A Swerve Drive Question

[sanddrag]

why would you have a swerve (rotating wheel modules) drive with omni wheels?

[Dave Scheck]

Originally Posted by sanddrag

why would you have a swerve (rotating wheel modules) drive with omni wheels?

After rereading your question, I think you're asking about a rotating module with an omni wheel mounted to it. The answer below answers the question of why would you mix crab and omni.

In 2004, we had crab on the front and fixed omnis in the back. You can see it in this picture. The treaded wheels are able to rotate and the omnis are fixed.

We learned from past experience (2002, 2003) that with 4 treaded wheels, zero-point turning was hard on the battery. In 2003 we ended up with a skid pad that was dropped pneumatically to turn quickly in place. In 2004, the omnis did the job. They allowed for quick turning, as well as crabbing with a little software magic.

[lndnny]

Originally Posted by sanddrag
why would you have a swerve (rotating wheel modules) drive with omni wheels?

I figured that rotating and moving would involve wheels at awkward angles to each other. The omni wheels would prevent scrubbing (bouncing).

Originally Posted by Noah Kleinberg
If you rotated the wheels so that diagonal wheels are parallel (like a holonomic drive), then it should be able to rotate and translate simultaneously.

I had figured you could just do that, but wondered if there was another way to do it. With a holonomic drive, isn't it best to have the wheels in a perfect square (or diamond)? The wheel modules in a design I've been working on are farther apart. I've been working on a swerve bot for the learning experience and as a cool demonstration robot .

Sorry I don't know how to do the quotes

Thanks
Dan
Team 808

[sanddrag]

What is your reason to put omni wheels on a crab/swerve drive? Is it so you can have 100% of your power going toward your direction of travel as opposed to 70% and still be able to rotate while translating? That's the only reasoning I can think of.

[Noah Kleinberg]

Originally Posted by lndnny

I had figured you could just do that, but wondered if there was another way to do it. With a holonomic drive, isn't it best to have the wheels in a perfect square (or diamond)? The wheel modules in a design I've been working on are farther apart.

They don't have to be in a perfect square, and not being in a square shouldn't make it much less efficient, but it might be harder to program it. See this thread for more: http://www.chiefdelphi.com/forums/sh...ad.php?t=38839.