed guidance and control method are verified,which can provide reference for the in-depth application of machine learning method in the field of controlled projectile technology.关键词
机器学习/制导炮弹/落点预测/控制方法Key words
machine learning/guided projectile/impact point prediction/control method分类
武器工业引用本文复制引用
曹文洁,周文涛,常思江..基于机器学习的制导炮弹控制方法研究[J].弹道学报,2025,37(2):60-68,9.
