Minimization of Power Losses in Cooperating Manipulators

Abstract

The control of multiple manipulators handling a common object entails the solution of an underdetermined system of linear equations which represents the system’s dynamics. In order to choose an optimal solution to this problem, various approaches have been proposed: minimum internal force and minimum power, among others. The present work investigates an approach for minimizing the power losses in these systems. It is shown that the power imparted to the manipulator/payload system cannot be optimized once the system’s motion is prescribed. However, assuming certain loss characteristics for the dc servomotors commonly used on robotic manipulators, it is shown that the minimization of power losses can be cast as a linear-quadratic optimization problem. Local and global performance indices are introduced to allow comparison of the minimum power loss and the minimum internal force approaches. An example of two Puma 560 robots handling a common payload is shown to demonstrate the proposed technique.