This was originally designed to drive one of the spinning battlebot type robots, but I believe it has applications elsewhere if it works. The idea here is that you have a limited amount of power to both move and to operate your weapon, which in my case was to have the entire exterior spin and have spikey objects projecting out to destroy other bots. Usually this is done by have one set of motors that move the bot and then another set of motors spin the exterior. My idea was to spin most of the exterior of the bot and then drive the bot using a system similar to the cyclic controls on a helicopter. To do this you need a circular shaped platform drive wheels placed evenly around the edges of the circle who are arranged to spin the entire platform. At the center of the flatform will be a circular platform that rides on a lazy-susan bearing and driven by the drive motors in the opposite direction from the rotation of the platform, so the central platform effectively does not spin. Now connect linkages so that each of the drive wheels always stays perpendicular to a point (the control point) on the center platform and drive the platform clockwise. At the start and when you don't need to move the point is at the center of the platform, the center of both circles. Since the control point is on the center the drive wheels will drive in a circular formation. Mount your RC control on the center platform and servos to allow you to move the center control point left/right & forward/back. Now if they can take the Gs, mount your batteries to the spinning platform, if not mount them to the center platform. Take metal and encase the entire bot in heavy steel up to the limit of your class. Start the motors and the bot will begin to spin. All of your spent energy is now turning into rotational energy. Once the bot is up to speed, move the control point to the right and the bot will move forward, left is back, forward is left and backwards is right. Since you are storing up all your energy as rotation, it can be expelled violently for acceleration or for destroying anything the bot touches. Also the entire bot will be gyro stabilized so it would be nearly impossible to turn over. The only downside is that the bot is vulnerable till it comes up to speed, so you may want to add a mechanism to let the outer skin freewheel.

I think this system would also give you regenerative braking as it would work as an IVT.
-- MisterQED, Dec 03 2007

This drive is used in Tug Boats http://en.wikipedia...iki/Voith-Schneider
[MisterQED, Feb 03 2011]

Rovio http://www.wowwee.c...resence/rovio/rovio
3 omni-wheels - could be programmed to spin and drive. [neutrinos_shadow, Feb 03 2011]

//it has applications elsewhere if it works// not just [marked-for-tagline], but near proof we are putting out our 37% today
-- lurch, Dec 03 2007


I have a friend whose second job is working for a Tug Boat Winch maker. He started telling me about the drives they used and I realized it was the one I thought up for this idea. (link) So I guess it did have applications elsewhere and it does work. Just put wheels where they have vanes and you have my idea.
-- MisterQED, Feb 03 2011


It's mechanically sound, but sounds pretty fragile by robot combat standards. all those axles are going to have to be mounted on mobile points, which means they can't have as much support as a fixed bearing, plus all the linkages necessary to keep them properly aligned. Additionally, since at least some of the weight is being kept non-rotational, you are losing some of the benefits of traditional differential drive implementations, which have 100% of the robot's weight being in rotation.

As a final problem, your design does not solve the problem of absolute heading. When (not if) the "stationary" central platform rotates, either through wheels slippage or impact, your controls will no longer move the robot in the expected directions (this is the critical problem of all such designs, and I know because I was on the team that built Cyclonebot, which had a more standard differential drive system). In principle the drive wheels could be manipulated individually in order to bring the control point back into alignment, but that's impossible with your setup, since all the wheels are mechanically linked to a single control point. In principle, they could all be controlled by individual independent actuators, but that's asking an awful lot both mechanically and computationally.

It's clever, but I think it would lose out to simple, more robust designs. [+-]
-- 5th Earth, Feb 03 2011


While the cyclical mechanical control might work, using just 3 omni-wheels and clever software would work better. See [link] for an existing (non-spinning) version.
-- neutrinos_shadow, Feb 03 2011



random, halfbakery