基于神经网络的卫星钟差预报及误差控制  

Satellite Clock Prediction and Error Control Based on Neural Network

作　　者：朱陵凤 李超 刘利 李晓杰 郭睿 

机构地区：[1]北京卫星导航中心,北京100094

出　　处：《全球定位系统》2016年第3期68-72,共5页Gnss World of China

基　　金：国家高技术研究发展计划(批准号:2014AA123102)

摘　　要：针对卫星钟差预报误差随时间不断变大问题,提出了基于BP神经网络的卫星钟差预报误差控制方法。该方法利用神经网络强大的非线性映射能力,对历史时段钟差拟合残差及当前时段钟差预报误差进行训练,得到拟合残差与预报误差的映射关系,据此根据当前时段的拟合残差预测未来时段的预报误差,将其补偿到未来时段钟差预报钟差结果中去,以实现未来时段卫星钟差预报误差的有效控制。为验证该方法的可行性和有效性,利用实测的COMPASS导航卫星钟差数据进行钟差预报精度分析,结果表明,神经网络可有效控制卫星钟差预报误差、提高卫星钟差预报精度。Aiming at improving the predicting precise of satellite clock error,method of controlling satellite clock error will be mentioned in this paper which is also based on BP neural network theory.Considering of the nonlinearity mapping ability of neural network,this method compares residual errors,which is created according to outdated statistics,with current errors to find the potential relation between these 2errors.After the comparison,this paper gets mapping relation between residual errors and current errors.This paper assumes this relation also exist between current errors and the future errors.To prove this assumption is workable,this paper analyzes the predicting precise of satellite clock error through the statistics of COMPASS satellite clock error.The analysis shows this method is feasible and effective,which means neural network can improve the predicting precise of satellite clock error.

关 键 词：神经网络 卫星钟差 COMPASS 预报 

分 类 号：P228.4[天文地球—大地测量学与测量工程]