Performance Assessments of Correction Models in GNSS Network-based RTK Positioning

Global Navigation Satellite System (GNSS) real-time kinematic (RTK) positioning is applied to determine users' positions at the accuracy of a centimetre level providing that carrier-phase ambiguities can be resolved to their integer values. Differential atmospheric errors can only be neglected for distances at a certain extent. The network-based RTK (NRTK) resolves differential atmospheric errors and orbit errors from GNSS observations of a network of reference stations before sending the computed corrections to requested users. These coefficients of correction models are explicitly determined based upon network observations on common ambiguity level and coefficient of correction models; namely, Virtual Reference Station (VRS) observations, Flächen-Korrektur Parameters (FKP) standing for area correction parameters, Master-Auxiliary Concept (MAC) and Individualised-MAC (I-MAC). This study aims to evaluate performances of four main correction models from a reference station network in Thailand principally structured by the Royal Thai Survey Department. In this experiment, an identical antenna is used whiles four receivers are observed in order to determine and evaluate performances concurrently. Obstacles are constantly a challenge in GNSS observations; hence, it is included in this positioning estimations. It is found that, in the observed area without obstacles, percentage of fixed solutions is higher than 95% when VRS and I-MAC correction models are applied and greater than 75% with MAC and FKP. With obstacles, success rate percentage of fixed ambiguities using FKP is better than 50% while other models are less than 50%. The horizontal positioning accuracy is better than 5 centimetres in unobstructed conditions.