A FAST APPROACH FOR THE ROBUST TRAJECTORY PLANNING OF REDUNDANT ROBOTMANIPULATORS

In this article, a fast approach for robust trajectory planning, in th e task space, of redundant robot manipulators is presented. The approa ch is based on combining an original method for obstacle avoidance by the manipulator configuration with the traditional potential field app roach for the motion planning of the end-effector. This novel method i s based on formulating an inverse kinematics problem under an inexact context. This procedure permits dealing with the avoidance of obstacle s with an appropriate and easy to compute null space vector; whereas t he avoidance of singularities is attained by the proper pseudoinverse perturbation. Furthermore, it is also shown that this formulation allo ws one to deal effectively with the local minimum problem frequently a ssociated with the potential field approaches. The computation of the inverse kinematics problem is accomplished by numerically solving a li near system, which includes the vector for obstacle avoidance and a sc heme for the proper pseudoinverse perturbation to deal with the singul arities and/or the potential function local minima. These properties m ake the proposed approach suitable for redundant robots operating in r eal time in a sensor-based environment. The developed algorithm is tes ted on the simulation of a planar redundant manipulator. From the resu lts obtained it is observed that the proposed approach compares favora bly with the other approaches that have recently been proposed. (C) 19 95 John Wiley & Sons, Inc.