Minimizing the Effect of IMU Misplacement With a Functional Orientation Method

To collect reliable data, it is important to determine how inertial measurement unit (IMU) sensor placement affects measurements of segment motion and to establish best practices for sensor placement and orientation procedures. We aimed to determine the extent to which a functional orientation method minimizes the effect of variations in sensor placement on IMU-derived segment excursions. Twenty healthy adults walked overground in a lab. Three IMUs were placed per segment on the pelvis, thigh, shank, and foot. Differences in estimated segment angular excursions between sensor placements were compared between an assumed sensor-to-segment orientation and two versions of a walking-based functional sensor-to-segment orientation. For time series data, differences in angular excursion between sensor placements were reduced from 60% of the gait cycle (assumed orientation) to 44% of the gait cycle (functional orientation) for the pelvis; from 31% (assumed) to 28% (functional) of the gait cycle for the thigh; and from 84% (assumed) to 0% (functional) of the gait cycle for the shank. Mean angular excursion RMSDs between sensor placements were <5{degrees} for most comparisons for the functionally oriented data. Functional orientation did not minimize inter-sensor placement effects when one sensor was located on a region with high soft tissue artifact (e.g., anterior thigh). Functional orientation reduced inter-sensor differences in segment excursion for the shank and foot and between-subject variance in inter-sensor differences for all segments. Functional orientation minimizes the effect of variations in IMU sensor placement, but care should be taken to select sensor placements that minimize soft-tissue artifact.