Follow the Line 2007

The Robotics Club at my Faculty organized a competition – Follow the Line 2007. Rules? Pretty simple – one has to build a robot that will try to make it through a course set by a white line on the faculty’s floor. Time limit – 3 minutes. Course length – about 25 meters. The fastest one wins… My colleagues were building a robot for some time and they decided to take part. As I was also building a robot I decided to compete as well.

Preparations

I switched from Keil to WinARM (no code limit, and I had to do it anyway, sooner or later)

I tested my AD converters – my ARM has 14 of them, but only 8 work (probably after my experiments with electronics…)

I (with help from my colleague experienced in SMD soldering) created a sensor board made from 5 TCRT1000 circuits (phototransistor and LED diode working in ifrared)

At first I decided to mount the sensors in the front, on a special adjustable arm designed by me… The stability and durability of this thing was a huge disappointment. In the end, I mounted them on the bottom of the robot, using a tape… Actually, I started liking tape constructions… Fast and reliable;) The algorithm was a very simple proportional regulator, with the addition of finding whether there actually is a line under the sensors and optional going back in a searching mode…

At home, the robot behaved pretty good, so I went to the faculty to try it out on the test-track – it was so slippery, that Hildorfr had about 1 or 2 seconds of inertia… The problem was overcome by attaching some rubber to its tracks… It clung to the ground pretty well, but also slowed my robot considerably… I didn’t have much time to test other options, so that was it…

Competition

I was the third one to start…

First run – 1:45 – giving him (a temporary) third place and the honour of being one of 5 robots that made it through the whole course during the first round
Second run – robot hung approximately 1 meter from the ending line…
Third run – robot went off track after ~10 meters…
Final result – 7th place, out of 20 (11 robots finished the whole course within the time limit). Much better than I anticipated, but… Disappointing given the first run. While others were “upgrading” their robots, I tried to do the same. But instead of “upgradeing” Hildorfr, I messed it up, especially the sensors… Conclusion – I shouldn’t mess with my robot “on field” when I don’t have an ISP programmer to debug it (no COM port in my laptop…)…

Next year it should be better:)

Here are some links:

The competition’s homepage, with gallery and official results
Youtube movie featuring my first (best) run – from 5:30 onward

Enjoy!