The great evolution of Micromouse

You might think that there are only so many ways a robotic mouse can run a maze, but in its nearly 50-year history, competitors in Micromouse events have time and again proven that assumption wrong. In the video after the break, [Veritasium] He takes us on a wonderful journey through Micromouse development competition bots.

The goal of Micromouse is simple: get to the destination (center) square of the maze in the shortest time. Competitors are not allowed to update their cars’ programming once the build has been revealed at the start of the event. Over the years, there have been many innovations that may seem obvious now but were groundbreaking at the time.

The first, most obvious challenge is to find the center of the maze. The simple wall in the first event in 1977 evolved into variations of the “flood fill” algorithm. At first all the robots stopped before turning until someone realized you could cut 45 degree corners and move diagonally if the robot was tight enough. The shortest route is not always the fastest as the turn loses a lot of speed, so it is sometimes possible to improve the time by choosing a slightly longer steerer with fewer angles.

More speed is only good if you can stay in control, so many robots now include propellers to suck them in, which increases traction. This has resulted in speeds of up to 7 m/s and cornering forces of up to 6 G. Even specks of dust can cause loss of control, so all competitors use tape to clean their wheels before they run. Many winning runs are now under 10 seconds, which requires many iterations of the design to increase controllable speed and reduce weight.

All of these innovations started as experiments, and the beauty of Microhouse is how accessible they are. It doesn’t take a lot of budget to get started, and the technical barrier to entry is lower than ever. We’ve looked at another Micromouse design before. Even if they aren’t tiny, we can’t get enough of tiny robots.