Omni Mania
 

Im a mentor this year for my team (bagelbytes) and im reseaching options for the team during the off season

Presently im working on the usefulness or the possiblities of using a omni drive. for all of you out there that have used or know about omni drives what are things that i should know.

ups?
downs?
Difficulty?
Speed?
Torque?
Complications?
Ups vs. Other drives?
Where to get them...?
or how to build them?

and any other thing i should know if i am to construct on before season start to try it out


I realize im asking broad range on questions, so take your pick at what to answer

Thanks in advance:yikes:

-Oscar-

Oscar,
You have a lot of questions. There are several different applications of omnidrives (sometimes called holonomic drives).
To date, the most sophisticated omnidrive system that I have seem in FIRST competitions have come from team 111 and 45 (or is it 47?). Team 111 (Wildstang) had a four wheel steering drive system that competed last year. You can get more information from their website.
The other bot from team 45 (or47) competed several years ago and functioned very well. It used two steerable drive wheels and caster.
Other applications have been used and each has advantages and disadvantages. You have to decide where your limitations reside. Manufacturing? Design? Cost? Weight? Software?
I know this didn't answer all of your questions, but I really recommend checking out the Wildstang website. It has a ton of information. A Google search on omnidrives will also inundate you with info.
Good luck.

Many teams have used swerve or crab style drive systems, but the two that Warren mentioned, ChiefDelphi (47) and Wildstang (111), are the leaders in this area. If I remember correctly, team 47 was the first to do this in 1998. Team 45 has never attempted swerve or crab drive... we're tank people like Warren's team.

Andy B.

Sorry for the false credit Andy.

We may not be tank people much longer, if the game keeps changing in the directions it did last year.

Treads were not as important since the majority of the game didn't require mondo pushing power.

Live and learn.

since we have no idea what the game will be this year
we have no idea what type of drive or powertrain will be optimum for this years game.

One danger here: if you put a lot of time and effort into developing an omni drive now, then you will be likely to use it on this years bot, even if its not a good match.

The advantages of prototyping a drive system like swerve or killough are numerous. The most important however is that it is just another way students can learn ;)

We used the 4 Wheel Steering in 2002.

The advantages are the extreme maneuverability. Most likely the only robot that can effectively block you is another 4 wheel steering robot. Or you may need the maneuverability for the manipulation which is what 47 used in 2000 to move parallel to the wall while the ball grabber pointed at the wall.

Disadvantages however are the extra motors required for steering, as well as the amount of power you can put down to the ground. Typically the wheels are not super wide or grippy so that you can effectively turn them in place.

To make it easiest we made the drive pieces modular and interchangeable. The steering stayed intact and you only removed the power module with one bolt. It proved to be very effective.

... and then there is always the possibility that people involved with the design of the game for next year actually read through these posts, looking for ideas...

Just imagine if the following hypothetical conversation were actually true:
W: "hey, it looks like all the veteran teams are building "Team Whizz-Bang"-style drive systems this year. They are all spending all summer building prototypes and perfecting the designs. All those poor rookies are going to be left in the dust!!!
D1: "well, won't they be surprised when they see the new game and see that "Team Whizz-Bang"-style drives are absolutely useless when driving across a platter of cooked rice." All that time and effort spent working on the wrong problem! And after we TOLD THEM not to make any assumptions about the next game!
D2: "heh heh heh heh!"

But, of course, none of the game designers actually read the CD posts, and even if they did, they wouldn't be that devious... :rolleyes: :rolleyes: :rolleyes:

I am just waiting for the game that is entirely dependent on climbing... monkey bars perhaps? :rolleyes:

::shudders:: Don't give them ideas!

Is it just me?
Or did anyone else read that and get a tingle?

*shifty eyes*

Good thing we've got that hovercraft/submarine/biplane drivetrain completed. ;)

Ahh yes but don't forget about the traction on cooked rice. That's a tricky one. However, I think rice would be to obvious, I'm betting they go with something more like a penne pasta.

keep the ideas omni ideas coming plz

lol its risk to invest in a drive ahead of time but its risk im willing to take so plz anyone wiht omni expirience speak up and by the way i looked up wild stang now and along time ago and i still dont see any omni related info, maybe someone could direct me with a link perhaps?

