Investigation of Motion Planning Algorithms on a Differential Robot

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

Nowadays autonomous robots are becoming more popular, for example, in the industry automated forklifts and loaders are used. Most of these use motion-planning algorithms to execute their task. In the last decades numerous algorithms have been developed, which proved to be efficient in solving the problems of motion-planning.

In this paper we developed two known motion-planning algorithms to a differential robot. The first is the Rapidly Exploring Random Trees (RRT) algorithm, which searches for a path to the given target on a geometrical map. The other is the A star search, which uses a grid map to plan the path. We examined the two algorithms using simulation.

We implemented a state-space regulation, so the robot could move more precisely.

We also wrote an exploring method next to the motion-planning algorithms, and modified a line extraction algorithm (Split and Merge), so it would work better in noisy environment.

Downloads

Please sign in to download the files of this thesis.