Effects of IMU Sensor-to-Segment Misalignment and Orientation Error on 3-D Knee Joint Angle Estimation

Accurate joint angle estimation using magneto-inertial measurement units (IMU) is a significant challenge for human movement analysis outside of the lab. Sensor-to-segment misalignment and sensor orientation errors are key factors in joint angle estimation. However, the effects of these factors on 3-D knee joint angle estimation accuracy are not well understood. This paper presents a general model to investigate the effects of sensor-to-segment misalignment and IMU orientation error on 3-D knee joint angle estimation accuracy. An analytical model was built to quantify the effects of sensor-to-segment misalignment and IMU orientation errors for 3-D knee joint angle estimation. Precise outcomes were quantified via simulation analysis and experimental testing in which fourteen healthy subjects performed drop landing and cutting tasks while wearing IMUs and optical motion capture markers for ground truth estimation. Sensor-to-segment misalignment accounted for 52% of the total knee abduction angle error and 40% of the knee internal rotation angle error but did not significantly affect knee flexion angle error. Conversely, IMU orientation error accounted for 55% of the knee flexion angle error but did not significantly affect knee abduction or internal rotation angle errors. These results indicate that sensor-to-segment misalignment significantly affects knee abduction and internal rotation angle estimation while IMU orientation error significantly affects knee flexion angle estimation. This study could serve to inform the development of accurate IMU-based 3-D knee joint angle algorithms for human movement analysis.