Bridging missing GNSS phase measurements through cycle slip detection for improved smartphone positioning

Abstract

Smartphone-related location-based services (LBSs) are among the most frequently used applications in daily life. The accuracy of smartphone global navigation satellite system (GNSS) positioning is limited to approximately ten-meter to meter-level precision due to poor observation quality and obstructed environments. For smartphone GNSS data, the phase measurements are frequently missing, which are treated as cycle slips in traditional algorithms, leading to the reset of ambiguity and variance, thereby decreasing positioning performance. To address this issue, this study proposes bridging the phase data gap in smartphone precise point positioning by utilizing previously available measurements before the signal outage to perform cycle slip detection, aiming to avoid excessive parameter resets in the filter. Indicators, including test values of geometry-free (GF) combination, code-minus-phase (CMP) combination, differences between Doppler and time-differenced carrier phase (TDCP), and pre-fit phase residuals, are selected to determine whether to reset the ambiguity and variance for a missing phase measurement. The proposed method is validated with 10 smartphone datasets collected in driving environments. The percentage of positioning errors within 1.5 m is improved from 65.1 to 99.1% for one of the Samsung S21+ phones. For a Mi 8 phone, the 68th and 95th percentile positioning errors are improved by about 2 dm. The overall statistics of the ten datasets demonstrate the correctness and efficiency of the proposed method, showing improvements in positioning errors within 1.5 m and 1 m by 4.3% and 8.2%, respectively. Moreover, the potential application of the proposed algorithm is validated with a geodetic dataset, showing significant positioning accuracy improvement when the signal outage occurs.

Jiahuan Hu

Postdoctoral Fellow

Research interests include GNSS precise point positioning, PPP-RTK, atmospheric modeling, and smartphone precise positioning.