Path following swerve drive

Hi,

My team is thinking of using a swerve drive, and we are trying to figure out how we would do path following. Does anyone know how we would do this, any thoughts are appreciated!

what path are you trying to follow? I generally recommend against trying to follow any lines on the carpet.

In auton, we would follow a path.

Don’t. You will be happier with a controllable 6wd robot than an uncontrollable and unusable swerve.

If by path following you mean route following the answer becomes something my team has dealt with in the past. In an independent swerve chassis each wheel can move individually. For following a pre-programmed route/path this can cause some difficulties with wheel alignment. if the wheels start the match slightly off from straight this can cause your robot to not perfectly follow the route you set. In 2015 the slight variations in wheel alignment between our modules prevented us from performing a 3 tote auto successfully. However 3360 was able to successfully perform a 3 tote auto that year(link to show amount of precision required). The key difference was the amount of precision our two harvesters required. Depending on the slop present in your swerve drive system you may have slightly less precise driving in auto.

Slop effects both tank and swerve drives systems. This is part of the reason why you can see teams driving their robots back and forth across ground before a match (helps with lining up the robot and making sure the gear systems are matched up reducing the effect of slop). It affects swerve systems slightly more in that you have two separate gear systems that can be effected by slop(drive and turning vs. just drive in tank).

Additional note on path following for swerve:
In tele-op following a path is relatively simple as any of the minor differences in wheel angles can be dealt with by your drivers.

Tldr;
Auto: Depends on slop in your gear trains and how closely you need the robot to follow the preset path.

Tele-op: you should be okay.

Thanks. Will definitely be helpful

I would like to specify what I meant by path following. I mean a set of time parametrized positions, a trajectory, if you will. If that clears anything up and changes it, please let me know.

I would love to see a swerve drive at both of the District Events we have with you guys!

With some good vision processing code, you can probably pull off a reliable auto. This would take some serious development though.

If you are not using vision in auto, you might be less reliable at placing a gear or shooting high. If you design your mechanisms to take into account the error caused by the swerve auto, as in increased shooter accuracy or increased range of where you can run into the wall and place your auto gear, you might find the costs of more difficult auto don’t outweigh the benefits of swerve in teleop.

If you’re gonna go swerve, this is definitely one of the better years to do it. If you pull it off, you will find yourself performing really well at the CHS district events.

Good luck, and see you guys at DC and CMD!

Hey,

We are leaning towards using a mecanum drive, due to the complexity of a swerve drive. Should still be pretty cool.

See you at comps!