Analysing Time Series of GNSS Residuals by Means of AR(I)MA Processes

Abstract

The classical least-squares (LS) algorithm is widely applied in processing data from Global Navigation Satellite Systems (GNSS). However, some limiting factors impacting the accuracy measures of unknown parameters such as temporal correlations of observational data are neglected in most GNSS processing software products. In order to study the temporal correlation characteristics of GNSS observations, this paper introduces autoregressive (integrated) moving average (AR(I)MA) processes to analyse residual time series resulting from the LS evaluation. Based on a representative data base the influences of various factors, like baseline length, multipath effects, observation weighting, atmospheric conditions on ARIMA identification are investigated. Additionally, different temporal correlation models, for example first-order AR processes, ARMA processes, and empirically determined analytical autocorrelation functions are compared with respect to model appropriateness and efficiency.