A sequential network approach for estimating GPS satellite phase biases at the PPP-AR producer-side

Ambiguity resolution (AR) in precise point positioning (PPP) requires precise satellite orbit, clocks, and phase biases corrections. Satellite phase biases are fractional hardware corrections which help to retrieve the un-differenced integer carrier phase ambiguities. Satellite corrections can be obtained from the international GNSS service (IGS) or estimated by correction providers called producer-side. We introduce a new PPP-AR observation model and a new sequential network algorithm (SNA) to estimate satellite phase biases. The new model is fully compatible with standard IGS satellite correction products, and it takes advantage of currently available IGS global ionosphere maps to improve the stability of corrections estimation. Furthermore, the proposed model is full-rank per-frequency and per-site and this method simplifies the integration of any additional frequency or site observables in the system of equations. The per-site satellite phase biases method allows users to customize their network solution. In many cases, users only have to estimate the phase biases of a few satellites estimated by few stations to resolve ambiguities of their observed satellites. The novel two-step algorithm provides a good balance between the computational burden, the computer memory load, the efficiency of handling parameters, and the precise estimation of correction parameters. The proposed PPP-AR model and the SNA performance is then validated by estimating satellite phase biases with 1 year of GPS data from a sub-network of IGS stations. A rigorous a posteriori statistical test is performed using data from an independent GPS network. As a result, the precision of WL and L1 ambiguities was improved significantly with the confidence level of P > 99.99% by applying the estimated phase bias corrections to phase observables.