Visual servo control for a class of mobile robot

Abstract

Application of vision based control of mobile robots is a very promising area, however, solutions to real problems are almost always hampered by the difficulties associated with computational effort and robustness of image processing tasks. In the paper, an implementation for visual servo control based on off-the-shelf software and hardware components for wheeled nonholonomic robots is presented. All the algorithms are based on well established methodologies selected with simplicity as a fundamental criterion. The image processing tasks are implemented using mathematical morphology operators and the transformation from image parameters to 3D velocities is based on the interaction matrix approach. The results obtained are satisfactory and there is still a good margin of improvement in terms of execution performance.