Visual Servoing Controller for Robot Manipulators

Abstract

This paper presents a new family of fixed-cameravisual servoing for planar robot manipulators. The methodologyis based on energy-shaping framework in order to deriveregulators for position-image visual servoing. The control lawshave been composed by the gradient of an artificial potentialenergy plus a nonlinear velocity feedback. For a static target wecharacterize the global closed loop attractor using the dynamicrobot and vision model, and prove global asymptotic stability ofposition error for the control scheme, the so called position-basedvisual servoing. Inverse kinematics is used to obtain the angles ofthe desired joint, and those of the position joint from computedcentroid. Experimental results on a two degrees of freedom directdrive manipulator are presented.