*There seems to be fairly wide agreement here on CD that “swerve drive” and “crab drive” mean different things.
Swerve means each wheel is independently steered and driven.
Crab drive does not have independent steering for each wheel. Typically all four wheels are steered in unison to the same direction.
In FRC, the early steerable drives were crab variants. Swerve came on the scene years later.
Surprised nobody has mentioned 1625 Winnovation yet. Ran a few super neat swerves. Aren’s brainchild 3928 also ran one previously. The CIM-in-wheel design was quite cool.
11 ran a half baked swerve back in 2010…it was not prototyped in the offseason and not enough development time went into the controls or drive practice.
Ask me after the season is over :rolleyes:
There are times where I think ‘wow swerve really helped us that match, neat.’ Other times I think ‘why did sink so much time and effort into this drive?’
It’s all a trade-off and we’ll be doing a post-season review to objectively evaluate where we want to go with it for next season.
Nailed it! It’s 100% trade off city. There is no doubt in my mind that it takes time and resources away from other tasks. That being said, we like it and hard to imagine us not running it given that we’re worse at making tank drives at this point.
Ironic fact: Both of their crab/swerve modules were mass produced for a while (Wildswerve for those who wanted motor-in-module, 118’s for those who wanted coaxial drive). The production outlived the usage by the respective teams. Go figure.
2005-2012 was the original run. As soon as they went crab/swerve, they stopped winning championships :D… Ok, not really. They nearly won 2007 partially because of their drivetrain.
Bill Beatty came back later for the 2015 (?) season and with it came crab/swerve, at least initially. But 2016 and later they’ve been running 6WD.
Dillon from 1625 posted a whitepaper about that awhile back (I’m sure you’ve read it already)… this one. Basically they went all-out for FRC’s (as far as I know) only 6WD crab drive ever (2 speed too, that was a big thing back then) in 2012, found it most useful for power-strafing (against 469 that year), and went for the non-crab but still genius lobster drive of 2011.
When you’ve swerved too long, there’s no getting straight.
Specifically, the WildSwerve module was exactly the modules contained in the telescoping swerve drive from our 2004 bot.
111 hasn’t actually ever run a true swerve drive. We tend to define the line between crab/swerve as independent control of the wheel headings, which we’ve never had. Closest we have come was independent front/rear wheels in 2008 in order to achieve ‘car’ steering’. We did build a prototype swerve base 2-3 years ago, but have since abandoned it due to software resources.
We ran modular cantilevered 6WD drive trains on our 2011 and 2012 bots. This concept was ditched due to weight management issues. Our 2017/2018 bots are essentially 100% COTS WCD’s.
Common misconception: WildStang 2005 wasn’t swerve, but rather a tank drive perpendicular to the ‘front’ of the robot.
Robonauts 118 ran swerve drive from mid-2000’s through 2008, half-swerve in 2009 and then tank drive from 2010-present.
Robonauts are still running swerve 
Modular Robotic Vehicle
Space Exploration Vehicle
we’ve wanted to do swerve for years, but for various reasons haven’t had the resources or bandwidth to get a drive base going. We keep falling back to mecanum when we need to strafe sideways, and it generally presents a more compact drive train layout.
The team members didn’t want mecanum this year because it’s “too slow”, so we organized a drag race between our Steamworks robot (mecanum) vs. our Stronghold robot (4 CIM 6 wheel drive, a very fast robot). The mecanum robot was not only faster, but easier to drive as well.
We will still prototype swerve at some point, but I doubt it will win a trade study against mecanum (just my opinion).
Cheers,
SS.
Is 968 currently using a tank drive? They looked like tank in their last 2 regionals, but they have swerve at Utah.
We’ll stick to our 2 speed 6 CIM WCD. We appreciate the power.
Watch them at Utah again. I just did.
If that’s a swerve drive, the modules are locked. Not a hint of swerve motion in early or late matches.
Matter of fact, I believe the photo from their TBA pagereveals their actual drivetrain: 6WD flat with omnis on both ends. (L.A. location for the photo.)
