摘要
全球定位系统(global positioning system,GPS)卫星钟差是影响精密单点定位(precise point positioning,PPP)的关键性因素。线性模型、二次多项式模型、灰色模型和国际GNSS服务组织(International Global Navigation Satellite System Service, IGS)超快速星历(IGS ultra-rapid,IGU)预报模型对于不同类型的卫星钟钟差具有相应的适用性。鉴于常用的单一预报模型存在不足,提出了一种针对超快速卫星实时钟差预报的模型——遗传算法(genetic algorithm,GA)优化的反向传播(back propagation, BP)神经网络组合模型(GA-BP模型)。选用灰色模型预测数据作为原始数据,分别采用BP神经网络模型以及GA-BP模型进行卫星钟差预报。经数据分析显示,GA-BP模型在不同时段的预报精度都有一定程度的提高。
GPS satellite clock error is a key factor affecting precise point positioning(PPP). The linear model, quadratic polynomial model, grey model and IGU prediction model are suitable for different types of satellite clock error. In view of the shortcomings of the common single forecasting model, this paper presents a model of real-time clock error prediction of the ultra-rapid satellites—BP Neural Network Optimized by Genetic Algorithm(GA-BP model). With the grey model prediction data as the raw data, the BP neural network model and GA-BP model are used to predict the satellite clock error. The data analysis shows that the forecast accuracy of GA-BP model is improved to some extent in different time periods.
作者
刘永辉
任彪
徐景田
LIU Yonghui;REN Biao;XU Jingian(Guangzhou Institute of Building Science Co.,Ltd.Gungho 510440,China;School of Geography and Information Engineering,China University of Geosciences(Wuhan),Wuhan 430074,China;Sichuan Electric Power Design and Consulting Co.,Ltd.,Chengdu 610041,China)
出处
《测绘地理信息》
2020年第6期35-39,共5页
Journal of Geomatics
基金
国家自然科学基金(41874037)。
关键词
超快速星历
实时全球定位系统
卫星钟差预报
反向传播神经网络
遗传算法
ultra-rapid ephemeris
real-time global positioning system satellite
clock error prediction
back propagation neural network
genetic algorithm
