Inertial Measurement Units (IMUs) enable accurate estimation of anatomical joint angles but require a sensor-to-segment calibration. Literature has presented several algorithms that address this gap; however, adequately comparing calibration performance is not trivial. This study compares 3 calibration methods: N-pose calibration (NP), functional calibration (FC), and manual alignment (MA) to estimate 3D wrist joint angles during single-plane and multiplane tasks. Thirteen healthy participants were instrumented with IMUs and optical markers to compute the range of motion error (e), root mean squared error, and offset between the joint angles from the optical reference and each IMU calibration (NP, FC, and MA) as dependent variables. We then performed 3-way repeated-measures analyses of variance on each dependent variable to evaluate interactions between calibrations, tasks, and joint axes. NP showed the worst root mean squared error (8.34° [7.41°]) performance in the calibration main effect (n2G = .095) and calibration × tasks interaction (n2G = .121). In an exploratory analysis, FC performed best (main effect root mean squared error = 6.52° [4.47°]) in the offset calibration × axes interaction in single-plane (n2G = .160) tasks. Therefore, we recommend FC to optimally perform wrist calibration and against NP. These findings are viable in aiding the development of portable IMU-based clinical motion-tracking devices.
© Copyright 2025 Journal of Applied Biomechanics. Human Kinetics. Alle Rechte vorbehalten.