As part of an undergraduate reading group led by Nancy Ouyang, http://nouyang.blogspot.com/, an awesome fellow course 2-er, we made some small, origami-like hexapods! We following instructions based on research done at Berkeley: http://robotics.eecs.berkeley.edu/~ronf/Prototype/index.html
The flat-pack piece. It’s two pieces of regular cardboard with a sheet of thin plastic glued in between. The slits that you see are precut with a laser-cutter in the cardboard before the plastic is glued in between, so when this piece is cut into shapes the slits will become hinges made of flexible plastic. The instructions provide the laser-cutting files.
As we were trying to bend the pieces at the hinges, sometimes the cardboard would pop off. They were unfortunately fragile, probably because the glue wasn’t strong enough. The instructions recommend using a regular gluestick, but I’m convinced a spray adhesive, like Super-77, would be much better.
The top and bottom pieces (here left and right) are folded to have flexible parallelograms where the legs will attached. These parallelograms are key to the degrees of freedom of the robot. You can fold the parallelograms in two ways — up or down. It matters! But only with regard to the other piece…
Here you can kind of see how the three pieces stack to make the robot. The parallelograms attach the to the center piece. The top and bottom piece have 3 parallelograms each — when they are stacked, the parallelograms are always facing inwards and are opposite to each other. i.e. Directionality matters!
Another indication of directionality. The top and bottom piece each have a set of hinged pieces on one end, i.e. the parallelograms are on the long side of the piece, these hinged pieces are on the short side — when it gets put together, these should be opposite. So you need to match the side the parallelograms are folded on the top piece with the side the hinged piece should be on with regards to the bottom piece.
Finally put together! Everything is glued with superglue, because it holds well with cardboard and sets really fast. There are a lot of very small parts to be glued together, and I recommend doing one at a time, waiting until it sets, and then move to the next part. The whole thing is flexible, so you need to hold it while it sets so that it doesn’t fall apart.
Here you can see the two degrees of freedom. It walks with a tripod gait — three feet on the ground, move the other three feet forward, then switch the legs that are on the ground, then repeat. The center piece can contract and expand because it has hinges in the center. Our plastic was probably thicker than ideal — sometimes the hinges didn’t bend very well when you tried to contract the center piece. But thinner plastic would make the whole thing less stable.
So it worked! But it wasn’t actuated. In the Berkeley project they used SMA wire to actuate it, but that was outside the scale of what we could reasonably do in the couple hours we set aside to do it. 🙂