Visual Servoing of Mobile Robots with Sphere Objects

The problem that using visual feedback to control the distance and orientation of the mobile robot with respect to a static sphere object is considered in this paper. Firstly, a unit virtual sphere is added on the classical camera model to obtain accurately the direction of the object. After measurable signal analysis, the kinematics model of the system is obtained. Then a switched controller and a continuous adaptive controller are developed to drive the mobile robot to the desired pose. Lastly, in order to estimate the distance between the camera and the object, a nonlinear observer is designed to give an exact estimation for the radius of the object, thus no metric information of the object is needed. Simulation results are collected to validate the effectiveness of the proposed method.