Pose estimation for servicing of orbital replacement units in a cluttered environment

This paper presents our work from an on going project on motion and pose estimation for supporting vision-guided fine manipulation of orbital replaceable units in a cluttered environment. The major challenges of this work include target objects, some with highly reflective or mirror surfaces, moving in a cluttered environment with changing illumination. These challenges are dealt with through closed-looped integration of video from multiple cameras and manipulator joint angle measurements, along with moving edge detection and model-based feature tracking. A powerful technique based on a Lie group and its associated Lie algebra is employed for pose estimation. Effectiveness of the proposed techniques is illustrated by good test results for data acquired in the laboratory.