Singularity-Robust Inverse Kinematics Using Lagrange Multiplier for Redundant Manipulators

Abstract

In this paper, a singularity-robust inverse kinematics is newly suggested by using a Lagrange multiplier for redundant manipulator systems. Two tasks are considered with priority orders under the assumption that a primary task has no singularity. First, an inverse kinematics problem is formulated to be an optimization one subject to an equality constraint, in other words, to be a minimization problem of secondary task error subject to an equality constraint for primary task execution. Second, in the procedure of minimization for a given objective function, a new inverse kinematics algorithm is derived. Third, since nonzero Lagrange multiplier values appear in the neighborhood of a singular configuration of a robotic manipulator, we choose them as a natural choice of the dampening factor to alleviate the ill-conditioning of matrix inversion, ultimately for singularity-robust inverse kinematics. Finally, the effectiveness of the suggested singularity-robust inverse kinematics is shown through a numerical simulation about deburring and conveyance tasks of a dual arm manipulator system.