I think what everyone really wants to know, Dave, is whether you're D1 or D2 :)

Woa.. that would be awesome to see this robot just kinda spinning out and sinking deeper and deeper into a 3 ft bin of cooked rice. :)

lots of info about various omni drives can be found in this thread.

This thread includes listings of all applicable drive trains, discussion and debate about the various strengths and weaknesses of each drive, a place to obtain good omniwheels, etc. etc. Even some argument over whether or not a 4-wheel killough can deliver 100% power.

read up in there.

Thanks that is more useful

Wildstang
 

I read in the past threads about 111 wildstang being a serve drive... i thought it was a true omni drive with omni wheels

Can someone verify of clarify?

In the 2002 and 2003 seasons Wildstang (111) has used a crab steering system. Their drive system was powered by 2 Drill motors and 2 FP motors (one motor driving a different wheel), and their modules were rotated using 2 Globe motors. Check here for 2003 pictures http://www2.wildstang.org/2003/inven...otographs.html

In 2003, The Sea Dawgs (258) used a similar system, but we used 2 Drill motors and 2 CIM motors to drive (1 per wheel), and 2 Globe motors to rotate each pair of modules.

<edit>
That link into Wildstang's Inventor Award submission is a really neat place to look at the intracacies of their robot. Their 2002 submission was great, as well.
</edit>

<edit_2>
I'm going to find a few pictures of 258's 2003 system and link to them a little later.
</edit_2>

Any chance the 2002 submission is still out there somewhere? I'd love to see it. Heck, I'd keep a copy on my computer if I could.

I'd love to see more teams publish inventor awards of this quality. I think we'd see an overall increase in bot construction and technology if everyone shared like 111.

which team had the ball drive - looked like a giant backwards mouse ball - was that in 2002?

I dont know how well it worked in the games, but that is what I think of when I hear omni-drive - it could go in any direction without first having to point the wheel (there were no wheels)

Not to get this off subject or anything.....

We (team 45) had a ball drive last season, but we did not use it an any competitions. Team 909 did use the ball drive while in Chicago due to the fact that their robot was not operational.

here are some links

Team 909 using the ball drive

Thread about the Ball Drive "Mighty Mouse"

Ken - TechnoKats (45) had the ball drive.

I'm looking into finding our 2002 Inventor submission. I'll post a link if I find it.

I wasn't on the team then, but in 2001 our robot (Rampbot) had BUPODS (omni wheels) on the front. Mike Soukup can probably fill in any details.

This picture shows the BUPODS. There may be closer shots in the 2001 gallery, these were the first two I saw.

http://www2.wildstang.org/gallery/2001_Build/P2100004
http://www2.wildstang.org/gallery/2001_Build/STUDEN_2

I've designed an "omni drive" that we probably will end up using this year. It can function as a Skid steer, Crab, Car, and Angle drive systems, all at the simple push of a button. It should be pretty durable too. I'm not going to post any specs on it until after we build it.
I've prototyped it several times using Lego's/edu bot parts, etc, and each time it's worked very well. Since it can function as any system, we won't lose anything walking into kickoff with the designs waiting. If there isnt any reason not to use it, then we'll probably go through with it.
It costs about $800 for the chassis, drive components, and some spare parts. It's estimated weight is around 40lbs. (w/o electronics)
Anyway, I'm putting together a video presentation on my system, to show my team. I might post it up on CD later...

I Think Im coming down with it...

I think ive gOT OMNI FEAVER!

