基于Takagi-Sugeno模糊神经网络模型的卫星钟差预报方法  

The Satellite Clock Bias Prediction Method Based on Takagi-Sugeno Fuzzy Neural Network

作　　者：蔡成林[1,2] 于洪刚[1,2] 韦照川[1,2] 潘军道 

机构地区：[1]桂林电子科技大学信息与通信学院,桂林541004 [2]广西精密导航技术与应用重点实验室,桂林541004

出　　处：《天文学报》2017年第3期111-124,共14页Acta Astronomica Sinica

基　　金：国家自然科学基金项目(61263028);桂林电子科技大学研究生教育创新计划资助项目(2016YJCX14)资助

摘　　要：卫星钟差预报精度的不断提升是精密导航的关键问题.为了进一步提高钟差的预报精度和更好地反映钟差的变化特性,提出一种基于Takagi-Sugeno模糊神经网络(Fuzzy Neural Network,FNN)的钟差预报方法.该方法首先根据钟差数据的特点对钟差进行预处理,然后以预处理后的数据建立一种高精度预报钟差的Takagi-Sugeno模糊神经网络算法.采用IGS(International Global Navigation Satellite System Service)不同采样间隔的精密钟差数据进行了短期预报试验,并与ARIMA(Auto-Regressive Integrated Moving Average)模型、GM(1,1)模型及QP(Quadratic Polynomial)模型进行了对比试验,分析结果表明:对不同类型原子钟,该方法用于钟差短期预报是可行的、有效的,其获得的卫星钟差预报结果明显优于常规方法.The continuous improvement of the prediction accuracy of Satellite Clock Bias （SCB） is the key problem of precision navigation. In order to improve the precision of SCB prediction and better reflect the change characteristics of SCB, this paper proposes an SCB prediction method based on the Takagi-Sugeno fuzzy neural network. Firstly, the SCB values are pre-treated based on their characteristics. Then, an accurate Takagi-Sugeno fuzzy neural network model is established based on the preprocessed data to predict SCB. This paper uses the precise SCB data with different sampling intervals provided by IGS （International Global Navigation Satellite System Service） to realize the short-time prediction experiment, and the results are compared with the ARIMA （Auto-Regressive Integrated Moving Average） model, GM（1, 1） model, and the quadratic polynomial model. The results show that the Takagi-Sugeno fuzzy neural network model is feasible and effective for the SCB short-time prediction experiment, and performs well for different types of clocks. The prediction results for the proposed method are better than the conventional methods obviously.

关 键 词：天体测量学 时间 方法 数据分析 方法 TAKAGI-SUGENO 

分 类 号：P127[天文地球—天体测量]