Optimized GNSS/INS Integration with Time Synchronization for High-Accuracy Smartphone Positioning

This study proposes an optimized, tightly coupled global navigation satellite system (GNSS)/inertial navigation system (INS) integration method for accurate smartphone positioning, developed and evaluated using data sets from the Google Smartphone Decimeter Challenge (GSDC) 2023-2024. In the GSDC literature, inertial measurement unit (IMU) data have been used in only a limited manner, primarily because of GNSS-IMU time misalignment between smartphone system clocks and GNSS time. To address this issue, we introduce a lightweight post-processing time synchronization scheme that selects a per-drive constant time offset by minimizing the root mean square of position corrections within a Kalman filter. The proposed method incorporates innovation-based adaptive measurement weighting to accommodate varying GNSS measurement qualities across smartphones and environments. Experimental results demonstrate that compared with conventional differential GNSS approaches, the proposed GNSS/ INS integration reduces the mean of the 50th and 95th percentile horizontal errors by up to 50 cm. The proposed method achieved submeter accuracy for the first time in a competition, securing first place at the GSDC 2023-2024 challenge.