You can get class credit for the coolest things these days. Take for instance, this Automatic Chessboard that [Brian] and [James] built for the final project in one of their classes this spring. We just looked at a robotic chess setup on Monday that used a gripper mounted on a gantry to move the pieces. This one’s a lot more user-friendly and borders on magical. That’s because the moving parts are all located below the board and could be hidden from view if a proper case were built around the edges.
There are two main components to this build. The first is a grid of reed switches that detect the moves made by a human. This works because each piece the human player uses has a weak magnet glued to the bottom which is just strong enough to actuate the reed switch and let the computer sense what move was just made. On the robotic side of things this works like a plotter. Each of the computer’s pieces has a metallic disc glued to the base. What basically amounts to a plotter under the board uses rare-earth magnets to grab the computer’s piece and drag it to the next playing position.
The use of two separate magnetic systems provides some interesting design challenges. You can see the device in action in the video after the break, and a full writeup and source code package is available at the blog linked at the top of this feature. But for your convenience we’ve also mirrored the PDF whitepaper after the break which lays bare all of the juicy details.
Sounds horrible… as in the motors…
Gotta love engineers. Build a brilliant and clever autonomous chess set, then support it with electrical tape and wheelie chairs.
On a side note, did this remind anyone else of Jumanji?
Would be neat if all the pieces had magnets, so then the computer could play either/both sides, and detect when someone is messing with its pieces! Quieter actuators would help too.
The robot arm looks cooler, but the reed switches are the way to go for sensing the players moves.
It sounds like they are grating some food. The noise is distracting.
Better title: Automated chess set prefers bottom.
Nice idea (albeit hardly original) but rather poor execution. More like a “mee too” (not unlike the 100000001 dinky “robot arms” whose only redeeming quality is that “this one’s made by me”). It would be essential to have both sides playable by the machine (not always the black) and it is also essential to make it as noiseless as possible, or else the magic goes away real fast and one might as well use a simple chessboard-on-a-screen.
So what happens when a horse has to pass though 2 other pieces?
I’ve had this idea for quite some time but never got around to implementing it… only I was going to make a Ouija board ;-)
@Bram,
Why don’t you watch the video and find out. It moves one in the first minute of the 2 minute video. (spoiler: it just slides it in between them.)
My grandfather had one of these. I’m not sure when it was made, guessing 70’s.
The noise was similar, but the pieces were correctly sized in proportion to the squares; it just moved the offending pieces to the side (then back) for a knight move.
I’m not finding the device on the almighty Google(tm), which is disappointing. Despite the noise and significant weight, it worked perfectly and beat my young self two to one.
sad that the piece always approaches the destination from the same side, rather than taking the optimum trajectory from where it came from.
Also, maybe it’s using the reed switches to stop the movement? The pieces never seem to quite make it to the center.
If it put a bunch of pieces of center, it may foil its clever “move in between pieces” plan.