Oh NO! NoT the Crab sTRAND Thats the Most Contageous!

i posted this earlier
 

here was our omni-drive killough from 2002, we loved this drivetrain. its super maneuverable and very easy to program, but we had no traction. we didn't need the traction back then though, so it wasn't a problem......this video is worth watch though.

we never got caught up in definitions when we built our drivetrain............

since there is so much talk about omni-drives, i thought i'd let you know that we're still waiting to use our next generation Kiwi Drive............stay tuned, its been two years in the works.

for now, check out our old video from the pre-season of 2002, this was a video we unveiled before competing with our reworked chassis. it was more fun to drive than to watch.

the first and last FIRST Killough, for now

http://stuweb.ee.mtu.edu/~alkrajew/FIRST/kiwi.mpg

anthony - that video is just awesome

im wondering why you say you didnt need a lot of traction in the 2002 game?! that was the one with the 180 lb goals that had to be dragged into your zone to score

and the finals pretty much turned into a goal dragging contest.

so...... im confused (as usual :c)

I've got a copy of it on my computer. My AIM is bi24ll, contact me sometime and I can transfer the pics/vids to you. I'll be out until late tonight, though. I'll be around tomorrow.

Ken,

I thought about addressing this when you made your other post in this thread, but never made the time for it.

Each year, there are many ways to play the game. Last year, you could stack, or you could play for the top of the ramp. The year before, you could grab, move, and steal goals, or you could do balls. Each year has had a break down like that. These past two years, the scoring has been biased toward having a powerful robot, but I don't believe that was intentional.

Last year, team 67 had one of the best stackers, and they also had a swerve drive which helped them immensely. Looking back, we (73) really could have benefited from a swerve also.

in 2002, the kiwi drive was used on a ball collecting robot. I never saw their full robot in action, but I can guess that they paid little to no attention to the goals. Did they win any competitions? No. But they did build an awesome robot.

Most teams find out the game, then choose a strategy, then build a robot to meet that strategy. However, other teams choose something new to try, and then figure out what part of the game they can use it best at. I can almost guarantee that there will be some part of the game next year where a very maneuverable robot will have an advantage.

Some of my favorite robots (Beatty 2002 being the best example) were designed specifically to suit the game. But who knows if their walking design was designed in the first weeks of the 2002, or the summer before, or 5 years before.

The Kiwi drive and the ball drive are both very beautiful designs. One was used in competition, the other wasn't planned to be. Who knows when their team might win the championship because of their ultra-manuverable drive system developed earlier.

I certainly wouldn't discourage 702 from working on a swerve drive, whether or not they design their robot around it. They have the added bonus that they would be only the 2nd team to have a working swerve on the west coast (258 being the first).

Bill passing Wildstang pictures & videos to John & the rest of FIRST....hmm....that sounds oddly familiar ;)

We're currently looking on the server for the 2002 inventor submission and will hopefully have it up again in a day or so.

SpaceOsc: Our 2002 robot used a 4 wheel swerve/crab drive without omni-wheels. Dave posted pics of our 2001 robot that had a set of omni-wheels. There may be better pictures around our gallery, and I'm pretty sure Matt Reiland has some good closeups too.

ajlapp: I'm still amazed at how cool your Kiwi drive is and also how simple & elegant the control system was. Great job with that bot.

Mike,
I'm sure I don't know what you're talking about, and... they were all beautiful and inspiring. ;)

John

thanks Mike!
 

to answer a question.......

we designed specifically to pick up balls and not manipulate the goals, we didn't win, but i'm not entirely sure it was because we ignored the goals.

thanks mike, by the way, i've got chris ankeny here diligently replicating the StangPS for me! hope you don't mind.

Re: i posted this earlier
 

Anthony,
Just so you know, we on 229 LOVE that video.
We show it at demos sometimes.

The kiwi-drive itself is wicked cool.
Hopefully we can emulate a quick one at some point. We also loved your *programing* method. :D ;)

John

thanks again
 

thanks again for the compliments.........

we made a quick kiwi out of the edu robot just for demos, worked pretty good, and believe it or not our programming worked without any modification.

speaking of modification, my programming guys are sweating bullets about having to re-write their kiwi code for the new controllers. :D

can you give a breif explaination of how the joystick inputs are used to control the 3 motors at 120° angles from each other?

its not clear from the video exactly how the bot was intended to be driven - lots of spinning and twirling going on there.

trade secrets
 

i'd love to explain.........what i can anyway.

we used three normal joysticks, connected together with a "y"-shaped yoke, and ball joints. the x-axis of each stick was unwired, and the threey-axis of each stick were wired into one coupler.......just to save space on a cramped IO.

