Swerve Drive: Pros and Cons?

A robot that is spectacularly manoeuvrable and cannot score well will still be sent to play defense if it gets picked.

This reminds me of something that I wanted to bring up. I’ve had a few people tell me that swerve is inherently bad for defense. Even including one team who’s robot was swerve drive.

I’m not sure why this would be. Perhaps it is because swerve drives are often geared pretty aggressively such that they don’t have the highest pushing power, but there is no fundamental reason why swerves would be any worse at defense than a typical drive train. In fact the ability to move any direction without needing to reorient can be an advantage while playing defense.

4 Likes

For reasons that I won’t get into, we went to one of our off-season events with nothing more than a swerve drivebase (with some extra ballast added) and therefore could only play defense. It proved to be an excellent defense configuration. The nimbleness of the drivebase in terms of being able to drive in any direction at any time allowed us to easily get between scoring robots and the scoring locations and then stay between them as they tried to get around us. And I can tell you that we were very successful at pushing back on teams that tried to push through us. And many times we were able to knock scoring elements (hatches or cargo) out of other robots through “aggressive” bumper to bumper contact.

We were geared for speed (6:1 to 4" wheels from NEOs), but were still able to push around other robots. There were several times where we pushed other robots sideways where their drive wheels were skidding sideways on the carpet and really had no issue with power to push them. it may have been all that pushing power we put in the power cube that was bolted to our frame…

4 Likes

That may have been our team that said we couldn’t play defense with a swerve drive. We were really pleasantly surprised how well playing defense with swerve worked.

We switched to 2910s swerve this year and we really loved it. It gave us some flexibility in design since we could put mechanisms on all 4 sides of the robot. We had some fantastic software kids that made the switch easier.

Watch the lower left display at 6:32 and 6:55 to see how playing defense against a swerve can be problematic.

JMH

2 Likes

I don’t doubt that a well driven swerve robot can be very effective at defense. It appears that most teams wanting to use swerve are wanting to improve their scoring ability. Only directing extra resources to a drive base will not improve lackluster scoring mechanisms.

Your speed also contributes to your effectiveness in defense. If I recall my physics classes correctly, kinetic energy = (M * V^2)/2. It is the jolt your robot imparts that causes your opponents to lose their game pieces. A robot that only pushes would not have the same effect.

You remember your physics correctly. Perfectly elastic collisions will conserve energy (imparted from one robot to the other). However, FRC collisions are generally closer to inelastic collisions which conserve momentum rather than conserving energy. Momentum is M*V. V is still a big player. So is M. I guess the 60 lbs of CIM motors that we put in the power cube to get the robot weight up to 125 lbs helped a bit, eh?

1 Like

A friend’s robot weighed only around 70 pounds without bumpers and battery. It had 4 775’ per side and was extremely fast. Occasionally, between scoring cycles, they would play defense and give an opponent a single, sledgehammer hit. Often, those single hits would cause the opponent to lose their game piece like you were able to. They would then go back to cycling game pieces.

1 Like

WELLLLL if you want to make this argument… Take 1323, for example. They used a COTS (or soon to be?) drive and implemented a much simpler (compared to turret) swerve and got farther than 254 at Houston Champs. 1323’s focus on making sure they could evade defense rather than just being able to drive forward and backward was arguably a better strategy than the rather short-term approach 254 took that got them to einstein finals. :smile:

16 Likes

Just my two cents. Swerve, like anything else, can perform extremely well, or it can be a complete failure. It is just a matter of making sure that you understand what you are doing before you ever consider using it in a season.

Take team 1640 for example, they design and manufacture prototype swerve modules in the off season, then they use nearly identical modules during the season. Also, most of their designs from previous years have cross compatibility, so they can use previously built modules in their robot until they are able to manufacture new modules.

You could also look at team 2767. They have a quite simple and very durable design that is extremely effective. It goes to show that you don’t need to have a crazy complicated swerve drive to be effective. Stryke force has by far one of the highest performance designs that I have ever seen.

