In this page you can find everything about my master thesis: "Study and design of a fuzzy hierarchical behaviour-based controller".
The focus of this thesis is on autonomous robot navigation in real-world, unstructured environments. This field has the goal of moving the robots, purposefully and without human help, in environments that have not been engineered for them. In particular, our work is focused on the study and development of a specific module of a robotic system, called control module.

We have developed a hierarchical, modular, efficient, behaviour-based controller, integrating it in a pre-existent architecture for autonomous robots. This architecture, and hence our controller, will then be used in the application domain of assistance to fire fighters in emergency response operations. Finally, we have experimentally validated the controller both in simulation and on a real robot in indoor environments.

I have worked to this project at the AASS center at Orebro University in Sweden, under the supervision of the professor Alessandro Saffiotti.

 

Click on the icon to download my thesis:
Click on the icon to download my presentation:

Here below you can find the videos that I made during my experiments with a real robot: to watch them you need the DivX 5.1.1 codec or a later version.

The first experiment:

The experiment reported here consists in one complex behaviour that let the robot to follow a corridor, face a door and finally cross it. The door is more than 9 meters far from the start position of the robot and in the corridor there are two obstacles to avoid. You can download the video that shows the experiment (the file is quite big, about 35 MB).

The second experiment:

This experiment is very similar to the previous one but the speed has been increased to 0.3 m/s and the first obstacle and the wood panels, that were in the corridor, have been removed. The video shows both the robot and a debug window where you can see which behaviours are active and what control values are sent to the robot. Note that to reduce the dimension of the file (now is only 9 MB), the video has only 6 frames per second.

The third experiment:

This experiment is totally different from the others because in this case the robot has to reach a particular point following a path that is computed on-line. As you can see in the video, at the beginning the robot doesn't have a complete map of the corridor and the path builded is unfeasible. After a while the robot is able to complete the map and to find and follow a good path to reach the goal. The video shows three different windows: one with the robot, one with the map of the corridor and one, called Local Perceptual Space window, where you can see the information collected by the robot using its sensors.