The evaluation of absolute position drift of inertial-based motion capture systems

Abstract

Inertial based capture technologies allow for direct capture of motions within a manufacturing environment and, may be used with human simulation software to perform ergonomics analyses. However, these technologies are negatively affected by metal environments where “drift” has been shown to cause error in capture accuracy. Twenty participants completed four multi-task events simulating real work, in a laboratory while instrumented with inertial and optical based motion capture systems. Participants began and ended each event by performing a static T-Pose posture in a known location. Lower-leg 3D position data were extracted and the position difference from the start and end T-Poses were analyzed. Results indicate significant lower-leg position error relative to the starting location of the T-Pose to the end, with the inertial systems as compared to the optical based system. These errors can impact overall accuracy and representation of work within a human modeling environment.