一种神经网络预测模型误差的地磁导航方法  

A Geomagnetic Navigation Method Based on Neural Network Prediction Model Error

作　　者：张涛 张文博 高东 陈晨 郑建华 ZHANG Tao;ZHANG Wenbo;GAO Dong;CHEN Chen;ZHENG Jianhua(School of Mechanical and Electrical Engineering,Henan University of Technology,Zhengzhou 450001,China;National Space Science Center,CAS,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 101407,China)

机构地区：[1]河南工业大学机电工程学院,郑州450001 [2]中国科学院国家空间科学中心,北京100190 [3]中国科学院大学,北京101407

出　　处：《航天控制》2024年第1期37-42,共6页Aerospace Control

基　　金：实验室基金项目(CXJJ-22S017)。

摘　　要：针对地磁场模型更新慢和误差大而引起的地磁导航精度难以提高的问题,提出一种基于人工神经网络的地磁模型误差预测方法,通过建立地磁场强度矢量要素与地理特征及时间信息之间的映射关系,结合在轨卫星的磁场测量数据,实现对地磁场模型误差的估计和预测;此外,提出了一种神经网络与滤波器组合的地磁导航方法。为了验证方法的精度和有效性,利用卫星的实测信息进行了仿真分析,结果表明:同现有文献中先进的导航方法相比,位置和速度精度由4.15 km和4.38 m/s提高到了1.34 km和1.47 m/s,明显改善了地磁导航精度和效果。A geomagnetic model error prediction method based on artificial neural networks is proposed in response to the difficulty of improving geomagnetic navigation precision because of the slow update and low accuracy of the geomagnetic field model.The mapping relationship is combined with the magnetic field measurement data of on⁃orbit satellites to estimate and predict the geomagnetic field model error by establis⁃hing the relationship among geomagnetic field intensity vector elements,geographical features,and time in⁃formation.In addition,a geomagnetic navigation method combining neural network with filter is proposed.In order to verify the accuracy and effectiveness of this method,simulation verification is conducted by using satellite measured information.The results show that compared with the advanced filter method in recent years’literatures,the position and velocity accuracy can be improved from 4.15 km,4.38 m/s to 1.34 km,1.47 m/s that significantly improves the precision and effect of geomagnetic navigation.

关 键 词：地磁导航 地磁模型误差预测 BP神经网络 地磁滤波 

分 类 号：V249.3[航空宇航科学与技术—飞行器设计]