This paper describes a procedure developed and validated to assess the accuracy of an infrared-based motion measurement system used to perform a kinematic analysis of the torso with respect to the pelvis during simulated lifting tasks. Two rigid reflective marker triads were designed and fabricated for attachment to the thorax over the 6th thoracic vertebra and the pelvis. System accuracy was assessed for planar rotation as well as rotations about multiple orthogonal axes. A test fixture was used to validate known triad orientations. The spatial coordinates of these triads were collected at 120 Hz using a ProReflex motion measurement system. Single value decomposition was used to estimate a rotation matrix describing the rigid body motion of the thorax triad relative to the sacral triad at each point in time. Euler angles corresponding to flexion, lateral bending, and twisting were computed from the rotation matrix. All measurement error residuals for flexion, lateral bending, and twisting were below 1.75°. The estimated mean measurement errors were less than 1° in all three planes. These results suggest that the motion measurement system is reliable and accurate to within approximately 1.5° for the angles examined.