I looked at their drive mechanism up close one year – it was a thing of beauty. I hadn’t realized until I saw this post and checked Blue Alliance that the team isn’t around any more. What a shame. Anyone here know what happened to it?
Several folks here know what happened to team 1717.
It was fun while it lasted! We had the honor of being next to them in the pits at their (and our) very first competition.
Things were a little different that year
From what I understand, their school used FRC as a ‘senior thesis’ of sorts – the seniors (maybe others as well) built the robot with immense support from the school as a project that was a part of the curriculum. After the 2015 season the school wanted to grow the program and realized that FRC was not scalable to include so many people involved at a very high level of the engineering, so they opted to switch to a mechatronics project instead.
Someone with more knowledge can elaborate.
They “retired”, and now do a mechatronics-based science fair instead. That science fair is very prestigious, and while not as glitzy as FRC, isn’t necessarily a downgrade.
Team 5818 (Rivera Robotics) was formed as a community team to fill in for the hole left by 1717 (and my understanding was that at least initially, there was some overlap in mentors and students when the new team formed?).
Against my advice our team did swerve drive last year. It almost worked right, just enough to make it our worst season ever, including our rookie season.
This year we used the kitbot tank drive (following my advice) and went 9-2 in qualifiers.
I kept one of the swerve modules, just in case someone wants to try it some time in the future–it’s heavy enough to be good for clubbing them about the head and shoulders until they admit it’s a bad idea!
The summer after it was announced that 1717 would move on from the FIRST robotics program, 2 students, Kally & Amy, that were to be on the team the next year (1717 only had seniors on the team) took it upon themselves to start their own team. I had mentored on 1717 a few years earlier and offered my support. With the help of a 3rd student, Graham, they formed a non profit corporation, got funding from the community, and signed on a great group of students for the teams first year. A couple other mentors, Lennie & Doug, from 1717 came on to help out, and Ty from 3512 also was a great supporter.
Jim, that is a great picture, neat to see the beginnings of 1717.
I thought it would be interesting to hear about 5817’s experiences doing swerve their first 2 years, especially as rookies!
If I remember correctly 4920 had a swerve for the first event of their 2016 season and then later in the season replaced it with a kitbot drive train. If that’s not a good example of abandoning a swerve drive I don’t know what is.
It’s kind of interesting how rare swerve is in Ontario. Their were some notable swerve’s early on (1114’s rookie year) but it’s pretty much died out since then. Most areas have a perennial “swerve team” (1323/1717 in Cali, 1640 in MAR, 2767 in FIM, 16/2451/2481 in the midwest) but Ontario has been pretty lacking, particularly in the last 5 years. This could be seen as a good or bad thing, potentially a win for strategic aptitude of the average Ontario team but it also makes me sad how some students may never be exposed to this unique drive train style at competition.
With the more recent advent of things like the Talon SRX, 775 pro and the mag encoder I honestly believe that there is some reasonable competitive merit to becoming a “swerve team” assuming you develop and practice for it in the offseason and use it every year. FIRST seems to be trending towards games without the need for two speed drive train. I could imagine that being able to easily adjust the dimensions of your drive base could come in really handy strategically at the start of the season.
The Talon SRX just makes it so much easier to make and tune robust closed loop mechanisms. Over the past couple of years I’ve actually had more luck with motorized closed loop mechanisms than I’ve had with pneumatic ones. Teams like 4678 in 2016 have been building multi jointed arms to great effect that actually end up simpler than the traditional more mechanical solutions
If someone like WCP were to come out with a robust integrated swerve module I could actually see it become fairly prominent in FRC among high resource teams.
Re closed loops on talons.
In truth - I very seriously contemplated rolling my own control loops to more effectively test the entire system virtually. The talons being closed source I can’t accuretely model their outputs while in close loop modes in a virtual environment which limited the testing I could do on software.
In the end I just treated them as black boxes and tested the functions that generate their inputs so I could skip out on modeling but it did limit what sort of systems I could easily test. I may revisit this in the offseason but that remains to be seen.