Re: The Perfect swerve
 

What I was thinking of is to use the lights as constellations. You don't need color or features, just a filtered matrix of 1 and zero's.

Re: The Perfect swerve
 

No, you're missing my point. All else being equal, a given design will work better with some teams and some games than others.

OP asks questions like 'are 3in wheels better than 4in wheels?' That depends a lot on what exactly the robot needs to do. Does it have to traverse any obstacles? Does it need the extra ground clearance of a 4in wheel? Or does it need the extra wheelbase size afforded by the 3in wheels? One isn't inherently better than the other, just different. OP also asked 'are 2spd transmissions worth the weight and hassle?' Well, that depends a lot on the game and style of play of the team driving the robot. Small field with relatively short sprints to get from one place to another? Single spd is probably fine. Wide open field, demands for strong pushing forces AND high speeds? Might want to go with a 2spd setup. Again, one isn't simply better than the other, just better suited to different games and play styles.

Is field-centric control better than robot-centric control? Ask your drivers. Whatever they want or prefer is the right answer, one isn't automatically better than the other. The effectiveness of a given control style depends heavily upon the person behind the glass.

Then there is construction style. Is it 'better' to have a swerve module made out of CNC milled plates that are bolted together than to have a swerve module made out of bent and riveted sheet metal? Well.. a team with 6 CNC mills would have a different answer than a team with a waterjet and press brake. One design isn't necessarily better than the other, but a given design will be better suited to a particular team. I know, I know, you want to insert 'unlimited resources' caveat here, but that really doesn't matter. The fabrication style is very secondary to a well thought-out design. And looking at what's been posted in this thread so far it is obvious that there are a wide variety of designs that have been used very successfully.

Re: The Perfect swerve
 

Integrating this with gyro readings using a Kalman filter would be a fun project.

I've thought before (it may have been mentioned before on CD) about trying to track field position based off of the driver station lights: look for 3 red/blue lights with predefined spacing between them.

Re: The Perfect swerve
 

This year several teams have developed swerve that works very well. Right now I think the biggest problem now is not the physical swerve but the human interface and the driver. Giving the driver an extra degree of freedom really loads up the drivers brain when they are in a match and performing in a match. We over come the the human problem part of the system by going to as many off season events as we can to train our drivers. Going forward with swerve improvements needs to focus on the human part of the equation. Rock solid field centric control would be one path. A swerve simulator was posted on CD. It's primitive. A real good simulator that captures the physics of swerve could be used to help train drivers. That's what the military does. Right now I believe that a reliable IMU based on inexpensive MEMs sensors is almost there. This summer I think we will look at the IMU' and algorithms that just came to market to see if they will work.

Re: The Perfect swerve
 

Not to change the subject, but is there any info about this swerve aside from the gallery on the Google+? Like maybe a disassembled version? Or a discussion here about it?

Re: The Perfect swerve
 

Presumably all the designs that are being reused over multiple years are publicly available somewhere per R13, right?

It would be a great thing if someone assembled links or other references to all of them in one place. Has anyone attempted such a thing?

Re: The Perfect swerve
 

YES!
I think that that the perfect swerve drive needs to be short, we suffored too much from a tall frame (even though we did quite well after we fixed the bugs), making the drive small and reliable are the most important.

Re: The Perfect swerve
 

It's a hobby of mine to design gearboxes and strange drivetrains, because those are things that we can count on being applicable each year mechanically.
Through extensive research into swerve drives, I have seen two swerve drives which in my opinion are supreme:
-Team 221's Revolution Swerve
-Team 2451's Cim-in-wheel swerve
I like the Revoltution because it is the most compact solution to a coaxial swerve I have ever seen. It's pretty much optimized for its design. For that reason alone I basically gave up on coaxial swerves because I decided if I wanted to make a coaxial swerve, I would just copy Team 221's CAD and modify it for smaller wheels (this is the route I think 368 took this year, or maybe they're just very similar) or just buy some from them. If we needed to have a lower gear or didn't care about COG, I would go coaxial swerve and turn the modules using 4 different motors with two centralized gearboxes like 3-cim ball shifters. Plus, because they are available from 221, it's extremely easy to implement one in a FRC robot in a short time (although still not during build). For example, 368 was able to get their swerve off the ground in one year according to their mentors.

Team 2451's Cim-in-wheel swerve (I forget the actual name) is based off of the "swerve god" Aren Hill's Cim-in-wheel swerve. The only reason I put 2451 as the team is because they released their CAD so it's easy to look at and modify. It is the best all-in-one module I've seen including all those shifting ones because of its compactness. It is the most space-efficient design I have seen, and there is even room for improvement to boot. If you don't need the shifting and care a lot about COG, this is a very good solution.
Team 221 has a picture of a Cim-in-wheel swerve (hollow pivot wild swerve), but the CAD for that product is just the regular Wild Swerve witha hollow pivot- the cim is mounted above the wheel.

Shifting all-in-one modules are pretty pointless IMO unless you have a lot of resources. This is because unless you really care about COG, a coaxial swerve makes shifting half as hard. Plus, in order to really make good use of these modules you need to lighten them as much as possible by boring out gears and plates and wheels... for most teams I think a coaxial swerve is the best solution.

I talked with an old mentor of ours who has done a lot of engineering, machining, and FIRST for a looooooong time (~25 years for the first two, 12 for the last). According to him, the hardest part of a swerve drive is the programming. It's for that reason that he heavily advises against our team doing a swerve drive, because he thinks that we would need professionals to program it for us. His opinion is based off of our team's sad efforts to build a swerve drive years ago.
I don't agree with that completely, because many teams have graciously released swerve code and whitepapers on the subject, so you don't really need an engineer to do it anymore because the hardest calculations are already written down. I think any team with interested and smart programmers can implement even a rudimentary swerve drive at this point.
However, we have a completely assembled crab drive (mistake #1 was to crab) sitting in our shed that weighs something like 80lbs from several years ago. We took it into our room in September to show the rookies, and I remember having trouble with the think because it was like a rock- solid, hard to hold and extremely heavy. It never drove because it was too hard to program. Nowadays there is a lot more info out there plus far lighter swerves that I want to use someday.

I remember Aren once had an idea in a paper of his about using a swerve drive that implemented a wheel tilted at a 45* angle to the ground. I don't think it was ever produced though.