paper: Cyber Blue 234 - Swerve Implementation Presentation

[Chris Fultz]

Thread created automatically to discuss a document in CD-Media.

Cyber Blue 234 - Swerve Implementation Presentation by Chris Fultz

[Chris Fultz]

This is the slide package from Cyber Blue's "Swerve Drive Implementation" presentation at the 2010 FIRST Championship Forum.

The presentation covers a brief history of omni-directional drives, the decision process the team used to select a drive type, the implementation steps, results and some lessons learned.

A companion "white paper" is in final review and will be posted soon.

Due to the size of the file, this has been converted to a PDF format.

Presenters were students Miranda Goelz, Marcus Hosler and Logan Milner and mentor Collin Fultz.

[Ether]

Hi Chris,

On Page 11 of the presentation, it states that "low traction wheels are required" for holonomic drive. Could you please clarify what you meant by this?

Thanks

~

[Akash Rastogi]

Originally Posted by Ether

Hi Chris,

On Page 11 of the presentation, it states that "low traction wheels are required" for holonomic drive. Could you please clarify what you meant by this?

Thanks

~

Depending on what motors are used, low traction wheels are better for turning the modules. WildStang uses kit wheels for example.

[Ether]

Thanks Akash.

I'd like to hear from Chris or Collin or someone on Team 234 with technical expertise who contributed to the paper if they are listening in.


~

[Matt Goelz]

Originally Posted by Ether

On Page 11 of the presentation, it states that "low traction wheels are required" for holonomic drive. Could you please clarify what you meant by this?

All this really means is that if you are trying to execute any omnidirectional drive other than swerve, you will have to use omni wheels or the likes of them. However with the swerve drive you can use any wheel you want: roughtop, wedgetop, rubber, plastic, etc. This is how you can have better traction than other holonomic drives.

It kinda refers back to the point of "the wheels have to slip for the system to move" under the holonomic drive column on that same page.

I did not contribute to this presentation but I was involved in the brainstorming, choosing, and testing of this project. This point came up a lot when we were discussing what type of holonomic drive to test.

I hope this helps.

[Ether]

Hi Matt,

Thanks for the response.

When you say "the wheels have to slip for the system to move" what do you mean by that? Are you referring to the rollers spinning, or are you referring to slipping between the floor and wheel contact surfaces?

If the latter, where did your team get that understanding?


~

[Chris Fultz]

The reference mainly applies to mecanum wheels. for these wheels to work and allow the robot to slide side to side, the small rollers must be low traction.

Same with an Omni wheel - they are made to slide side to side, so they have to be low traction.

With a swerve drive, since the whole wheel set rotates to change direction, a higher traction wheel material can be used. In these systems, the wheel does not slide across the floor.

For how did we learn this, it was just through analysis of the different systems and an understanding of how and why they work.

[Matt Goelz]

sorry i didnt notice this was already responded to, but this explains the mechanum and holonomic drive situation in a little more depth if needed.

Its both actually. This understanding comes from any holonomic drive that uses omni wheels. Each wheel has 2 directions it wants to go. The way the system goes one direction specifically is by each direction canceling out with an opposite direction from the adjacent wheel except the direction it will travel. You can only achieve this type of drive that uses this implementation with wheels that slide easily or omni wheels.

A mechanum or a holonomic drive where the wheels are angled in the four corners in a diamond pattern for example:
the front right wheel wants to go forward and right
the front left wants to go forward and left
the back right wheel wants to go forward and right
the back left wheel wants to go forward and left

This results in the system moving forward because the lefts and rights canceled out. To be able to cancel out directions like this, you have to use a very low friction wheel or omni wheel. In theory, if you have wheels that slide very easily, they will slip on the floor and move the robot. I have never actually seen this done, as all the holonomic drives I've seen use omni wheels.

[Ether]

Originally Posted by Chris Fultz

The reference mainly applies to mecanum wheels. for these wheels to work and allow the robot to slide side to side, the small rollers must be low traction.

Same with an Omni wheel - they are made to slide side to side, so they have to be low traction.

With a swerve drive, since the whole wheel set rotates to change direction, a higher traction wheel material can be used. In these systems, the wheel does not slide across the floor.

For how did we learn this, it was just through analysis of the different systems and an understanding of how and why they work.

Hi Chris,

Thanks for responding.

A mecanum vehicle does not require low traction tread surface on the rollers in order to move sideways. The sideways motion of the vehicle is accomplished by the combined action of the wheels turning and the rollers rolling. It does not depend on "sliding" at the roller-to-floor contact interface.

If you have access to a good university library, an excellent discussion of this is provided in chapter 3 of Introduction to Autonomous Mobile Robots by Siegwart and Nourbakhsh. You can also probably get the book via inter-library loan through your local community library.

I've posted a short paper here...

http://www.chiefdelphi.com/media/papers/2390

... which derives the inverse kinematic equations which show how to properly program a 4-wheel mecanum vehicle's individual wheel speeds in order to accomplish any desired combination of simultaneous vehicle motions (fore/aft; sideways; rotation) without scrubbing the wheels against the floor.


An explanation why a mecanum vehicle has less traction than a standard-wheel vehicle (even if both vehicles have the same tread material) is given here:

http://www.chiefdelphi.com/media/papers/2385


I'd be interested in discussing this with you if you disagree after reading the book or the papers.



~

[Collin Fultz]

Ether -

Thank you for the response and questions. Matt and Chris have pretty well captured our team's thoughts going into the decision making process. Our largest reason for not selecting a holonomic or mecanum drive for the fall project was the loss of pushing force commonly experienced by robots with those drive systems.

I haven't been able to read the links you posted (dang job...), but will put them on my to-do list.

Thanks again for the comments. I'll get back to you once I am able to read your links.

-C