A Study of Navigation Algorithms on a Real Robot

OData support
Supervisor:
Kiss Domokos
Department of Automation and Applied Informatics

Industrial applications of mobile robots have became widespread, and nowdays these machines can be found in our close environment, even in our homes. Taking automated warehouse forklifts or robot vacuum cleaners; they are required to interact with their surroundings. In order to fulfill this requirement they must gather information about their environment, detect the obstacles, and control their movement accordingly.

A navigation algorithm is the software solution, which leads the robot from it's actual position to a given destination. By processing the environmental data, these algorithms must navigate the robot safely - without any collisions - and they have to be convergent, they have to reach the goal position when possible. During the last decades numerous navigation methods had been introduced, there is significant research activity at our department too.

The Eurobot contest is an international autonomus robot contest, where our team participates with a robot, every year. In this paper, a study of navigation algorithms is presented. Our aim was to develop a working, real-life implementation, in order to use it at the competition.

The first part of this paper discusses the field of the navigation algorithms, especially the Dynamic Window Approach. The second part describes the implemented methods. A simulator tool had been developed for testing these solutions. After they became stable, the next step was the real-life testing using our previous robot (which was made for the 2012. Eurobot contenst). These tests proved, that these navigation algorithms are well suitable for the contest.

Downloads

Please sign in to download the files of this thesis.