1st Place Winner of the Smartphone Decimeter Challenge: Two-Step Optimization of Velocity and Position Using Smartphone’s Carrier Phase Observations

This paper describes a method that was declared as the winner of the Google smartphone decimeter challenge 2022 (GSDC2022), conducted on the competition platform Kaggle from May 3 to July 30, 2022. GSDC2022 was conducted to estimate the driving trajectory of a vehicle based on the raw global navigation satellite system (GNSS) data from smartphones. The GNSS data from smartphones have lower signal levels and higher noise compared to commercial GNSS receivers. Consequently, direct application of existing high-precision positioning methods, such as precise point positioning and real-time kinematic GNSS, is challenging. This study devised a two-step optimization method to optimize the velocity and position of a smartphone. In the first step, the states to be optimized were 3D velocity and receiver clock drift, which were estimated through the framework of factor graph optimization (FGO) using GNSS Doppler observations. The outliers of the optimized velocity were excluded. Further, the sections where the velocity were not obtained were interpolated to estimate the velocity of the entire driving trajectory. In the next step, the states to be optimized were 3D position and receiver clock bias, and the velocity and clock drift obtained in the previous step were used as loose constraints between the states. In addition, the time-differenced carrier phase (TDCP) was used as a high-precision relative position constraint, and an error-corrected pseudorange using a GNSS base station was added as an absolute position constraint. Following the implementation of the proposed method and its evaluation performed at the competition, the final score, mean(50%, 95%), was 1.382 and 1.229 m for public and private, respectively, which won the first place at GSDC2022.