Analysis of MEMS IMU accuracy metrics for Android smartphones

Regarding the relatively limited research on the accuracy evaluation of Android smartphone inertial measurement unit (IMU) sensor data, particularly the ambiguous aiding role of GNSS signal tracking in complex scenarios, this paper quantitatively assesses IMU data quality from 11 Android smartphones through three aspects: bias estimates, scale factor, and navigation error accumulation. The results demonstrate that: 1) The built-in accelerometers generally show better accuracy and stability with bias estimates typically below 100 mGal; 2) The embedded gyroscopes exhibit relatively poorer precision and stability with bias estimates ranging 5°～20°/h; 3) Significant scale factor errors exist in gyroscopic observations with varying magnitudes, and navigation error accumulation shows marked improvement after scale factor error calibration; 4) When IMU navigation errors are required to be within 10 m, the standalone navigation duration generally does not exceed 30 seconds; 5) Newer smartphone models generally exhibit significantly better IMU performance metrics than older ones. This research provides technical support for understanding smartphone IMU sensor accuracy levels, further mitigating error impacts, and enabling assisted GNSS signal reacquisition along with precise user positioning in complex scenarios.