神经网络在卫星钟差短期预报中的应用研究  

Application of neural network in satellite clock bias short-term prediction

作　　者：郭承军[1] 滕云龙[1] 

机构地区：[1]电子科技大学电子科学技术研究院,成都610054

出　　处：《测绘科学》2011年第4期198-200,共3页Science of Surveying and Mapping

基　　金：航空科学基金(20090580013)

摘　　要：本文针对卫星钟差的特点,提出了基于神经网络的卫星钟差短期预报模型,给出了基于径向基函数(RBF)网络进行卫星钟差预测的基本思想、预测模型和实施步骤,并对比分析了神经网络模型与灰色系统理论模型的区别。为验证本文提出的预报模型的可行性和有效性,利用GPS卫星钟差数据进行钟差预报精度分析,并与灰色系统模型进行对比分析。仿真结果显示,该模型具有较好的预测精度,可为实时GPS动态精密单点定位提供较高精度的卫星钟差。Aiming at the characteristic of satellite clock bias, a prediction model based on artificial neural networks was presented for clock bias short-time prediction in the paper. The basic ideas, prediction models and steps of clock bias forecasting based on radial basis function （RBF） network were discussed respectively. The differences between neural network and other statistical prediction method, such as gray predicting model, were compared respectively. To validate the feasibility and validity of the proposed method, it made a careful precision analysis for satellite clock bias prediction with the performance parameters of GPS satellite clock, and made comparison and analysis with Grey system model and neural network model. The results of simulation showed that the prediction preci- sion of the novel four-stage mode] based on wavelet analysis and artificial neural networks was more better, could afford high precise satellite clock bias prediction for real-time GPS precise point positioning.

关 键 词：径向基神经网络 卫星钟差序列 预报 滑动窗 正交最小二乘 

分 类 号：P228.41[天文地球—大地测量学与测量工程]