Adaptive Visual Servoing Control of a Manipulator with Uncertainties in Vision and Dynamics

Abstract

In this paper we present a new adaptive dynamic controller based on vision for tracking objects with a planar manipulator robot in fixed camera configuration, considering the orientation of the camera assembly, the depth of the object and the dynamic parameters of the robot to be uncertain. The control scheme is designed with a vision-based adaptive kinematic controller in charge of executing the task of tracking the object even with unknown parameters of the vision system. This controller provides speed references to a dynamic cascade adaptive controller whose objective is to generate the final control actions of the robot with imprecise knowledge of its dynamics. Finally, it is shown the stability of the proposed system based on the Lyapunov Theory and the performance of the two adaptive controllers through simulations are shown.