Position control of an Omnidirectional Robot through Visual Servoing

Abstract

The position control task of a mobile robot can be approached using different techniques. Among them, the Visual Servoing technique presents a control scheme using the data acquired through a camera, that in many cases simplifies the principal task, as generating a law control from image coordinates is more practical to implement. Considering these advantages, in this paper is presented a proposal of an ImageBased Visual Servoing control scheme to generate a trajectory between an initial position and a desired position that an omnidirectional robot would have to follow. The initial and desired position are identified as two images observed through a camera. Also, it is implemented a fuzzy logic system to manage gain values because of the possible changes in the system response speed. Finally, a simulation of the control scheme is presented, where it is showed that it is possible to reach the desired image, and it is concluded that the omnidirectional robot reaches the desired position. Also, it is shown figures related to the trajectory generated, the error functions, and the gain values during the process, as well as the error between the total distance desired and the distance obtained. Recommendations related to the kinematic model of the robot and the configuration of the fuzzy logic system are also presented.