Automatic Discovery of Redundant Constraints for Globally Optimal State Estimation

This paper proposes an automatic method to find a set of sufficient redundant constraints to obtain tight semidefinite relaxations for globally optimal solutions to state estimation problems in robotics.