Mecanum vs. Swerve Drive

[cadandcookies]

Originally Posted by jbsmithtx

mecanum wheels are a great and easy option, provided you build them right.

Isn't this true of almost any drive train (or system in general?),

If you do it right and know the limitations you're working with, you can be successful with nearly any drive train.

That isn't to say success doesn't come easier to certain methods and systems.

[efoote868]

Originally Posted by sentientfungus

Isn't this true of almost any drive train (or system in general?),

If you do it right and know the limitations you're working with, you can be successful with nearly any drive train.

That isn't to say success doesn't come easier to certain methods and systems.

There are many adjectives I'd use to describe swerve drive trains, but easy is not one of them.

[EricH]

If I needed omni-directional motion, and I did not have a swerve drive prototyped already, I would probably turn to mecanum--and adopt a strategy that used lots of offense with some "trick" defense. I bet that some of those FCS blockers could have been pretty effective on a mecanum base, for example--but so could have been the FCS, for the same reasons.

However, if I had a swerve drive prototyped that I was "happy" with--that is, that I figured was pretty close to where I'd want it for competition--I would take that over the mecanum.

Of course, I'm assuming here that I actually need omni-directional motion, but that's a discussion for another day, say a couple of days after Kickoff.

Here's what I see as the advantages for each side, just mecanum vs swerve:
Mecanum:
-Potentially lighter
-Easy to turn into 4WD should I decide at the event that I don't need omni motion
-Fewer moving parts, so less likely to break
-Somewhat less control complexity than swerve, on the feedback side mainly

Swerve:
-100% of power goes in the direction I want, no vectoring.
-several ways to set up, each with its own pros/cons (somewhat easy "tuning")
-more possible variants--witness 1625's 6-wheel, the 16/148 3-wheel, and 3928's CIM-in-wheel in addition to more "traditional designs"
-looks cooler

[Kernaghan]

1114 has really good information on drivetrain designs and advantages here.

[Chris is me]

Originally Posted by LeelandS

The only reason I can see not picking Mecanum for a third robot is if the team is trying to push with them. If a team builds a robot on Mecanum wheels and is pushing as part of their strategy, then I really wouldn't want them for a third robot.

How often are you picking a third robot with the intention of running an identical strategy to the one they ran in quals? Yes, sometimes it's similar, but there's a lot to be said for flexibility.

The thing is, mecanum isn't only worse than tank drive at pushing. More importantly, it doesn't resist being pushed very much. For a lot of defensive roles (including most of the "creative driving" roles you allude to in your post), if the defender can just be plowed through, the strategy doesn't work.

Here's an example situation. We had two decent choices left for the sixteenth pick at WPI. (The rest had <4 CIM drives, <100% powered wheels, or were not consistent at anything we needed) One was a mecanum drive robot with a full court passer (no scoring) that, while unreliable, could really increase disc throughput for our ground loading partner. The other option was a shifting tank drive with a 10 point hanger and a smart drive team. Yes, we could gamble and pick the unreliable mecanum shooter... but if they jam, what do they do? They can get in the way, but not enough to make much of a real difference in the final score. The ten point hanger could play any kind of defense needed - man to man, line defense, cover-the-disc defense.

Neither of these teams were running the strategy we wanted from them for most of Friday. One was a lot more versatile than the other despite having way less going on simply because they could resist being pushed and push well.

I won't go so far as to say I will never select a robot with mecanum wheels - but it will be despite their presence on the robot, not because of it. I've even seen mecanums play decent defense - it's just that the wheels are no help at all.

(by the way, despite my Mecanum Hater reputation here, I do think 2013 was an *okay* year for mecanum, and teams like 1100, 4134, 3824 played to its strengths)

[SoftwareBug2.0]

Originally Posted by Nemo

6WD pseudo code:
M1 = y + x
M2 = y + x
M3 = y - x
M4 = y - x

Or you can use this pseudo code:
M1=y1
M2=y2

[Gdeaver]

Over the years I have only seen a few Mecanum bots that have true omnidirectional movement on the field. They could strafe at best. We love to have our way with them. There have been many times where we had to defend a good scoring Mecanum on our alliance from the opposing alliance defense. I have yet to see a Mecanum on the field have control of chassis orientation and drive direction. I would choose a well developed 6 or 8 wheel tank over Mecanum any day.

[efoote868]

Originally Posted by Gdeaver

I have yet to see a Mecanum on the field have control of chassis orientation and drive direction.

How is this not a problem with swerve drives?

[Taylor]

Originally Posted by Gdeaver

Over the years I have only seen a few Mecanum bots that have true omnidirectional movement on the field. They could strafe at best. We love to have our way with them. There have been many times where we had to defend a good scoring Mecanum on our alliance from the opposing alliance defense. I have yet to see a Mecanum on the field have control of chassis orientation and drive direction. I would choose a well developed 6 or 8 wheel tank over Mecanum any day.

The problem is not with the drive train. The problem is with teams that don't properly know how to use it.

Mecanum is a very easy system to implement and use. Code is readily available; we had ours built and programmed within a day when we used it in 2011. This makes it a viable option for low-budget, low-experience teams looking for a quick fix.
In my experience (and this is a broad generalization - I'm not talking about any particular team), teams that use mecanum drive aren't as advanced regarding gameplay strategies. Driving is all about anticipation. Most mecanum-based offensive strategies are all about avoiding defense, not getting into pushing matches.
Mecanum systems aren't as good at pushing because they don't need to be.
Also, if a team does find itself in a pushing contest and is outmatched, it needs to find a way to disengage without pushing back. Mecanums, if used properly, can do this well. Unfortunately, most teams don't utilize this*.

