Estimation of Uncertainties in 3D Motion Analysis Data

A complete understanding of motion is important for assisting clinicians and engineers in the prevention, diagnosis, and treatment of people with physical disorders. According to an ISO standard, three dimension (3D) motion measurements are always related to some errors and therefore always only obtained values. For 3D motion measurement systems, when applied to measuring human motion, motion data is sometimes reported with a standard error of measurement reflecting the uncertainty of the measurement caused by this skin movement artifact during motion as well as system error and marker noise. This lack of information seriously limits the ability to investigate motion analysis for the researchers. The purpose of this paper is to quantify the uncertainty caused by skin movement artifact as well as sensor noise and measurement error of the motion analysis system. We present a method for incorporating the uncertainty of the matching between a model of the calibration device (3D robot) and measurements, obtained with a motion analysis system (Optotrak Certus), in a Kalman filter.