An Outlier Processing Method Integrating Data Snooping and Quasi-Accurate Detection for Incomplete Carrier Phase Measurements

Global Navigation Satellite System (GNSS) radio signals are extremely susceptible to interference in urban environments, and even multi-frequency geodetic receiver observation data are often incomplete. It usually happens that receivers that can output multi-frequency measurements only output carrier phase measurements for a single frequency, and there may also be multiple unstable small cycle slips. Therefore, this paper proposes a method for outlier detection and estimation suitable for incomplete carrier phase measurements. It is based on the principle of data snooping to eliminate special cycle slip combinations that cannot be detected by the classic dual-frequency combination method, thereby obtaining a highly reliable subset of quasi-accurate measurements and then combining the quasi-accurate detection mechanism to estimate all outliers. Experiments show that this method can solve the problems of the classic ionospheric residual (IR) and Melbourne-Wübbena (MW) combination method being insensitive to special cycle slip combinations and unavailable when measurements are incomplete, and can also make up for the insufficiency of the pseudorange phase combination method in detecting multiple single-frequency continuous small cycle slips less than 2 cycles. In addition, it can also detect and estimate multiple outliers at the same time, avoiding the error transfer problem that often occurs in the detection, identification, and adaptation (DIA) data snooping method, and the absolute error of the estimation of outliers is less than 0.3 cycles, the mean error is less than 0.01 cycles, and the RMSE is less than 0.04 cycles. Especially when partial carrier phase measurements are damaged and incomplete due to urban environments, it can accurately handle all outliers using only GNSS measurement information, supporting subsequent short-baseline carrier phase differential positioning.