Evaluating Accuracy of IMU Motion Capture System for Whole Body Movements

Abstract

Recently, the IMU motion capture system has been developed as a system that can be used in sports practice. In this study, we examined the accuracy of the kinematic variables measured/calculated with a commercially available IMU motion capture system for three whole body movements (squat, walking, and jump) by using an optical motion capture system for determining the corresponding kinematic variables as reference. The result demonstrated high correlations (> 0.94) between the two systems for the positions of the center of mass on the main motion axis. In the walking task, the IMU system overestimated the total distance traveled but the amount of the overestimation was limited to about 3%. In the center of mass velocity data of IMU system, a large ‘spike’ of measurement error was observed at the instant of landing and/or the instant of take-off. This might be due to the algorithm of the IMU system used for detecting whether the foot was placed on the ground. The rate of change in the vertical velocity of the center of mass was found to match well the theoretical value (gravitational acceleration) during free fall. The vertical angle of the long axis vector of each segment showed an average difference of 11.5 degrees compared to the value calculated from the optical system. The results indicate that the selected kinematic variables can be measured by the IMU motion capture system with an acceptable accuracy, so that it can be applied as a new measuring device in measurement in an environment where it was difficult to measure with an optical system, especially in sports motion analysis including the aerial period.