Nowadays robots have proven their usability in more and more areas, even in industrial and commercial applications. One subcategory of robots is mobile robots, which are capable of movements. One critical task for these robots is motion planning. In my thesis I present my development results on this topic, I analyzed path planning algorithms in simulated and in real environments.
To accomplish my tasks, I used some previous research results from the Department of Automation and Applied Informatics. To integrate the path planning algorithms into the Gazebo simulator, I used the Robot Operating System framework. In the simulator, I built the model of the real navigation environment at the department. With the help of a camera based localization system, I managed to set up the real world layout of the robot and obstacles in the simulator. To implement motion planning on the real robotcar, I had to modify its low level firmware components. With these modifications the robot is capable to navigate among obstacles using the camera based localization and motion planning algorithms.
In a motion planning software library developed at the department, I made improvements on an existing path planner algorithm. The modifications are based on a Matlab version of the algorithm. After the improvements the path planner is working faster and generates better quality paths than originally.