Variance-Covariance Modeling of Atmospheric Errors for Satellite-Based Network Positioning

Abstract

: This paper investigates the problem of variance-covariance modeling of atmospheric errors for the purposes of improving the accuracy of positioning using global navigation satellite systems, as well as improving estimates of the accuracy of such positioning. A method of modeling all variances and cross-correlations among observations made in a network of positioning receivers is presented. This is done by generating a theoretical model of covariance behavior based on the physical nature of tropospheric and ionospheric errors, respectively, and then using data observed at a network of reference receivers to derive key parameters of the theoretical model. The applicability of this method is demonstrated for a network of 10 receivers with a total extent of 250 km. It is shown that proper modeling of covariances improves positioning accuracy an average of 22 percent.