Improved time-differenced cycle slip detect and repair for GNSS undifferenced observations

GNSS undifferenced signal processing, in which the individual signal of each frequency is treated as independent observable, has drawn increasing interest in GNSS community. However, undifferenced signal processing brings new challenges for cycle slip detection and repair. One important feature is the carrier frequency identification of cycle slips since observations are processed separately. An analysis of real cycle slips in a BDS triple-frequency baseline dataset illustrates the deficiencies in the cycle slip detection process commonly implemented, in the case when cycle slips occur in just one specific carrier frequency. Hence, we propose an improved cycle slip detection and repair approach based on a time-differenced model. Two major advantages characterize this proposed approach. The first one is a significant reduction in false alarms due to carrier frequency identification of cycle slips. Having access to a reliable cycle slip detection method significantly reduces the number of ambiguity parameters to be estimated. The second advantage is the benefit of separating the OSS (Observation at the other frequency of Same Satellite without cycle slip) and OSE (Observation at the Same Epoch from other satellites without cycle slip) from the OCS (Observation with Cycle Slip). The simulation results indicate that separation of the OSS and OSE can significantly improve the model strength of cycle slip estimation, especially OSS. The proposed approach is validated by cycle slip estimation with a real data set. Smaller biases and larger ratio values jointly demonstrate that a much stronger model strength can be achieved. Finally, the cycle slip repair procedure is applied to triple-frequency PPP. The stable and fast convergence, as well as the reduction in standard deviations, proves the efficiency of the proposed approach.