A modified threshold model of HMW combination for detecting cycle slips in high-latitude regions during ionospheric scintillation periods

Abstract

Currently, solar activity is in its 25th peak year, with frequent occurrences of ionospheric scintillation, resulting in a significant increase in cycle slip events in Global Navigation Satellite System (GNSS) observation data and significantly affecting the reliability of precise positioning services. Accurate cycle slip detection is key to improving positioning performance. However, the fixed-threshold detection method often fails to adapt to ionospheric scintillation variations, resulting in high false alarm rates, frequent misjudgments, and subsequently reduced positioning accuracy. To address the limitation that the existing Hatch-Melbourne-Wübbena (HMW) combination cycle slip threshold cannot adapt to variations in ionospheric scintillation intensity, this paper analyzes the variation patterns between the scintillation factor Rate Of Total Electron Content (TEC) Index (ROTI) and the cycle slip detection values, i.e., Extra-Wide-Lane (EWL), Wide-Lane (WL), and Narrow-Lane (NL), and finds that the fluctuation range of the detection values increases with rising ROTI. Based on previous findings, a regional threshold model is proposed. Under low scintillation conditions, a fixed threshold is selected; under strong scintillation conditions, an adaptive cycle slip threshold detection model is constructed by performing quadratic fitting based on the relationship between ROTI and the detection values. The proposed model is validated using GNSS observations collected from nine high-latitude stations during 2023–2024 and is compared with both the conventional fixed-threshold method and the dynamic threshold method proposed by Zhao et al. (2019). Compared to the fixed-threshold method, the false alarm rate of this scheme is reduced by 39.1%, and the missed detection rate is reduced by 32.6%. Under strong, medium, and weak scintillation conditions, the positioning accuracy is improved by approximately 40%, 30%, and 10%, respectively, while the number of re-convergences is reduced by an average of 42.6%, 25.33%, and 19.57%. Experiments demonstrate that the proposed regional adaptive threshold model exhibits good stability and applicability under different intensities of ionospheric disturbances, effectively enhancing the accuracy and reliability of precise point positioning in high-latitude regions.