This is a cheap and easy to build ROV (Remotely Operated Vehicle). I came up with it because I wanted to create a minimalist aquatic robot that avoided the traditional annoyance of needing to get / make waterproof motors for ROVs. This ROV uses a new form of aquatic propulsion that works via the creation of multiple standing vortices around the ROV's asymmetrical hull (if this isn't new please let me know). The propulsion system doesn't require contact of any actuator with the surrounding water, and is very easy to build and is very robust.
Previous bio-inspired ROVs have used flapping fish-like fins to move through the water. Although these fins work by creating vortices along their surfaces, the ROVVor creates constantly maintained, standing vortices around its hull. These standing vortices are maintained by a high frequency vibration of the vehicle, induced by an internal pager motor. The vortices seem propel the vehicle through the water due to the vehicle's asymmetrical hull shape which must create some imbalance in the vortice's forces applied to different sides of the vehicle.
Due to its ease of implementation and robustness, this propulsion mechanism could potentially be useful for creating nano-scale ROVs (would need some way to vibrate the 'actuator molecule'), although at the molecular scale vortices might not work in the same way. Due to the ROVVor vortex-drive's minimal actuator displacement (compared to spinning propellers of flapping fins) it could also be useful for military applications to avoid detection. On a lighter note, the cheapness and ease with which the ROVVor can be created, and the interesting-ness of the way in which it works would make it very conducive to introducing kids to robotics in general and specifically to ROVs.
Step 1: Vortex-Drive Design Evolution
So I put a pager motor into a centrifuge tube (see figure I), and put it into a pool of water. It vibrated and I could see standing waves created all around it, but it didn't translocate (it didn't move through the water).
So then I tried adding little bristles to its sides (see figure II) and still nothing happened.
Finally, I starting thinking about adding a 'wave-sail' to the robot (see figure III) to take advantage of the waves created by the vibrations. I based this idea on the fact that one can use a fan on a sailboat to push the sailboat (see this great video: http://www.youtube.com/watch?v=0CrXvOKPymk). When I added this 'wave-sail' the robot started moving through the water.
However, when I trimmed the 'wave-sail' down a bunch (see figure IV) the robot still moved through the water, even faster than it had before. That's when I realized that the trimmed down 'wave-sail' was not acting as a wave sail at all. So I added sawdust to the water and saw a bunch of standing vortices all around the sides of the vehicle. I drew the vortices as I saw them as can be seen in figure IV). I then added some weights into the tube (little pebbles) and saw that the ROVVor could move even when completely submerged.
Then I thought that maybe if I put the same bendy straw on both sides of the vehicle that it would move even more powerfully through the water (see figure V) ,but when I did that it didn't trans-locate anymore! So it seems like the ability to move through the water is dependent on the vehicle's asymmetry and on the standing vortices created around it.
In the future I'm planning on trying a design in which multiple ROVVors are attached to each other (see figure VI) via a semi-flexible connection. In this way, by selecting which ROVVors to activate or deactivate one could selectively steer the whole structure in a desired direction. Because the connections are semi-flexible they would not propagate the vibrations from one vortex-drive subunit to another.
I haven't yet created a fully self contained version of the ROVVor because I couldn't find batteries small enough to fit into my centrifuge tube, and it's hard finding a water-tight tube that would fit a battery and the motor. I'll update this instructable if I find a right-sized tube and am able to implement a fully self contained version.