Path-planning and Navigation of Mobile Robots using Real-time Vision

OData support
Dr. Prohászka Zoltán Ferenc
Department of Control Engineering and Information Technology

The aim of my diploma project was to make a complex, mobile robot directing system. As an antecedent event I would like to mention that I have participated several times in Competition of Applied Scientific Engineering. There the competitors have to solve different tasks with the help of own-built robots on a specially made environment – imitating the Mars. During the accomplishment of the task the competitors cannot have a view of the field, but they can follow the happenings through the image of the webcam (“satellite”), that is placed above the field. The mobile robot has to be directed through wireless connection. But all of the commands that are initiated by humans are necessary to delay with 30 seconds. Beside conditions like these, those autonomous functioning mobile robots arrive at advantageous positions that require the minimum of intervention.

In my diploma I solved a task on the basis of an earlier announcement of competition, in such a way that the components of the implemented system can be utilised in the future too. According to the task the aim of the mobile robot is to transport back the earlier scattered pieces of lego from the locality to their place. The weight of pieces can be 1, 2 or 3 unit. At the same time the robot can transport maximum 3 unit weight.

In favour of task solving I developed a strategic module that plans the movement of the robot between certain locations, in such a way that it can do the least distance to transport back all of the (lego) pieces. On behalf of optimal route finding between certain locations without an impact I made a route-planning module. However on the marked route the path-following module leads along the mobile robot. For the accurate path-following it is necessary to determine the current position and the orientation, for this reason I deposited markers on the robot.

The implemented marker detection module is capable of recognizing the position of the robot by the picture of the webcam. Therefore it provides exact figures for path-following module.

In the first chapter of my dissertation I specify more exact the task mentioned above and its usage cases. The second chapter deals with functioning of certain modules, while in the third chapter I represent functioning of the robot directing program that can solve the task.


Please sign in to download the files of this thesis.