Inertial Measurement Unit (IMU) sensors, including accelerometer, gyroscope, and magnetometer, are used to estimate the orientation of mobile devices. However, indoor magnetic fields are commonly distorted, which causes the magnetometer’s readings to deviate from true north, resulting in inaccurate orientation estimation. Existing solutions either ignore the magnetic distortion or avoid using magnetometer when distortion is detected. In this paper, we develop MDR , a Magnetic Distortion-Resistant orientation estimation system, which fundamentally models and resists magnetic distortion in orientation estimation. MDR builds a database to record the magnetic directions at different locations, and uses it to correct orientation estimation from magnetic distortion. MDR tackles two unique challenges. First, to avoid the overhead of database preparation, MDR adopts practical designs to automatically build and update the database via self-exploration. Second, to reduce errors in database operations, we design a set of techniques, and thus further improve orientation accuracy. Extensive experiments on arm motion data of 27+ hours show that MDR outperforms the state-of-the-art method by 44.5 \(\% \) .
Yang et al. (Wed,) studied this question.