Now for the not so good example
One of the teams that I work with made the decision to use swerve drive last year. We had developed a nice and functional drive system in the off season that was reliable, and was fairly durable. We had planned on using this version during the season if we decided to do swerve, since it would just be a matter of duplicating the modules that we made in the off season, but one team member managed to convince our coach that we needed something more compact, since he saw another team with a smaller swerve drive. Lets just say that things did not go very well, as there were tons of problems with the new design and we did not have time to make the improvements necessary. As a result, we had a bad season. It was definitely not the fault of swerve drive as a whole, it was our fault for deciding to experiment with a new and untested design during our competition season.

Bottom line, there is nothing wrong with making a swerve drive or using it during the competition season. However, you absolutely have to do all of your learning during the off season and test the ■■■■ out of it before you ever even think about using it during the season.

1 Like

The same thing happened with detroit finals. Technodogs were able to evade defense bots like they were nothing. There are obvious advantages to swerve drive, it is just a matter of making sure that you understand the system, and have the resources to properly use it.

Also, I beg to differ about 254’s turret arm being simpler than a swerve drive. That arm is one of the most complex pieces of hardware that I have seen in FRC and the code that they had to create to have it function safely is far more complicated than your run of the mill swerve drive code.

I think that we also need to consider that different games have different demands for drivetrains. Obviously swerve wouldn’t work well for a game like stronghold, but other games that didn’t need terrain navigation, swerve could be much more useful.

Arent turret arms kinda wonky to operate too?

Very much so. You can every once in a while see 254’s arm have to do a full 360 right when the drivers are about to place a hatch or something. Sure it isn’t that much of a delay, but it can definitely throw off a driver who is not expecting it to happen.

Also, 254 ended up making a swerve drive robot in the off season that used the same elevator minus the turret. It honestly seemed to outperform their turret arm by a solid margin, and it was likely more durable, easier to code, and easier to manufacture than the turret.

Pretty much, who needs a turret when the whole robot is the turret…

4 Likes

This sums it up perfectly

1 Like

This is our experience and thoughts as well. We’re opting for no turret as we already have code from 2017 that makes the center of radius of the robot the target. While we had a less-than-effective shooter (well, more so indexer), the alignment system we had with our swerves worked a treat. If we’re ambitious, we may be able to consider robot speed in that alignment calculation as well.

Bypassing a turret has a bunch of benefits, not least of which is lowering the amount of mass needed high up and overall simplifying the design and mounting of the shooter. It also helps ensure that power cells will be fed into the shooter more consistently.

2 Likes

Agreed! We got our 2019 drive base moving the other day, and it was awesome to see it move. Hoping to have our index/flywheel on it shorty to start running practice data for a moving shot.

1 Like

1717 would like to have a talk.

3 Likes

I don’t think anyone here is arguing that having a turret isn’t useful. I certainly think a turret is an ideal solution, and I believe most teams at very high levels of play will have a turret.

Speaking from personal experience and knowing the resources of my team, if resources can be freed up instead of working on a turret (design, manufacturing, programming, etc.) with a marginally less flexible alignment method, it’s well worth it to skip the turret.

That is what I was getting at as well. There isn’t anything wrong with a turret, especially in this game, but many teams don’t have the resources to work on a turret and a swerve drive simultaneously.
However, this year’s game limits the effectiveness of a swerve drive without a turret because you will still have to rotate the robot to intake power cells, so you won’t be able to immediately shoot without a turret mechanism. With this in mind, swerve can much more effectively replace a turret in games like power up or deep space where the robots have to shuttle game pieces around the field instead of launching them.

Also, I will admit that the release of the grayt turret on WCP and the armabot turret system has made a turret mechanism more accessible for teams this year, but I would not trust either of those designs with a heavy arm mechanism on top of it.