Application-oriented joint rate minimization by cell and tool optimization for an R6-manipulator

A robot's joint rates increase when passing near-singular kinematic setups, increasing the required axes currents and wearout of the bearings. Exceeding a manipulator's axes acceleratory limits will abort its motion and therefore make an off-line-planned trajectory undriveable in a real setting; Latter also applies to hitting the axes' angular constraints. Given a desired trajectory and aiming for reducing joint rates, we discuss an application-oriented offline approach for multi-dimensional optimization of the tool definition using a heuristic method for an industrial R6 manipulator, a KUKA KR 60/2. We analyze the different kinds of singularities this robot's kinematic involves. As optimization criteria, we present a way to identify and quantify their impact on manipulability and prioritize optimization with respect to the arm and the wrist. We also show that decoupling may be applied for choosing an axes configuration that avoids hitting their angular constraints. The presented method is adapted to a plasma-cutting application. It avoids common approaches' draw-backs like pose or speed deviation or the need for an auxiliary axis.