一种基于遗传算法和BP神经网络的对流层延迟改正模型

摘要

Tropospheric delay is one of the main factors that affect the accuracy of Global Navigation Satellite System.Aiming at the low accuracy of the tropospheric delay correction model region established by using the global meteorological data,based on the genetic algorithm and BP neural network technique, a high-precision regional fusion model(GA-BPEGNOS model)is established on the basis of EGNOS model.By selecting 41 observation sites in North America from 2010 to 2014 and taking the zenith tropospheric delay data provided by the International GNSS Service as the true value,this paper compares the tropospheric zenith delay value calculated by EGNOS model and fusion model.T he result show s that the root mean square error of EGNOS model is 80.38 mm and the root mean square error of fusion model is 34.44 mm.Compared with the EGNOS model,the accuracy of the fusion model is improved by about 57%and good result is obtained.%对流层延迟是影响全球卫星导航系统定位精度的主要因素之一.针对全球气象数据建立的对流层延迟改正模型区域精度较低这一问题,文中基于遗传算法和BP神经网络技术,在 EGNOS模型基础上建立一个高精度的区域融合模型(GA-BPEGNOS模型).选取北美洲2010—2014年41个观测站点,以国际GNSS服务中心的对流层产品作为真值,分析比较 EGNOS 模型和融合模型的对流层天顶延迟.研究表明,EGNOS 模型的均方根误差为80.38 mm,融合模型的均方根误差为34.44 mm.与 EGNOS 模型相比,融合模型的精度提高约57%,取得满意效果.