Dynamic task-priority allocation for kinematically redundant robotic mechanisms

This paper presents a flexible redundancy resolution strategy based on the task-priority method. A dynamic task-priority allocation approach has been motivated by the fact that the performance may degenerate for improper fixed-priority assignment to various task components. Recursive local kinematic inversion is applied, which, along with a full task-decomposition approach, guarantees the flexibility of the approach. It is further shown that the damping technique is easily implemented yielding a scheme that performs well also in the neighborhood of singularities. Thereby, the computationally inefficient singular-value-decomposition has been avoided