Our team has been brought up a couple times as one who did mecanum well this year. Teams that didn’t get a chance to play with us have asked to see a video. Here it is! Hopefully after you see this video mecanum won’t be such a "dirty word
The first part of this video focuses a little more on our pushing power and the rest are clips from matches that really show off how we used mecanum to help us score. Please watch the whole thing so you can see all our moves. I promise you won’t regret it Let us know what you think and post time stamps of your favorite parts!
If you’ve posted it elsewhere, please direct me, but what did you do differently that meant you did mecanum well? Was it placement of center of mass, some sort of suspension on the wheels, specially stiffened chassis, gearing, control scheme, or something else? It’s a great video, but it doesn’t give much insight into your design.
I’m clueless on the specific design, but I’d toss driver training in there. They’re definitely awesome via the video! The individual movements don’t seem so different (perhaps pushing better axially?), but in terms of outcomes, I’d venture much of the performance benefit is in tactical usage. Is this a reasonable characterization? (It may also have a special control scheme element, and is certainly reliant of well-executed mechanics.)
What happens when someone catches you sideways? I saw a few spin-offs, but sometimes you seem to prefer turning rather than crabbing, even when it exposes you to transverse pushing.
Driver training is a huge part of it. Our driver was a veteran and had a lot of practice with mecanum. Great job Matt! The specific clip you reference shows just one example him getting into the feeding station as fast as possible. He has learned to go with the push and spin while still providing force in the direction he wants to go. This is not always possible with tank. What we have found is that when you push a tank robot sideways the only thing stopping them is their static friction with the carpet. When we get pushed sideways we can provide a force in the opposite direction to the push. I can’t site any equations as I don’t have friction coefficients but I would bet we are at least as hard if not harder to push sideways when we are strafing into the hit than a tank drive. That would allow us to play defense while facing different directions than tank drive bots. Note I’m only comparing us to tank here and not swerve as far as being pushed sideways (we were generally an offensive bot so we don’t really have footage of this)
…Are you saying you don’t slip your wheels (or rollers) when you strafe into a colson/nitrile tank drive?
I’d take those equations: a tank’s going to have basically double your CoF. AM lists their mecanum as 0.6 sideways (0.7 axial), and we’ve got nitrile at around 1.2 AM puts HiGrips a bit lower, while VexPro puts Versas right there. Colsons are even higher–I recall people modifying colsons to like 1.5. (EDIT: colson link)
I think this is one of the biggest takeaways here. Notice in the video that we never get into a pushing match per se; the robot is always still moving in the direction we want to go. The mecanum wheels allow the robot to provide force in the opposite direction we are getting pushed (thus preventing us from getting slammed into the wall and stuck) while simultaneously slipping away towards our target. We are strafing in a circle around the pushing robot.
This provides a clear benefit, as there isn’t much a tank drive robot can do to counter it. They can’t rotate and keep pushing us nearly as well as we can slip away from them, and it would take a lot more maneuvering for them to reposition themselves in front of us as we slip away.
Our reaction to a head-on pushing match is never “keep pushing back and hope we win.” Our reaction is to slip out and around the robot.
We don’t have a practice field or correct carpet so we haven’t done any of the specific tests you’re talking about. All we have is evidence from match videos that we are tough to push sideways when we apply force in that direction. We also haven’t really investigated the different material of wheels of robots after they attempt to push us. I would guess that yes our rollers do slip some as we drive but sideways into a robot trying to push us from the side. We’ve never got into a direct side pushing match where someone started to push us and we applied a force and still slid, which would get our wheels into more of a kinetic friction.
That actually just made me think of another point. We were playing defense once (we don’t do it too often remember). But a robot was using a ramming technique to break our static friction and every time they backed up to ram we drove sideways towards them and actually gained ground on each hit haha.
I’m not claiming mecanum is better than the higher traction tank drive bots at sideways collisions, they’re almost certainly not. What I was saying is that it’s tough to add in force of the mecanum in the sideways direction, along with static friction of the rollers, along with slipping of the rollers, and get a correct equation for mecanum because I don’t have those numbers. I bet every team has a different slip in their rollers making it hard to calculate as well. I do think that our mecanum was better than many of the robots we faced at taking sideways hits. Some of the mecanum bots were much, much worse at taking sideways hits than tank robots so that’s not a blanket statement. Those robots are probably the ones people picture when they think of collisions as well. Maybe in the off season we can do some testing
I’m puzzling over what your intended implication is.
Here’s what I think you are implying:
It is harder to push a mecanum robot sideways when said robot is trying to resist the push by attempting to push back than it would be if said robot simply had its wheels (wheels, not rollers) physically locked.
