Study on Stochastic Model of Baseline Estimation of GPS Reference Station Network

Abstract

Baseline estimation is one of the most important links in the data processing of GPS reference station network. Exact definition of functional model and stochastic model of baseline estimation must be required to achieve high precise baseline solutions. The effects on precision of GPS long baseline estimation of three stochastic models are analyzed and compared by computation experiments using observation data of GPS reference station network. Calculation results show that using refined stochastic model can reduce convergence time of baseline solution. For baselines about 100km long in GPS reference station network, baseline precision of float and fixed solutions can be improved about 0.10m and 3mm respectively by satellite elevations compared with standard stochastic model using 10~40 minutes’ observation data and baseline precision of float and fixed solutions can be improved about 0.15m and 5mm respectively by estimated stochastic model based on theory of stationary stochastic process compared with standard stochastic model.