On the other hand, most teams that use swerve drive, and have iterated it to the point of usefulness on the field (I've seen more bad swerves than I have bad mecanums), have a degree of experience and wiliness about them. They're usually well-practiced and well-coached so that they don't get into bad situations in the first place, and if they do, they can get out of them with ease.

*Watching matches, I'm often reminded of the cartoons where Wile E. Coyote is trying to outrun a rocket. He could simply dive to the side and the rocket would fly by, but that never occurs to him.

[Siri]

Originally Posted by efoote868

There are many adjectives I'd use to describe swerve drive trains, but easy is not one of them.

Yeah, that'd be a matter of "they're easy provided you build them right" (true) vs "building them right is easy" (false).

Originally Posted by efoote868

How is this not a problem with swerve drives?

Chassis orientation is controlled independently of translational direction. e.g, Just push in the defender (because they can't push you), spin around them, and then spin around again until you're lined up the way you want. Like this or this. Or after a brief pin. Or before not being pushed at the unprotected feeder. Or after rocking an FCS blocker.

Taylor - absolutely agree. It's not a theoretical issue, it's a game play one. For instance, there's no way we'd be putting all the swerve work in at this point (nor will we continue to), if it wasn't providing such a huge in-game advantage. Still, I'd think that after 7 years I'd have seen someone riding mecanum that really embraced this. Is coaching and practice really bringing this to (near? video?) zero? I've seen some really excellent drive teams on them, but they just don't seem to weigh this heavily. I know a lot of the biggest benefits for chassis orientation rely on holding traction, but I'd venture others do not. Is the default (or otherwise reasonably doable) code amenable to it?

[Andrew Schreiber]

Originally Posted by Chris is me

How often are you picking a third robot with the intention of running an identical strategy to the one they ran in quals? Yes, sometimes it's similar, but there's a lot to be said for flexibility.

Err hmm, DC 2013. 383 was a Mecanum bot picked to play 3 point shooter. Exactly the role they played in Quals. I think it worked out pretty well for us

Regarding not being picked in 2012 with mecanums – This was a function of their difficulty balancing.

Honestly, this debate can be decided pretty quickly.

Does your strategy dictate that an omni directional drive system is needed?

Does it dictate that you need absolute traction in all directions?

With those two questions you can decide if you need tank, swerve, or mecanum. Remember, strategy needs to be dictated by game play and resources available to your team. If you lack the resources to build a swerve your strategy should not ride on one. If your strategy doesn't ride on it you shouldn't build it

[Madison]

Originally Posted by Gdeaver

Over the years I have only seen a few Mecanum bots that have true omnidirectional movement on the field. They could strafe at best. We love to have our way with them. There have been many times where we had to defend a good scoring Mecanum on our alliance from the opposing alliance defense. I have yet to see a Mecanum on the field have control of chassis orientation and drive direction. I would choose a well developed 6 or 8 wheel tank over Mecanum any day.

This, in my mind, is a fundamental component in the debate about the effectiveness of mecanum drive in FRC. It's difficult to make a determination about how effective they can be when a significant number of teams that implement them do so very, very poorly.

The conventional wisdom has been that implementing mecanum has become easy. The wheels are available as COTS parts, the code is shared freely, etc. The reality has been, though, that there's a significant performance discrepancy between the few good implementations and the piles and piles of really terrible ones. That suggests that mecanum really isn't that easy to get right and we should probably be putting it into the same category as WCD or swerve -- perhaps still not in terms of effectiveness, but certainly in terms of knowledge required to do it right.

[avanboekel]

Originally Posted by Madison

This, in my mind, is a fundamental component in the debate about the effectiveness of mecanum drive in FRC. It's difficult to make a determination about how effective they can be when a significant number of teams that implement them do so very, very poorly.

The conventional wisdom has been that implementing mecanum has become easy. The wheels are available as COTS parts, the code is shared freely, etc. The reality has been, though, that there's a significant performance discrepancy between the few good implementations and the piles and piles of really terrible ones. That suggests that mecanum really isn't that easy to get right and we should probably be putting it into the same category as WCD or swerve -- perhaps still not in terms of effectiveness, but certainly in terms of knowledge required to do it right.

I completely agree. But why? On paper, it should be an easy drive to execute.

Edit: What I'm trying to say, is what's the difference between a good and bad mecanum drive?

[Boe]

Originally Posted by avanboekel

I completely agree. But why? On paper, it should be an easy drive to execute.

Edit: What I'm trying to say, is what's the difference between a good and bad mecanum drive?

Driver Practice.

[bbradf44]

Originally Posted by avanboekel

I completely agree. But why? On paper, it should be an easy drive to execute.

Edit: What I'm trying to say, is what's the difference between a good and bad mecanum drive?

In my experience, a good mecanum drive tikes time to perfect, not nearly as much time as swerve though. My team used them for the first time in 2013, we ended up having to re balance our wheels and straighten them out and fix programing on several occasions. In the end we did have very very good but un traditional mecanum drive (un traditional on the control end). At our regional we saw some good examples of bad mecanum drives where they were basically tank drive, they could go forward, reverse, and turn but couldn't strafe to save them.