Inertial sensor based inverse dynamics analysis of human motions

Abstract

Inverse dynamics analysis, which enables the information about joint force/torque of human body from the measured motion data, is a traditional tool in biomechanical studies. Conventionally, measurements of body motion are obtained by motion capture cameras, of which the bulkiness and high cost make it hard to capture some movements that need broad range of space. As an alternative, inertial motion sensing using cheap and small microelectronic mechanical systems (MEMS) based inertial sensors has been proposed. Inertial motion sensing allows ambulatory movement and is also free of occlusion problem. It has been actively studied for recent 2 decades, but until now, most researches have focused on measuring only the kinematics of body segments. In this research, the performance of inertial motion sensing especially aimed at inverse dynamics analysis is studied for squat motion. Kinematic data from inertial motion sensors are used to calculate the normal ground reaction force (GRF), which is evaluated based on the force plate measurement (ground truth) and additionally compared to the estimation result using optical motion sensing system. Squat motion, which can be simply modeled as a 3-segment linkage moving in sagittal plane is targeted. The normal GRF estimation result showed high correlation and low RMSE (R=0.93, RMSE<0.02 of body weight), which guarantees enough accuracy of inertial motion sensing to be used for estimating joint forces/torques through inverse dynamics analysis about squat.