A parallel solution to robot inverse kinematics

Abstract

Introduces an algorithm by which the inverse kinematics of a robot manipulator with closed-form solution can be computed in parallel to reduce the computational complexity roughly by a factor of n, the number of joints of the manipulator. The authors study the errors introduced by the algorithm by constructing statistical models of the errors. They evaluate the algorithm on a PUMA 260 manipulator, which has six degrees of freedom with six revolute joints and a reach of approximately 0.5 m.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>