Collision Avoidance in Multiple-Redundant Manipulators

Abstract

A new objective function is introduced to solve the under- specified inverse-kinematic problem of multiple-redundant manipulators. The objective function, Maximum Distance Criterion (MXDC), is based on the sum of the distances be tween key points along each of the manipulator kinematic structures. These points are called configuration control points (CCP), and are selected to represent sufficiently the manipula tor's configuration in Cartesian space. The objective function is used as a global criterion in multiple-redundant arm-joint con figuration optimization, enabling safe joint configurations to be determined while the robots are tracking desired end-effector trajectories. MXDC is tested using computer simulations, and compared with other existing criteria for optimizing multiple- redundant manipulator configurations. The proper selection of a set of configuration control points is vitally important in the application of MXDC. This is investigated in a series of simu lation examples, and qualitative guidelines for choosing CCP are presented.