On the intricacies of multi-epoch filtered PPP-RTK corrections and their impact on GNSS ambiguity resolution

Precise point positioning real-time kinematic (PPP-RTK), by capitalizing on its state-space representation (SSR) and its associated flexibility, is naturally emerging as one of the prevalent Global Navigation Satellite System (GNSS) techniques for high-precision positioning. The determination of unbiased ambiguity-resolved positional parameters becomes possible once single-receiver users get access to SSR corrections. However, the fact that such SSR corrections are often estimated in a recursive manner, based on a set of assumptions and a singularity-basis (S-basis) choice made in an arbitrary fashion by the correction provider (i.e., a GNSS network), may lead to serious pitfalls which the PPP-RTK user should be aware of when interpreting the delivered corrections. In this contribution, we will present the intricacies inherent in multi-epoch filtered PPP-RTK corrections and address the consequence of the corrections’ dependency on the provider’s S-basis. Through illustrative examples, it is shown how one can be misled by merely analyzing the estimable satellite clock solutions’ temporal characteristics, and how the distributional properties of the satellite phase bias solutions can be affected in case only their fractional part is delivered, contrary to the users’ usual expectation of being equipped with Gaussian-distributed phase biases. Next to this analysis, the important roles played by the correction latency and time correlation are addressed in both ambiguity-resolved positioning and the associated ambiguity-float and -fixed confidence information reported by the user estimation process.