Inverse Covariance Intersection Fusion for Lie-Group-Based Pose Estimation

This letter is concerned with the pose estimation problem on Lie groups. In general, robots naturally move on the special Euclidean Lie groups, which provides the motivation to model the measurement uncertainty on Lie algebras and project it onto Lie groups. Then, a Lie-group-based pose estimation method under inverse covariance intersection fusion is proposed, in which the obtained estimates are more accurate than in vector spaces. Since the unknown common information shared among the measurements is taken into account, the proposed Lie-group-based pose estimation method has the advantages of higher precision and better consistency. Finally, the effectiveness of the proposed method is verified through simulations.