I had this image of 'micromouse' competitions being very similar to humans trying to figure their way around a big maze. But, when I started to dig into it, I began to realize how far my perceptions were from the real world.
First, I naively thought that the robot mice would be down inside the maze and would only be able to see the walls immediately surrounding them. But, when I started to review the specifications for the maximum mouse size versus the dimension of the corridors (cells) in the maze I was surprised to find that the mouse could be larger than the channel it had to fit in... How is this possible?
My mental image of the competition was flawed. The base of the mouse has to fit within the maze channel - but portions of the mouse can rise above the walls and extend out for quite a ways. In fact, one of the most common approaches to sensing the walls is for the mouse to have a sensor array that tracks the top sides of the walls. Many of the mice operate as if they are looking down on the local part of the maze rather than being down inside it.
My second misconception was based on the old "Mr. Wizard" television program. In one of the episodes, Mr. Wizard taught me, and probably a hundred thousand other kids tuned in at the same time, that the easiest way to solve mazes is to put one hand on a wall and then follow it around until you come to the exit.
That approach works well for most mazes - assuming that they aren't specifically designed to defeat it. But at some point in the evolution of micromice a competition was held and the hands down winner was a simple wall following design that didn't even include a computer. At that point the rules were changed to make sure that the maze goal would be dead center in the maze instead of located along one of the edges. And steps were taken to make sure that just following a wall wouldn't lead you to the goal.
It will be interesting to see how many other misconceptions I have, and how many doors to the future are opened by correcting them.