so each y-axis of the joystick represents the velocity vector of one wheel. the joystick does the complex task of converting cartesian coordinates to polar coordinats inherently.

the robot goes in the direction the stick is pushed, or spins if the stick is rotated in the center.

the attached pictures show how the joystick is a vector based model of the real robot......instead of a motor and gear set, you have a potentiometer. you manually input the velocity of each motor and the robot responds. (both pictures were prototypes)


problems!
our drivers created what they called slip. this was their method for driving. it was based on a general direction rather than a finite point in space. they kind of glided the robot to its position. the joystick doesn't make it super easy to go in straight lines, but straight is irrelevant when you can go anywhere instantly.

we worked around this by making multiple faces for our robot to operate on......front didn't really matter.

thanks for the info.

simply awesome!

Actually, team #360 from Seattle, WA (my former physics teacher's current team) had a double shaft swerve in the 2002 competition. Also, team #998 from Alaska had a double shaft swerve this past season (2003).

At least I told you before more than 8 IPs downloaded the pics/vids. Don't I get any credit for that? By the way, it wasn't me who passed the url to John. There was in intermediary ;).

Here are a couple of pictures from 111's 2002 Inventor Award submission.
http://www.seadawgs.com/111/eng.jpg (two engineering drawings)
http://www.seadawgs.com/111/DSCN0736.jpg

Here are a few pictures of 258's 2003 drivetrain.
http://www.seadawgs.com/258/cimgearbox.jpg
http://www.seadawgs.com/258/cimassembled.jpg
http://www.seadawgs.com/258/cimassembled2.jpg
http://www.seadawgs.com/258/cimassembled3.jpg
http://www.seadawgs.com/258/cimassembled4.jpg
http://www.seadawgs.com/258/cimassembled5.jpg
http://www.seadawgs.com/258/drillgearbox.jpg
http://www.seadawgs.com/258/drillmounted.jpg

I've got drawings for all of 258's drivetrain components if anyone's interested, as well.

<edit>
I forgot to mention that I'm working on a paper describing the entire design process that led to 258's swerve drive this year. My goal is to use this in the WRRF's Motors/Drivetrain class in November and December.
</edit>

Also, team 159, Alpine Robotics from Fort Collins, Colorado, for all intents and purposes, had as much of a swerve system as we did (I consider them far enough west to be "west"). It's too bad, though, that they had some technical difficulties at the Silicon Valley Regional, in 2003, and were forced to lock two of their gearboxes, and install casters, if I recall correctly.

Team 668 had the preliminary workings of a swerve system, as well, but again, due to technical difficulties, it wasn't functional for competition.

Credit where credit is due...

its all because were all too lazy to turn our bodies in the rite directioin

Bill,
We're thinking about a swerve for 2005 (when Eric and I are in Senior Integrated Design class, and get to do FIRST for a grade). So I'm already gathering ideas/concepts. I'd love any information you'd share, and would be VERY interested in you're paper.

(brief aside: I'm working on a similar paper for some of the technology we're currently developing.)

I like to think of your swerve drive as "the working man's swerve". Unless I'm missing something, it seems like one of the *easiest* to fabricate. 229 has CNC manufacturing, but nothing really above that.

John

Hmm, that's a really cool design!
Mine does not have the motors on the wheel assemblies, however, and it uses more chain than gears...
I'll let you know how it turns out...all the prototypes so far have worked...
I'm not really sure who has had a crab/swerve drive here on the east coast...
I'm pretty sure there was 1 I remember from this past year...maybe we'll be the second?!
Heh...

Go 702 !!!!!!!!!! lol lil team pride


But anyways thanks for alot of the help you guys have lead me into odd and weird directions with this thing.. ill try wiht one of the mentors to build a prototype from a pre existing frame except were not trying a crab drive despite the obivious love affair you all have with it.. 702 going with a omni drive using all omni wheels in standard omni setup tho i do realize the obvious set back of less traction its alot simpler to build for now and the teams a lil weary on difficult new desings so they needed to be wowed right now with omni drive maybe well head in the crab direction later

does anyone one know where to get already made omni wheels.. i found a few places but does anyone one have somethign they already know is good?

