Data-driven Self-calibration Algorithm for IMU Sensors on Diverse Mobile Devices

Inaccurate IMU sensor data from mobile and wearable devices poses a significant challenge for various applications. Improving IMU data accuracy is critical for enhancing the performance of many applications on mobile and wearable devices, and these applications have wide-ranging impacts in the area of mobile and wearable computing. The main challenges in obtaining more accurate IMU sensing data include device-dependent and time-varying errors such as bias, scale factor, and non-01ihogonality misalignment, which can severely affect data quality. Traditional calibration methods are often unable to effectively address these dynamic errors. This project proposes a new method to generate high-precision reference data for IMU calibration. By adopting data-driven calibration techniques, the research aims to develop a more accessible and accurate solution to improve IMU sensor performance.