卫星观测信号盲源分离算法  

Blind source separation method of satellite observation signals

作　　者：刘国光 LIU Guoguang(Guangzhou Nanfang Surveying and Mapping Technology Co.,Ltd.,Guangzhou,Guangdong,510665,China)

机构地区：[1]广州南方测绘科技股份有限公司,广东广州510665

出　　处：《测绘技术装备》2022年第1期50-56,共7页Geomatics Technology and Equipment

摘　　要：为研究GNSS观测数据中多路径误差分离方法,本文从盲源分离的角度出发,提出了一种基于广义特征值的全新盲源分离算法,并通过仿真实验进行算法验证。此外,还在仿真数据的基础上增加多路径误差组成多路源信号,通过广义特征值进行源信号分离,仍然可以得到很好的效果。采用相似系数、二次残差评价指标对分离效果进行评价,结果表明,与ICA和MAX-SNR盲源分离算法的效果相比,广义特征值算法相似系数最高,二次残差最低。In order to study the multipath error separation method in GNSS observation data,a new blind source separation algorithm based on generalized eigenvalue is proposed from the perspective of blind source separation,and the algorithm is verified by simulation experiments.Based on the simulation data,the multipath error is added to form the multipath source signal,and the source signal separation through generalized eigenvalues can still get good results.The separation effect is evaluated by similarity coefficient and quadratic residual evaluation index,and the results showed that compared with ICA and MAX SNR blind source separation algorithms,the generalized eigenvalue algorithm has the highest similarity coefficient and the lowest quadratic residual.

关 键 词：GNSS 观测信号 广义特征值 盲源分离算法 多路径误差 

分 类 号：P228[天文地球—大地测量学与测量工程]