Visual Tracking for an Eye-in-Hand System with the Unknown System Dynamics Estimator

In this paper, a visual tracking controller with an unknown system dynamics estimator (USDE) is presented in order to deal with the visual tracking for an eye-in-hand (EIH) visual servoing system. We first propose the USDE to estimate unknown dynamics in robotic visual servoing system by deploying filter operations and convenient algebraic calculations. In this line, the information of acceleration in the image space can be avoided. Moreover, the USDE is incorporated into a visual tracking controller. The uniformly ultimately boundedness of the closed-loop system is provided. Besides, the tracking error and the estimation error both exponentially converge to a small compact set around zero. Finally, the effectiveness of the proposed controller is shown by the simulation of the EIH visual servoing system with a two degree-of-freedom robot.