Do GNSS parameters always benefit from integer ambiguity resolution? A PPP-RTK network scenario

Abstract

All rights reserved. In this contribution we study the impact of integer ambiguity resolution (IAR) on the precision of GNSS estimable parameters. In doing so, a multivariate formulation of the GNSS undifferenced observation equations of a PPP- RTK network is presented, where appropriate one-to-one transformations to the parameters ease our analysis. Functions of the estimable slant ionospheric delays are shown to have distinct responses to IAR, ranging from zero (of a satellite-/receiver-averaged type) to very significant (of a double-differenced type). While the satellite-/receiver- Averaged components of the estimable clocks and phase biases on a particular frequency remain unaffected after IAR, the precision of their single-differenced (SD) counterparts experiences improvement. In particular, the precision improvement gets larger as the geometry of the model gets weaker. This is, however, not the case with the wide-lane satellite phase biases, since their precision is almost insensitive to the geometry of the model. Supported by numerical results, our analytical expressions show that the precision improvement of the SD satellite clocks and phase biases increases as the number of stations increases. We also highlight the zero-correlation property between the double-differenced ambiguities and the code biases that get estimable on the third frequency and beyond.