Mecanum vs. Tank Drive?

[svince]

Our team is currently trying to decide between using a mecanum drive and a basic tank drive from the KOP in order to effectively make a cycle for placing gears. Would mecanum be a viable alternative because of the precision boost, or would tank style overpower us in our runs across the field to the point where it isn't worth the change in style?

[euhlmann]

Mecanum
Pros: precise maneuvering
Cons: requires precise robot mass distribution in order to work. Also harder to program

Tank
Pros: easy to build
Cons: go up to feeder station. You're not aligned. Drive back and turn. Go up again. Still not aligned. Drive back and turn again...

[RoboChair]

Tank.

[aguzauckas]

Originally Posted by svince

Our team is currently trying to decide between using a mecanum drive and a basic tank drive from the KOP in order to effectively make a cycle for placing gears. Would mecanum be a viable alternative because of the precision boost, or would tank style overpower us in our runs across the field to the point where it isn't worth the change in style?

I would strongly suggest to go with what you are stronger with. If your team can code and drive mecanum, then that would be a great idea. There are downsides to a tank drive as you would have to reverse and go back forward if you aren't aligned correctly.

[Taylor]

Great question!
There are two schools of thought regarding mecanum wheels. One is that mecanums are an easy way to gain an extra degree of freedom when driving (easy as compared to swerve). It does have a bit more complexity than a standard 4 or 6wd setup - more refined versions of it require use of a gyroscope. However, this does allow a team to get many of the translational properties that swerve allows, at a reduced cost and a fraction of the resources (labor, motors, programming, electrical).
Another view is that mecanums are a lesser drive system - by virtue of the rollers on the wheels, they are subject to heavy defense, and cannot play defense themselves. This view is usually the more loudly proclaimed, usually by people who heard this view from somebody else, and by people who have no direct experience with mecanums.
I will say that a number of high-end teams have experimented with mecanum drives and felt they have found a better solution elsewhere.
I will also say there is a stigma with mecanums, with questionable merit, and some teams say they would not pick a mecanum robot to ally with in eliminations. However, I have noticed that the people who proclaim they'd never ally with a mecanum robot, aren't usually in a picking position.
The general rule of thumb is if somebody doesn't know how to spel "mecanum" then their opinion is mostly based on hearsay.
Welcome to Chief Delphi!

[AndyBare]

A tank drive will give you the ability to literally push your way across a field and hold your ground if you need to. Mecanum is a really good drive system if there's a reason. The field this year is flat, making mecanum a viable drive system to pursue. With limited visibility though, there is a possibility of you getting into situations such as being pushed around (perhaps while on the way to get fuel or gears), and not even recognizing what's going on. 2014 was a strong year for mecanums because the field was flat and clear. 2015 was a good year because there was no defense. This year, on the other hand, I'm not so sure. My vote is for tank.

[Tom Bottiglieri]

[FrankJ]

The stigma against mechanum is very real and a bit undeserved. You generally going to have to seed well or be an alliance captain to play in the afternoon.

Mechanum is very good for precise aiming and locating for tasks like placing gears.

Mechanum will do poorly against a properly geared high traction robot. Varying results against others. On the other side of the coin, you can't be pinned by being pushed sideways like a high traction bot can. You can use defensive strategies that involve deflection rather than pushing matches.

Effective use of mechanum takes a different mind set than tank drive. Key is lots of driver practice in a full size space. But then effective use of tank drive requires lots of practice in a full size space.

[Lord Basket]

The biggest downside by far to mecanum drives are that they can and will get pushed by everyone. In a game like Steamworks where there's an open field which most robots will regularly have to traverse, and where teams without functioning devices have nothing to do other than play defense, that one downside overshadows any and all minor advantages that mecanum drives have.

[svince]

Originally Posted by FrankJ

Mechanum will do poorly against a properly geared high traction robot. Varying results against others. On the other side of the coin, you can't be pinned by being pushed sideways like a high traction bot can. You can use defensive strategies that involve deflection rather than pushing matches.

I've heard that we can use gearboxes with a different ratio to increase our force, but is there any truth to that rumor?

[TheNerdJedi]

Do a decision matrix with members on your team. List various traits you want your drivetrain to have, like "mobility", "speed" , "complexity" , etc. and compare the various drivetrains in question. Assign a number for each trait to each drivetrain and see which drivetrain has the highest at the end.

Think about your team capabilities and what would be best for the team.

[Ben Martin]

tank

[Kevin Sevcik]

Since our team has used mecanums a couple of years, I'll chime in.

First, exactly centered weight distribution isn't 100% necessary. Things will go wrong if your CG is 4 times closer to one pair of wheels than another, but a CG that close to the edge is terrible on any drivetrain. Follow normal CG location principles and you'll be fine.

Flat floors matter somewhat more. We used mecanums* in 2011, and the undercarpet plywood support for the towers would often often skew us when we hit it crooked. Closed loop yaw rate control will help a lot here. (*Technically half octocanum. Only 2 sticky wheel modules because weight.)

You're not going to get pushed entirely across the field by every robot, but you're definitely going to lose to heavyweight well designed tank drives. You do need to account for this in your drivetrain evaluation. Expect it to add extra point-to-point seconds to your cycle relative to a tank drive. On our 2011 bot, even just the two extra sticky wheels helped. The driver would drop them to stop a sideways shove and power out of it. They didn't get fast and good at this till Champs.

Since you're probably pitting a 4-CIM mecanum vs a 6-CIM tank drive, the mecanum is going to have less push and acceleration even before tracion comes into play.

Mecanums are not actually complicated to program. There are labview, C++ and Java blocks/classes for open-loop mecanum control. I know there's also C++ and Java code for closed loop yaw control and field centric control as well.


All that said, I'd still go with a 6-CIM or 4-CIM + 2 miniCIM tank drive. It's pushier and has less moving parts and breakables than mecanums. It will inevitably get you where you want to go faster. If you need a system to quickly line up a gear without backing up, add something to slide the gear side-to-side on your robot. Moving the gear side-to-side for alignment is the goal. You don't have to move the entire robot to do so.

[GeeTwo]

Originally Posted by euhlmann

Mecanum
...
Cons: requires precise robot mass distribution in order to work. Also harder to program

Not really, on either. In 2013 we did mecanum. Using the canned mecanum drive class, it was no more difficult than a four motor tank system. (OK, an extra joystick axis). We ran it on our prototype robot, "woody", which had at least 4x the weight on one corner wheel than the opposite, but within a half hour of driving practice, drivers had learned to compensate so well that they weren't doing it consciously.

Actually, woody was the only robot we have ever built that I have over an hour of stick time on, due to a demo I did at work. All I recall was having to account for a slight pull to one side.

[ldsedam]

I would strongly recommend not using mecanum in any game where defense will have the potential to have an impact. The ability for a robot to be completely shut down by defense with little effort is a huge drawback. There is a reason why there has never been a mecanum drive on Einstein (Except 2015).