Short Circuit at the Java Storms Challenge

Nederlandse Tekst

A team consisting of two students and a teacher Electrical Engineering (EL) finished as third at the Java Storms Challenge, a contest organized by the IT-company Chess-iT. On Friday 8 and Saturday 9 February Chess-IT organized the Java Storms Challenge for students for the second time.

The participating teams had to program LEGO Robots with a Java programming environment developed by Chess-IT. These Robots were used on a track, which consisted of a chessboard with fields of 25x25cm. The objective was to move an object by to places, but before getting there, all kind of obstacles should be avoided. Time and accuracy were essential for this track.

The Twente team consisted of Jan Broenink (Teacher at Control Engineering, EE), Bert van den Berg (Student EE, 4^th year) and Edwin Dertien (Student EE, 5^th year). The name of this team was “Short Circuit”, obviously to take the shortest path on the track, but also named after a famous robot movie.

Bert had some experience with the Java-system and Edwin had quite some experience with the LEGO Mind Storms system. 

LEGO® Mind Storms™ consist of a programmable LEGO brick (the RCX), two motors, a light-sensor, two micro switches and a lot of normal LEGO technics bricks. 

Friday at 18:00 h the contest started at Chess-IT with coffee and sandwiches. Employees of Chess made the objective clear and also gave some idea of the kind of track we could be facing. The final track with the precise position of the obstacles would be announced at 13:00 h the next day. This made the design of the robot harder, and there was little time left for calibration.

TheLEGO Mind Storms kit just contains three sensors: a light sensor and two micro switches. These sensors are not always reliable; therefore the obvious solution would be an ‘open loop’ system. As Electrical Engineers we were in favour for a ‘closed-loop’ system and therefore we wanted to avoid the obvious ‘open-loop’ system. Thus we decided to build some kind of a closed-loop system. 

After the first evening (and night until about 3:00 h) a prototype was made, which uses the two micro switches to make rotation-sensors. These rotation sensors were used to measure the distance (odometry) and can be used on any track.  The first control engineering hurdle to take was to let the robot move along a straight line, while using two non-identical motors to dirve the two big wheels and a castor for stability. Lego also has rotation sensors, but these do not belong to the Mind Storms kit. 

The sensors we made generated too much “events”.  This caused the communication to have an overflow of data, and important events were lost. It was therefore impossible to make a reliable controller. Unfortunately, it took the whole morning and a part of the afternoon to find out that this solution did not work. 

At 14:00 h the final track was announced. We had just lost the hope of making a nice closed-loop system and we were about to make an open-loop system. This ‘time-measuring’ robotic system was of course against our principles..

After carefully looking at the track, we concluded that we had another option. We could get to the target by using the left wall as a guidance system (left-wall-following). With this option it seemed relative simple to get to our target object (which had to be moved by two fields). The construction needed for this solution only consisted of two motors and an extended bumper operating the micro switch.

In the last available hour and after three major changes of the robot, we finally had a working solution. These robot changes were necessary to obtain more points, as two of the criteria for the jury were construction and the lovability of robot. We added some parts to the bumper so it would not get stuck in the different obstacles. Also a part was mounted enabling the robot to catch the target object.

After some calibration the robot succeeded to get to the target. We did not succeed in the programming the last part of the assignment; the robot was not able to move the target to the target field. 50 Points were awarded for the fact that our robot reached the target. The points for performance were calculated with the following formula: 50 – (10*n) points, where n was the distances in fields of the robot to the target. We could have got another 50 points when the robot actually moved the object to the target field.

Summing all points given for source-code, lovability (ours was most lovable), construction and performance, the team “Short Circuit” got a third ranking, just after the Hogeschool Amsterdam and TU/e. The last mentioned was defending their title of last year, and as I might add, they succeeded nicely.

We made some silent movies (mpeg) with our Digital Still Camera. You can see the left-wall following, including a right turn in the corner. (605 kB, 6 s). You can also see the robot avoiding obstacles and finally getting to its target, the target is moved against a blue obstacle, but our robot could not find its way to the target field (6 MB, 1:06 minut). These movies where made during the try-outs, when enough light was available. Therefore you will miss the lovable look of Short Circuit.