Abstract
The influence of the mounting position of a magnetic-inertial measurement unit (MIMU) on the accuracy of posture estimation for a shank has not been extensively studied and remains unknown. In this study, we conducted comparative experiments using three MIMU positions: the lateral and frontal positions, which are commonly used, and the medial tibial position, which is less affected by muscle protuberance, considering the anatomical structure of the body. To determine the optimal MIMU mounting position on the shank, we repeatedly performed plantar-dorsiflexion and relaxation of the ankle joint in a chair-sitting position and examined the effect of muscle contraction on the posture of the MIMU (Experiment 1). We also performed posture estimation during gait and compared the three-dimensional shank posture measured by the MIMU and optical motion capture to evaluate the estimation accuracy for each mounting position (Experiment 2). In Experiment 1, the orientation change at the medial tibia was significantly smaller than that at the other positions, showing an 80% reduction compared with the anterior tibia during dorsiflexion. In Experiment 2, the medial tibia achieved the highest estimation accuracy, showing a 13% lower RMSE than that of the anterior position. The results of these two experiments suggest that the medial tibia is the optimal position on the shank, as the posture estimation accuracy was the highest when the MIMU was mounted on the medial tibia, where there was no muscle under the mounting surface. Moreover, the posture estimation accuracy was less affected by muscle protuberance under these conditions.