Adaptive visual tracking with uncertain manipulator dynamics and uncalibrated camera

Abstract

A control-theoretic solution is proposed for designing a stable visual servoing adaptive controller for a planar robot manipulator based on a monocular and fixed camera. The controller should allow tracking of a moving target, under uncertain knowledge of the robot/camera parameters. In order to solve the direct parameter adaptive problem related with the camera calibration uncertainty, a novel stable adaptation scheme is introduced. The scheme is then combined with a robust or adaptive controller for the manipulator, taking into account its nonlinear dynamics, leading to an overall stable adaptive system.