Application and Evaluation of Lighthouse Technology for Precision Motion Capture

Abstract

Lighthouse technology, commercially known as SteamVR tracking, is a 3D motion capture system developed for virtual reality applications that couples light-based and inertial tracking methods. In this paper, we benchmark its performance in motion tracking for possible applications in biomechanics and robotics. The tracking performance of the system was evaluated to determine the accuracy and repeatability for pose measurements from dynamic studies. In order to establish a baseline for tracking and to cross-validate the results, a seven degree-of-freedom industrial robotic arm was used. For precision, repeated motions were performed using the robotic arm with a motion capture tracker mounted on the end-effector of the robot. Accuracy was evaluated by comparing poses of the motion tracker from the lighthouse technology with the poses obtained for the robot's end-effector. Key to the evaluation approach was the use of an additional tracker which served as a ground reference frame for tracking. This improved the evaluation results by resolving global frame inconsistency issues observed during initial data acquisition and findings documented in referenced literature. The results obtained show accuracy and precision of the system to be millimeter and sub-degree level for position and orientation measurements respectively under good tracking conditions. Post evaluation, the system was applied to a phantom limb (prosthetic arm) for an inverse kinematics analysis. Recording a tracker attached to the initial and final segments of the prosthetic limb, the experimental data was applied to a kinematic model of the limb to solve for the joint angle values.