Hmmm how about this......

Omni Mania
 

Greatly appreciated

But i shoulda made it clearer I WANNA BUY WHEELS!!!!!!

lol thanks for the the pic tho ::drools :::D

Re: Omni Mania
 

www.kornylak.com


Good omniwheels. bearings in the rollers, polyurethane rollers, aluminum construction, the works. Expensive, though.

Hey Bill, how 'bout the real thing ;)

After shaking off some of the dust, here's a link to our 2002 Inventor submission (thanks Kristin).

http://www.wildstang.org/ws/Inventor/2002/

hey thanks aflak

thats the site i found i preffer the cheaper transwheels alot better on the site

anyone have any where else to buy them?

Erm, careful with those cheaper wheels . .
One type (50 dollers) is all aluminum and has built in bearings, both illegal for FIRST use and not easily driveable. The others are all plastic. Be careful that whatever you buy will meet your weight/stress requirements. A 2 inch plastic omniwheel will not support a 130 lbs robot.

Moneys Worth
 

but 12 four inch 4000 series wheels just might hold :: plans & schemes :::

Omni wheels
 

We started the 2002 season with 4WD and skid steering, but not enough power. Immediately following a regional, we manufactured two omni-wheels for the front of our machine and added a swiveling tow-bar. With this progression I can offer a before and after viewpoint.

Pros:
Manueverability: Our 2002 machine went from an unturnable beast to a spinning, driving wonder (comparatively, at least).

Cons:
Traction loss. Our design had to replace existing wheels, resulting in 20 nylon rollers arrranged on the perimeter of an 8.5" aluminum disk. At any time there were perhaps 3-4 rollers 'biting' into the carpet (we left sharp edges on the rollers for just this purpose), but there was some traction loss compared to the original, rubber wheel.

Rough ride. As a new roller came into 'bite' there was a bit of a bounce and a lot of clacking noise. On carpet it was sufficient.

Maneuverability: Double edged sword. Another robot could ram us on either front side and redirect our efforts.

All in all, those omni wheels made for a MUCH better machine. Consider carefully what you give up for what you lose.

Good Luck

Out of curiousity, does anyone know of any team that ran 4WD skid steering with all omniwheels?


I can't recall ever seeing that done before, and while I understand the problems associated with that strategy, I'm interested in learning how it might handle.

Omni History
 

Do you mean a robot that was entirely composed of omni wheels... cuz as far as i know it hasnt been done ,of course i wont mind being the first:D

but if there has been id like to know from who and when and how it went

I believe that in 2002 MiM (343) had a mode that did this.

MiM had 6WD Omni.
They had 2 wheels, 2 wheels, 1 wheel, and 1 wheel on the sides of the bot.

So this bot wouldn't perform like the one Maddie is asking about.

Maddie - I've never heard of anyone making a bot like this. I can't imagine why anyone ever would... it would be too easy for someone to man-handle you all match simply by pushing you sideways.

John

You should know me better than that by now ;)

We started the first regional with that setup in 2002, it let the robot fly around the field but as stated above, it also let the robot get pushed laterally easily. We ended up putting two traction wheels on the front and left the omni's on the back and that worked well

Pushing too hard
 

this being pushed around, is it as bad as they say for teams using omni wheels or even crab drives?

i saw alot of footage involving 111 in alot of the regional matches and they zoomed across the field tho i did see some bots give them a few pushes in the wrong directions 111(wildstang) did lil work of them and zoomed thru the field evading and getting out of tight places and rough bots? so i mean is it that easy to take out such fast and maneuverable bot? video says contrary but i suppose it also depends on how good the bot ends up being

Re: Pushing too hard
 

We are talking about omni-wheels.
Wheels that allow for lateral motion.

Basically, if you use 4 omni wheels there is NO friction in the lateral direction. This means someone can push you sideways across the field with little or no effort.

As far as Wildstang's drive...
They had a decent amount of traction, they just weren't geared low enough to win a "straight pushing match" they relied on maneuverability and their cool wedges. They could have pushed, if they were geared lower. (Shifting swerve drive anyone?)

