The great problem
We quickly finished JohnBot's body, as we only had one more thing to add after our first session. Today, the concerns trascended the physical plane, as we had to worry about giving JohnBot a mind. As explained before, JohnBot needs to complete the Wumpus task; consequently (and obviously), we need to develop an algorithm that fits the rules of this task.
Mostly, we were concerned with having JohnBot be able to navigate through this Wumpus world 6x6 grid, find some "gold," and be able to kill the Wumpus in case it's found (In case this doesn't make sense, go to the "home" page for more details). An easy task for a human, but could our newborn JohnBot do it?
Most of the time during this session was spent discussing a seemingly unsolveable complication, which could pose a great threat to JohnBot's task. The problem consists on any board where we could have the following the situation:
On these two situations, our robot would received the same signs when making its second move, after starting on the bottom left space. That is, even though the pits have different positions on Figure 1 and Figure 2, the robot would be unable to realize there is a difference. We decided to call this a "false positive."
However, and luckily, we were assured that there would not be any unsolveable problems in the actual puzzle. This possibility consists of an unsolveable puzzle, so it was quickly discarded. Thus, our main concerns was quickly dissipated by the rules of the Wumpus world itself.
After doing a lot of hard thinking, we decided that the best path to take was to finish JohnBot's mind (or at least the "skeleton" of its mind) outside of class. We did, and JohBot, although he might not be able to state it, was pleased with it.
Comments