Student "edition" found at {csi dot journalspace dot com}.

Maybe I shouldn't have started this blog now, not with everything that's been going on.

In the first meeting of my Introduction to Robotics classes for the ninth week of the third term, the students came up with another suggestion for renovating the maze to suit their needs to be able to get out of it.

There was this one corner or turn that the robot is supposed to maneuver, after which it comes to a T intersection.

If it moves to the right, it can fall into the trap of any one of three dead ends.

If it moves to the left, it goes straight to the exit, albeit after three more turns, but no more forks.

Since most of them programmed that the robot will go into a random direction if both touch sensors are pressed (meaning that the robot has a wall in front of it) or to go opposite the last direction it turned (which would always be a left), they wanted to stack the odds in their favor that the robot would always turn to the left.

The loophole they found to do this, and to which I agreed, was for them to put an extension to the last wall of the turn, but still large enough for any of their robots to go through straight.

This ensured that it increased the possibility that the robot would hit that wall before rounding the turn completely, so instead of heading towards the intersection head on, it would turn leaning to the left.

This wall the right bumper will hit first, and their robot will then follow its natural programming to go to the left.

This was the only dead end lead that I agreed for them to change though.

The other dead ends are in tighter areas of the maze and thus cannot be reconstructed.

The group that suggested this though, has unique bumpers in that it’s shaped like a Y in front of the robot, and has the possibility of getting caught if the robot were turning backwards or at an angle.

Session 1029 is at a dead end. Class dismissed.