Hope this helps.

John

Since we're on the subject of Omni-wheels, I should mention that there is a new white paper on this subject. It's called TechnoKat Trick Wheel.

This paper includes a detailed mechanical print package for an "omni-wheel" (we call them Trick Wheels). The one in the print package was used in 2003. The designer & developer of this module is Mark Koors, TechnoKat & Delphi engineer.

Andy B.

Woah
 

Nice:D :p

Re: Woah
 

I have a quick question. Ok, of all the research ive been doing on omni drive systems i like the holonomic the best ... but since 2002 FIRST has only given 2 matching motors ... How can you accompish a true holonimic drive with only 2 matching motors ... oh yeah and i hate the 4 wheeled holonmic drive, it limits manuverability to fowards/backward and side to side :rolleyes:

Re: Woah
 

with only 2 motor i dont know but it can definitly go in front/back and right/left and anything inbetween.... just dont know how to turn tho. with simply two omnis in 90degrees apart. and 4wheeled omni drive is true holonomic. which means it can go front back right left anywhere inbetween and rotate similtaniously.

Re: Woah
 

All you're concerned with is that the output speeds match eachother. This can be done with gear ratios to get them close to the same speed and shaft encoders to take care of minor variations. It's very similar to the way teams use shaft encoders to make sure their tank-style drivetrain goes straight.

"Limited maneuverability" isn't in the vocab of a holonomic drivetrain. Both our and SPAM's 'bots were of the 4-wheeled variety last year and they sacraficed nothing in terms of maneuverability.

Take a look at Evette and SAM if you'd like.

Re: Omni Mania
 

before the 2002 season we built our prototype kiwi drive with two drills and one fisher price motor. as with any other motor matching we geared the fisher price to meet the drills free speeds. we had never recieved three matched motors before but were determined to use our kiwi drive and controls. when the kit arrived and we had three matched drills we were psyched........but our prototype worked awesome even with one odd man out.

because of the nature of the kiwi drive the slight differences in motor matching didn't affect performance, you're constantly using all three mtors and wheels and it just buzzes along. we've never used it again, but neither has anyone else. it's a really great drivetrain....we've built all kinds and the kiwi is still my favorite.

Re: Omni Mania
 

is the kiwi drive defined by the fact it has three omni wheels? does that seperate it from other omni drives? lets say using 4 to 6 omni wheels. or is it something else that makes it kiwi

Re: Omni Mania
 

Hmm.. even I too am victim of misreading thread titles.

I thought this was a thread about potential guests during the Championship event of the Omni hotel in Atlanta.

Oh well...

Carry on.

Re: Woah
 

Let me add this little experiment to George's comments:

Start with omni wheels that have the rollers in the plane of the wheel.

Build a rectangular chassis with a pair of omnis "front and back" - aligned with the chassis's latitudinal axis - and a pair "side to side" - aligned with the logintudinal axis. (we created a prototype with a piece of plywood and 4 rechargable Ryobi drills ... with a little creative soldering to wire the speed controller outputs to the battery contacts)

Load the IFI default single stick code into the FRC

Rotate your chassis 45deg AND rotate your single joystick 45 deg.

Now try out the maneuverability of 4-wheel omni ;)

OK, now, rotate the joy stick back, and do the rotation in software
:cool:

Finally, add a second stick, that will superimpose rotation left or right on top of the translation stick output
:p

Add a yaw rate sensor to detect drift (rotation) due to drivetrain irregularities and output additional "opposite rotation" for correction ... and viola!

That is the essence of what we used last year...
eric

Re: Omni Mania
 

kiwi was a code word we used to refer to a killough drive base. the original killough base has three wheels 120 degrees apart. it is the simplest configuration for true holonomic function, at least in terms of quantity of motors and such.

though building a true killough is often considered too challenging, mainly due to programming complications, control can be solved in many novel and simple ways.....our solution in 2002 required no software whatsoever! i'd love to build another, only time will tell. perhaps in the meantime another team will attempt it......no one has fielded a true killough since ours in 2002. :(