Study on integrated cycle slip handling using GPS/Galileo combined observations

Quality control theory in terms of testing residuals has been proven to be an effective approach to deal with cycle slips. The integrated processing of multi-GNSS observations can further improve the reliability of cycle slip handling when using the residual snooping method. In this study, we investigated the integrated technique using the combined GPS/Galileo time-differenced observations for cycle slip detection and correction. Cycle slips in phase observations are detected utilizing snooping residuals from a least squares adjustment, in particular, with an assumption of student distribution for the residuals as the variance factor is unknown. For reliable cycle slip resolution, we take into account the correlation between the estimated cycle slip parameters for different satellites and systems instead of resolving the cycle slips in a satellite-by-satellite manner in case more than one satellite is flagged. These concepts were implemented within a precise point positioning (PPP) application and applied to real GPS and Galileo data collected at MGEX stations under different levels of ionospheric activity. We also tested other cycle slip handling techniques based the Hatch–Melbourne–Wübbena (HMW) and geometry-free (GF) combinations for comparison. The results demonstrated that PPP using the GPS/Galileo combined residual test detection approach can achieve better position accuracy and keep convergence even under active ionospheric condition (indicated by Kp index) compared to using the HMW- and GF-based detection methods. Moreover, cycle slip resolution using the integrated resolution scheme can lead to more precise PPP results compared to the satellite-by-satellite scheme, indicating that the correlation of cycle slip parameters between satellites could significantly affect the outcome of integer cycle slip resolution. However, more investigations are required when integrating observations from full GNSS constellations since a large computational load may degrade the efficiency at the GNSS data preprocessing stage.