Angular Acceleration-Free IMU-Joint Vector Calibration Based on Measurement Integrals

Inertial sensor-based human motion tracking technology has many applications, including but not limited to sports, cinematic rehabilitation, and gaming. Human pose estimation is highly sensitive to the accuracy of the IMU-joint vector, which is defined from the center of rotation (COR) of joints to the inertial measurement units (IMUs). The sensor measurement error, caused by soft tissue artifacts (STAs), and the joint motion substantially impact the IMU-joint calibration. This article proposes an IMU-joint calibration method based on measurement integrals that does not need the derivative of noisy IMU measurements for angular acceleration. A number of influencing factors such as the STAs noise, IMU-joint distance, and the joint angular velocity are investigated by simulations and real tests. The proposed method demonstrates not only better calibration accuracy but higher robustness than the state-of-the-art methods.