Effect of sensor to segment alignment errors on IMU-based running joint angle estimates

Engineering of Sport 15 - Proceedings from the 15th International Conference on the Engineering of Sport (ISEA 2024)

Body-worn inertial measurement units (IMUs) are increasingly utilized to capture human performance in real-world scenarios (e.g., during training or competition). For example, many physics-based sensor fusion methods combine raw IMU data with kinematic and anatomical constraints to estimate human joint angles. For running and sprinting, lower-limb joint angles show importance in many traditional indoor camera-based biomechanical studies; however, translation to real-world studies requires further advances in the IMU-based methods. Beyond the well-known challenge of correcting integration drift errors, a primary challenge towards accurate IMU-only estimation of joint angles is the accurate estimation of sensor-tosegment alignment (S2S) parameters (i.e., the relative orientation and location of the IMU and anatomical ‘xyz’ frames). However, the minimum S2S accuracy that is required for accurate joint angle estimates is largely unexplored due in part to the presence of other confounding error sources in experimental studies (e.g., soft tissue artefacts). In this study, 96 000 simulations are used to better investigate the impact of S2S parameter errors on the accuracy of IMU-based lower-limb joint angle estimates during running.