My thesis has been written in the Robotsoccer Project running on the Department of Control Engineering and Information Technology. This project has been starting for a few years in the Intelligent Robots Laboratory of the department. The fundamental objectives of the research project are identical with the robotsoccer projects running on other places in the world, I mean the major goal is to create a robotsoccer team, in which the robots are able to play against each other without all human interventions. During the game only the controller computer can send information independently for the robots, the operator can only instruct the computer to start the game and to stop the game.
To achieve the goals in robotsoccer the effective control of the teammates is indispensable. To this, there is a need for a visual system, which is able to define the status of the robots on the robotsoccer ground. This observational system consists of two cameras fixed above the ground, which provide image information about the position and the orientation of the robots. These information can make it possible to lead the robots by visual feedback. The aim of this thesis is to develop calibration and control procedures which are based on visual feedback.
The aim of the calibration is to define wheel drives which keep the robots on a straight orbit. I developed two kinds of calibration procedure during my work. The off-line calibration defines the necessary drives with the movements of the robots. For this, I elaborated a procedure based on circle fitting. The on-line calibration mends the results of this pre-calibration procedure during the movement of the robots with the support of a fuzzy controller.
After the review of the calibration procedures I present a developed procedure for the position and orientation control. The presented procedure allows to lead the robots to an optional position and to add a given orientation. Finally, I present two methods of leading the robot along an orbit defined with edges. These methods are as follows: the control point by point and the orbit-following control.