Image-based visual servoing of a Gough-Stewart parallel manipulator using leg observations

In this paper a tight coupling between computer vision and parallel robotics is exhibited through the projective line geometry. Indeed, contrary to the usual methodology where the robot is modeled independently from the control law that will be implemented, the proposed method takes into account, from the early modeling stage, the fact that vision will be used for control. Hence, kinematic modeling and projective geometry are fused into a control-de voted projective kinematic model. Thus, starting from a vision-based kinematic modeling of a Gough-Stewart manipulator a visual servoing scheme is presented, where the image projection (edges) of the non-rigidly linked legs are servoed, rather than the end-effector pose or the leg directions.