Integer Ambiguity-Resolved Detection and Identification of GNSS Model Biases

GNSS model validation constitutes an important part of any GNSS data processing scheme. Statistical tests are then employed to test for the occurrence of model misspecifications, such as pseudorange outliers, carrier-phase slips, neglected atmospheric delays, and constellation faults. The very precise carrier-phase measurements are generally used to achieve the highest possible accuracy in GNSS parameter estimation. However, since the interferometric carrier-phase observables are integer ambiguous, the carrier-phase GNSS models are of the mixed-integer type, consisting of real-valued and integer-valued parameters. Although the estimation theory for mixed-integer models is reasonably well developed, this is not yet the case for the statistical validation of such models. It is the goal of this contribution to present and evaluate statistical detection and identification tests that exploit the integerness of the carrier-phase ambiguities, thereby significantly improving the power over the customary employed ambiguity-float tests. Various applications of the tests are provided, together with graphical and numerical analyses of their performance.