20-09-2015 19:57

Knufire

Re: pic: Top View

Why a VEX omni wheel? For a follower wheel carrying minimal torque, it seems like it's rather overkill.

20-09-2015 20:42

GeeTwo

Re: pic: Top View

I don't see that a third follower wheel parallel to the other two would give you any additional information. In particular, if your robot strafes (I'm guessing that you're using a skid-steer and that would only happen if you're rammed or otherwise lose lateral traction), you wouldn't know about it. A third wheel with a different orientation would. Useful orientations for three wheels would be delta (kiwi), H, I, and U. A Y configuration would tell you about strafe, but not report rotation.

Edit: Re-reading your description, perhaps you do have what I would call an "I" configuration. My initial reading of "each end of the robot" was left/right, and I thought you were doing differential to get rotation. I now believe that you mean "front/rear", and are orienting those transverse to forward travel.

20-09-2015 23:47

Ari423

Re: pic: Top View

If your are using them for what I am assuming (trying to calculate the robot's position on the field) then good luck. You will probably need it. Even with good encoders the error in measurements (of distance and rotation) will quickly build up and throw off your data. Plus everything is easy until you account for turning in a WCD, then there are about a billion other factors to account for. I will be very happily surprised if you get it working well and will most likely grovel at your feet.

21-09-2015 01:14

Munchskull

Re: pic: Top View

I agree with Ari, it seems like a very odd way to gather data. Giving the benefit of the doubt that it will work then it seems like it will be a rather heavy answer to gathering directional data. Assuming west coast drive, I would personally use one encoder per side and a quality IMU to track the acceleration, position and rotation of your robot. I have to had it to you though, it is a rather creative solutions for data gathering. Reminds me of how a ball mouse works.

21-09-2015 10:42

carpedav000

Re: pic: Top View

Dumb question; Why not put an encoder in the gearbox[es]?

21-09-2015 10:54

saikiranra

Re: pic: Top View

Having two of these modules perpendicular to each other help see if wheels are slipping or how far a robot has traveled if an opponent is pushing it sideways.

21-09-2015 10:57

Knufire

Re: pic: Top View

I'm surprised at the reactions in this thread, I had always thought that follower wheels were a way to get more accurate odometery than encoders in gearboxes.

The problem with encoders on your drive wheels in a skid steer drive is, well, the fact it's a skid steer drive. You're relying on wheel slip to turn, which means your encoders are going to read farther than they do.

With two follower omni wheels (perpendicular to each other, one facing forward), the wheel will (theoretically) only spin while moving forward and slip, but not rotate, when you move sideways.

I know 33 had spring loaded follower wheels in 2014.

21-09-2015 17:12

Darkseer54

Re: pic: Top View

I have also seen plenty of teams get success in this. If you look at 488's octocanum they use this type of follower wheel. It seems to be a tested method, so not sure what the sudden uproar against it is...

21-09-2015 17:49

Munchskull

Re: pic: Top View

I was not aware that this had been tried before. If this is a bonafide method then I would love to know more. Could some one explane the requirements a little more?

21-09-2015 18:22

Darkseer54

Re: pic: Top View

Did a search for follower wheel, the way that most teams reach success with it seems to be that they use two of them for translation in combination with a gyro for rotation. Though this is used for omnidirectional drives, so I am unsure pf the effectiveness on standard skid steer. It would also help for drives such as 4 omni's which have a tendency to drift.

21-09-2015 20:46

EricH

Re: pic: Top View

A follower wheel is a common thing, in certain drive systems. (So common that, contrary to popular belief, Lunacy wheels were NOT the only thing allowed to contact the regolith--follower wheels were specifically allowed.)

But, as noted, the standard configuration is two, at right angles, with a gyro for rotation, and usually not on a skid steer (Lunacy being the exception).

And, for some teams, an IMU is even more preferred: two accelerometers at right angles to each other and a gyro. Just needs some calculus.


BTW, a more common configuration of a given follower wheel--and a simpler one, I think--is a small wheel (like a VEX omni, and not a VEXPro omni)on the end of a rotating support structure, slightly spring-loaded so it'll stay in contact with the ground.

21-09-2015 23:10

Doug Frisk

Re: pic: Top View

Out of curiosity, has anyone ever attempted to modify an optical mouse to do movement tracking for an FRC robot?

Something like this: http://hackaday.com/2013/04/03/addin...omous-vehicle/

I would think that it should be very inexpensive parts wise and should track well on carpet.

Though it would get a bit more complex if you wanted the device to not be in physical contact with the field.

21-09-2015 23:37

Munchskull

Re: pic: Top View

Back in 2005 (I think) my team tried this but with a ball mouse. It worked till the robot went to fast for the sensor.

22-09-2015 00:55

Doug G

Re: pic: Top View

and just make your robot drive like a mouse ball like technokats (45) used to do!!

http://www.chiefdelphi.com/media/img...c6cf6b31_l.jpg