Game Board Mark I
The first step in designing the board was trying to figure out the size. Since the robots turned out to be around four-inches at their largest dimension, I decided the board size should consist of a 12×12 grid of five-inch squares laid out on eighth-inch plywood.
With the help of some metric conversions, a chalk line, and some luck we managed to lay out a grid of magnetic tape. The board was further divided into nine pieces of 4×4 squares, this was to make assembling and transporting the board easier.
With the board assembled, the robots were placed on the board and their movement was tested and fine tuned.
The next step was to design and print the board with all the elements on it. I started by recreating a classic RoboRally board from scratch (with some modifications) so that it would be high resolution enough to be printed in the ~5×5 square feet needed. Since the bots had trouble climbing over the joints between the board pieces, printing the board on vinyl helped smooth out the bumps. My first attempt at designing a board left the vinyl a little too small, but it was big enough to test and tune the the bots on it and play a few test games.
Game Board Mark II
After some experimentation and observation we determined that two separate issues seemed to be causing our robots to go off-course: First, the board itself had warped, and second it seemed that the intersections weren’t distinct enough. After some head scratching and fiddling we determined that the board needed a complete redesign, and thus we set about making RoboRuckus Factory Floor Mark II!
The warpage issue was caused by using thin materials that could warp simply from humidity and how we were storing the board squares. These warps made the transitions from one board square to another bumpy and uneven and sometimes threw the bots off the line. If the bots left the line they’d travel until they detected another one, sometimes several squares away from their original target. We decided that we’d have to go with a thicker, sturdier material for the board squares and settled on 1/2 inch MDF (Medium Density Fiberboard). This had the advantage of being stiffer and less likely to warp in storage, which would allow the board squares to lay flat and eliminate any bumps.
One of the other navigation issues seemed to be that the robots often moved too many spaces. Our robots navigate by sensing the magnetic tape that’s laid out in a grid pattern under the graphic. As they pass over the intersections where the strips of tape crisscross, the robots note the increased magnetic flux where the strips overlap and log each spike to count the number of spaces traveled, but two layers of magnetic tape under the board wasn’t quite distinct enough for the robots to reliably detect as an intersection. We solved this issue by setting the tape into the top of the board instead of running it underneath (bringing it closer to the sensors) and embedding a rare earth (neodymium) magnet at each juncture to ensure that the magnetic field at the intersections was significantly higher than the lines. These two changes significantly improved the performance of the robots and made runaways much less frequent. As a bonus, the same vinyl mat from before could be reused.
