Instantaneous PPP convergence with a Decoupled Clock Model

In recent decades, Precise Point Positioning (PPP) has become a highly effective Global Navigation Satellite System (GNSS) positioning method and promising alternative to well-established relative positioning techniques. PPP is characterised by the use of precise satellite data (such as satellite orbits, clocks, and biases) and the accurate modelling of various error sources. PPP allows us to achieve position accuracies at the centimetre or even millimetre level. However, its widespread use has been limited due to significant convergence times. Among other approaches, innovative PPP models and PPP with integer ambiguity resolution (PPP-AR) have proven to be the most effective in reducing this time.

The decoupled clock model (DCM) provides an elegant way to perform PPP-AR in an uncombined approach for any number of frequencies. The idea is to estimate separate receiver clock errors for code and phase observations, resulting in a receiver code clock error and a receiver phase clock error. Additionally, the unknowns are reparametrized in such a way that the integer property of the phase ambiguities is conserved, with a datum satellite being used to set the phase ambiguities' datum. Unlike other PPP-AR approaches, direct differencing with a reference satellite is not necessary. PPP-AR can therefore be performed in a straightforward manner.

In this contribution, we discuss the DCM and its key characteristics. We present PPP results achieved with the uncombined DCM and GPS, GLONASS, Galileo, and BeiDou observations on three frequencies at thirty-second and one-second intervals. We then evaluate convergence behaviour, coordinate accuracy, ZTD estimation, and ambiguity fixing rates. The PPP investigations were conducted using the open-source software raPPPid. Our findings show that instantaneous PPP convergence to centimetre-level accuracy can be achieved within two to three measurement epochs.