Have I correctly inferred your implication, or have I misunderstood?
Sort of. I understand that really the only force resisting is still static friction of the carpet and the wheels. Our force we apply with the motors I feel cancels out the slip of the rollers, leaving still static friction with the carpet. Which makes us just as good as tank if they have the same CoF.
Again the benefit of that is when a robot breaks contact with us to try to ram we continue to push against them and we back up as they do, not allowing the ram and gaining ground.
You are the numbers expert here Ether, feel free to give some numbers!
Our force we apply with the motors I feel cancels out the slip of the rollers, leaving still static friction with the carpet. Which makes us just as good as tank if they have the same CoF.
Given a tank drive bot with the exact same tread material on its wheels as are on the rollers of a mecanum bot, said tank drive bot would have better traction than the mec bot (assuming equal weight of both bots). The reason for this is explained here.
If my understanding of a wheel’s mechanics is correct, static friction from the wheels is the only force ever really pushing the robot, and slip occurs when the force from the motors exceeds that static friction. That contradicts what you’re saying here:
Would that be correct? It seems counter-intuitive that force from a motor could overcome slip – I’m imagining Lunacy-style wheels, where more force from the motors just meant more slip.
You are correct in that static friction is the only force pushing the robot. maybe slip is not the word I’m looking for. What do you call when the rollers spin, not the wheel itself. Slip of the ROLLERS is different than static friction being broken between the rubber and the carpet. When I use the term slip for the rollers I mean the rollers free spinning.
I don’t see how to make this a ‘cancel out’ situation: how does it overcome the phenomenon that mecanums break traction before any tank drive with at least the same* CoF and weight? If the issue is slipping (which is why I asked), it’s not CoF vs CoF issue; it’s the geometry of the wheels. The fact that mecanum CoFs are then 40-75% that of most tanks only serves to compound this.
Roller slip will vary by specific DT, but not nearly as much as between mecanum and tank in general, and it’s not helpful in strafing. It’s also (thanks to Ether), not overly difficult to model the description, at least to the point of illustrating its negative effect. As the paper explains, the friction that keeps rollers from slipping adversely effects mecanum’s strafing traction. If this is what we’re talking about, I don’t understand the implication that driving the wheels produces more tank-like results.
All very good points! I don’t claim to be a physics expert. I am not. That’s why I asked Ether to give me some numbers ::rtm::
Again we were never trying to push in the sideways direction. Our video does not show us winning too many big pushing matches. Mecanum allowed us to avoid those situations. We did not choose mecanum for its pushing power haha, nor do we generally use it for such things. The video is just meant to show some examples of how we used mecanum to our benefit this year, and to hopefully show it’s not as weak of a drivetrain choice as it is in most CD’ers minds. No one (hopefully) chooses mecanum for traction purposes. We just wanted to highlight some of our match videos because people cited us as a good example of mecanum.
No problem. I think the discussion just got sidetracked because I (perhaps also Ether) was worried that teams–who would otherwise be influenced by the video in very positive and correct ways–might take these statements the wrong way:
Particularly the clarity of the last one, as it’s contrary to the actual geometry of mecanum wheels.
Also, I seem to have distracted the thread from Mr. Elston’s very good questions:
Great discussion here and thanks for everyone’s input.
KnightKrawler used 2 AndyMark Nano Tube 20’s with hex shafts in the stock 12.75:1 ratio. One CIM per wheel and each wheel driven directly from the gearbox. Since we need a drive motor per wheel, we used this as an opportunity to leave out chain. As noted before, the wheels are AndyMark 8" HD mecanums that come with side rollers which we do think help to smooth out the movement of “Silverfish” as it transitions between mecanum rollers. The robot weights in just under the 120 lbs. limit with the 84" blocker and GoPro installed. I bet “Silverfish” was just over 118 lbs. for its fighting weight.
While we did develop field oriented drive code this year, it was not utilized. The robot is driven in open-loop as we do not use any sensors on the drivetrain. We did have a gyro on board and tinkered with auto robot orientation for loading and hanging, but it was never used in the competition code.
Here is a picture of the drivetrain: http://i233.photobucket.com/albums/ee153/natecmiel/drivetrain.jpg](http://s233.photobucket.com/user/natecmiel/media/drivetrain.jpg.html)
I hope this clears up some of the questions, let us know if you have any more.
Thanks,
Nate
2052 KnightKrawler Mentor
Does anyone know if there is an engineering drawing available – or even just a concept sketch – which shows how the axial forces on the rollers are reacted in this design?
Let us know what you think and post time stamps of your favorite parts!
