Study of singularity robust inverse of Jacobian matrix for manipulator

A new solution to the SR-inverse (singularity robust inverse) of a Jacobian matrix for serial manipulators is presented. The joint axis linear dependence and the task space feasible motions at singular configurations are determined by using classified line varieties and reciprocal screws, respectively. By dropping the rank of the Jacobian matrix near singular configurations, an ordinary and systematic arithmetic is obtained. Numerical simulation of PUMA manipulators demonstrates the feasibility of this approach.