4 ms·

Min7 micromouse robot solves maze in 3.921 seconds

• ColinWright 15y ago
  > [The exploration phase is] slower, it's > still impressively fast considering there's > only a 20MHz processor running inside That's what people think, but the processing is trivial. The hard parts are the mechanical grip and the control theory that lets you know exactly where you are, where you're pointed, and how fast you're going. My experience of building and running one of these things is that computing the exploration, solving the maze, and computing the fastest path (which isn't always the shortest[1]), is all a complete doddle compared with coping with wheel slip, bumps in the surface, and precise control of the speeds of the wheels. Granted, my mouse never ran at that speed, but I had an 8 bit 4MHz processor and it was mostly idle. [1] A longer route can be faster if it has fewer corners. I was at the competition where a mouse first ran a diagonal, and thereafter the designs of all the mazes changed to account for this.