Infosystem 2009 – Showcasing the mazebot

Presenting the mazebot robot in Infosystem 2009 an annual exhibition held in Thessaloniki. Photos and videos of the robot in action are included.

Robot in the maze

Even though a little late I am making a blog post regarding Infosystem 2009 an international exhibition about information technology and communications taking place each year in my home city Thessaloniki, Greece. The institution from which I graduated, the Technological Educational Institute of� Thessaloniki, Department of Informatics(hereafter referred to as TEI) tends to take part in this exhibition every year and it did so this year too. The exhibition took place in 26-29 of November. It was requested of me to bring the mazebot to the exhibition and showcase it to the visitors of TEI’s stand. It made sense since the mazebot project was actually my graduation thesis and so I accepted. If you do not know what the mazebot project is about follow the link above to learn more about it and understand what this post is about.

Of course I did not regret accepting the invitation. It was a great chance to test the mazebot in a completely new and different environment (as far as lighting conditions are concerned) and also to show it to more people outside of TEI. Not to mention have fun while I am at it. Below you can see a few pictures of the stand the Department of Informatics had in the exhibition. Note the reconfigurable maze that the mazebot works in, located in the middle of the stand.

Infosystem stand 1

Infosystem stand 2

As can be seen the maze holds the lion’s share of the stand’s space, which makes sense since it alone attracted many curious visitors asking what it was. For the duration of the exhibition I tried different maze configurations and set it as my goal to avoid training the robot anew wanting to make sure that its previously trained neural networks can generalize in new and different mazes and lighting conditions.

Robot in the maze

In the image below you can see me adjusting the robot in the maze while at the same time explaining its function to some visitors to the stand. I will have to admit that after the first day of repeating the same thing over and over and over again it became really tiring and that I should have written it all down in some poster for people to read. That would not only have saved me lots of repeating myself but it would have given me ample time to tinker with the robot and test additions/changes to the robot’s functions. I was not able to test anything, or look at any source code for the duration of the exhibition since I was constantly answering questions and repositioning the robot to make yet another run inside the maze 🙂

Me and the robot

As far as the movement of the robot inside the maze is concerned it went a lot better than I had anticipated. I expected that I would need to re-calibrate many variables for the robot to function correctly under these totally different lighting conditions. Instead just after 1 hour of remembering my own code, since I had not really spent a lot of time working on the mazebot for months, I started seeing results. We witnessed many successful maze runs for the duration of the exhibition, 3 of which I managed to capture in video. After editing them to add information about both the graph search algorithm uploaded to the robot and about the navigation algorithm which utilizes neural networks I uploaded two of them on YouTube. I apologize in advance for my amateur editing but I think that if anyone has read the mazebot project page he can figure out what both the graph and the code means. So if you have not already done so, go and have a quick read.

In the video below I uploaded the A* graph search algorithm to the robot which it interprets with the built in C interpreter and tries to find the exit to the maze using that algorithm. You can see which parts of the algorithm’s code run while the robot moves in the maze. This video is a nice demonstration of all of the mazebot’s functions.

In the second video the robot follows quite an unorthodox route to the exit of the maze on purpose. It was to test a new maze configuration. While the robot traverses the maze, shots of the navigation algorithm appear in the screen showing which parts the robot executes at any given time.

Infosystem 2009 was a great experience for me and a very nice test for the mazebot. In my opinion (a quite biased one ) it performed really well in an untested environment. Robots, coding and surrounded by friends. What else could a person ask for?

All the guys

Finally here is a picture of me at the end of the exhibition. Admittedly I have a very crazy look in this picture, I guess I might have been tired. You can see me holding the robot while at the same time videos of it moving inside the maze are playing in the background.

Me and the robot crazy look

My thoughts regarding Infosystem are mixed up. Compared to previous years traffic was quite lower and the number of stands has also diminished. So Infosystem as an exhibition seems to be dying. On the other hand the stand of TEI, Department of Informatics was as alive as ever and I really enjoyed myself for the duration of those 4 days. I am very happy to have been able to test the mazebot especially among friends whose input I really valued. I would grab another chance to do so without seconds thoughts should such a chance be presented to me.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.