A Study of Localization and Map Building Algorithms on a Real Robot

OData support
Kiss Domokos
Department of Automation and Applied Informatics

Nowadays autonomus robots have increasing role, in industry but also in daily life we confide a lot of things to them. Mobile robots have the capability to move around in their environment. The localization problem of mobile robots is an active research field, so there are a lot of algorithms from the past decades. To calculate the accurate position of the robot, it is neccessary to build a model from the environment, in which we can locate the robot.

First I studied localization and mapping algorithms in simulation, after that I implemented these algorithms on a real robot, that is available at the Department of Automation and Applied Informatics. Before this I had to study the available tools, such as the omnidirectional robot, the laser rangefinder and the ROS operating system which is responsible for running algorithms.

The goal of this thesis is to implement algorithms that on the one hand can calculate the robot's real position, on the other hand can create a map from the robot's environment. After that I studied the well known SLAM (Simultaneous Localization and Mapping) problem and implemented an algorithm, that includes these two functions at the same time. Finally I implemented an algorithm, that navigates the robot fully autonomusly in the previously totally unknown environment.


Please sign in to download the files of this thesis.