Our team's unique holonomic drivebase, feedback wanted!

[KJARVIS]

Hey guys, during the offseason my team was discussing ideas for interesting drivebases that would help us stay competitive in the 2023 season, should holonomic be advantageous.

For context, we (team 4788!) live in West Australia and we dont have the resources to buy COTS swerve modules. We have experience with West-Coast-Drives and Mechanum, but both have obvious drawbacks that make them less competitive (in most situations) than swerve.

Ideally, the alternative drivebase would allow us to achieve holonomic motion without impeding on normal tank motion- in the case of an experimental drive system failing, we want to be able to just take it off and have a tank drive. It also means we can maintain the pushing power/defence of tank.

The Plan

This lead us to a 2-state actuated drivebase we called WASP (West Australia Sideways Propulsion). The basic idea was to place two pneumatically-actuated omni wheels on the back of a west coast drivebase, with omni wheels on the WCD’s front two wheels.

When actuated, none of the (highlighted orange) traction wheels touch the ground.

After further research, we found it was really similar to 1939’s drivebase in 2017

which used the same concept to drive sideways

It seems to be forced to move in arcs instead of straight lines since, unlike H drive, the sideways force isn’t applied to the center of the rotation, creating a moment. However, this could just be what they did in the video.

To achieve straight sideways motion, our solution is to add 4" mechanums to the front wheels of WCD.

Testing

To test this, we mounted a perpendicular omni to the back of our 2022 robot, and put mechanums on the front, which worked well, even with full manual controls (tank drive on the thumbsticks, sideways wheel on the triggers).

WASP Drivebase Prototype testing

The video only shows us moving in arcs but we could go straight, it felt identical to driving mechanum. We also tested driving with the omnis ‘undeployed’. The mechanums do not greatly affect how the normal WCD drives.

CAD version

We’ve also CADded the module on our WCD to check the piston geometry was possible with 60PSI in a reasonable form factor.

In this version we’re using two 32mm bore x 25mm stroke cylinders to lift the robot about 2°. both rear omnis are driven by a single neo through an in-tube gearbox (We’ll also reconsider the use of the vex gears and mechanum wheels😬)

Heres a link to the inventor files if you wish to check it out.

Evaluation

In conclusion, with a proper control scheme and the right game field layout, we may consider this drivebase for 2023.

PROS:

• Cheaper than swerve

• Can be used as normal tank in case of module failure

• less motors and complexity

CONS:

• Requires switching between traction and agility, can’t be simultaneous and more driver dependant

• Basically unproven in competition

• Requires pneumatics and consumes some air to deploy

Of course, we haven’t thought of everything. Any feedback is greatly appreciated and we’re happy to answer any questions!

[pryland]

Checkout butterfly and/or grasshopper drive.

We had worked on a Colson/Mecanum without the sideways omni, but never got past the prototyping due to time constraints and increased development/availability of swerve.

pic: Butterfly Drive

[a475ab733ead0c44da5260d5a7a75b64_l.jpg] A new butterfly drivetrain that we hope to test in the offseason. This allows for strafing in addition to a “high/low speed configurations. FESTO pistons are connected to the wheel pods, which serve to switch between a 4” omni wheel and a 2.5" colson wheel. In “colson-mode,” the sprocket and gearbox reduction allows a speed suitable for lasting some time in a pushing match, and in “omni-mode,” the drivetrain has a high speed, and the omnis also allow fo…

pic: VersaDrop GrassHopper Drive

Good point, I wasn’t thinking about that aspect. The lowest speed option with the single speed double reduction is 10.39:1 for 7.25ft/s. Does that speed sound more effective?

[diswed0226]

This is cool! Having the wheels in the back makes the packaging a little easier than H drive. I’d be a little concerned about those mecanums holding up if FIRST decides to put any sort of bump on the field. Maybe look at some beefier options?

[chadr03]

I know Spectrum prototyped something similar to this back in 2015. I don’t know if it ever made it on their final robot, but they used omni wheels on the front and just counter rotated them to go sideways.

2015 Recycle Rush Drive Train Test

Also see the build blog.
Spectrum 3847 Blog: Day 17: Drivetrain Video

[KJARVIS]

We were concerned about durability bieng an issue, but we’re planning on using these wheels instead of the vex ones, which have support on both sides of the roller.

[Joey1939]

Wow! You guys have done a very nice job of implementing this concept. I was programming lead on 1939 in 2017 when we implemented sidewinder, and It’s great to see someone else trying something similar. I especially like the actuation mechanism which resolves the tight ground clearance that we had with our sidewinder drive.

The one thing I would add is that our sideways motion got significantly better through out the season. In our robot reveal video you can see how it tended to drive in an arc as you describe. I added a PID controller that tried to use the gyro reading to keep us moving perfectly sideways by causing the front two wheels to spin in opposite directions while we strafed. During the season this software control didn’t work very well. However, the driver got pretty good at apply this counter-rotation manually to improve strafing.

If I remember correctly though, it turned out that there was a software bug that kept the PID controller from working correctly . After the season, I fixed that bug and it was able to strafe basically perfectly straight side to side. Here’s the code for that, it is any help.

[KJARVIS]

the other insect drives were inspiration for the name “wasp”!

Since it’s our first year doing something different, we wanted to make sure that, if our stuff fails, we still have a good west-coast drive. constraining wasp to a little module allows us to remove it if we find it isn’t effective later in the season.

[Goanna]

Thank you so much for your help! We tried to implement something like this but it didn’t work that well, knowing it worked for you it may have just been some bugs so I’ll give it another shot.

One thing that I found slightly worrying with our prototype was the amount of wheel slippage, which might make autonomous pathing difficult as I can no longer rely on encoder readings, did your team have this issue + any fixes?

[AronRubin]

I was just telling my son about how the Adam West era (“bam”, “pow”) batmobile was depicted to have a pneumatic piston extend from the bottom to spin on a dime. (teaching to drive )

[D.Allred]

Interesting concept. I’d say weight is also a significant con for this approach.

Try comparing this configuration to some of the “butterfly” concepts. Also take a look at 624’s modified H-drive from 2014 (Team 624 CRyptonite 2014 Reveal: Sidewinder) and 148 “nona-drive” for other configurations.

David

[mrnoble]

As a learning exercise this is fantastic, I am pretty sure yours is the first robot to use this exact setup, and that took some creative thinking from your team. Great work!

However, as an actual competitive concept, I would probably not run this. Because of the way your first post presents the rationale for the design (can’t afford COTS swerve, still looking for holonomic effectiveness) I wonder why you can’t remanufacture an existing swerve design, or come up with a design of your own? Frankly you’ve exhibited the skill, drive, and machining capacity to do it.

That’s me though, and my advice is based on limited information and a lot of bias. Your team sure looks capable of doing the same kind of work that teams like 3005 have done in the past, and that’s a lot to be proud of, and the foundation for a very successful future. Best of luck to you.

[boof19]

It may be worth looking at FTC’s implementation of deadwheel odometry. Sticking a couple of dead omni’s attached to encoders in a vertical and horizontal configuration. We have run it the past 2 years with some success.

[KJARVIS]

As much as i’d like to think we could make our own swerve, sadly i don’t think we’re there yet. We only have access to CNC tools during build season, which is obviously not the ideal time to try and develop our own swerve modules. We’d also like to at least have our normal WCD if our holonomic doesn’t work out.

[Tim708]

708 did a butterfly drive for the 2022 season (and 2018). With our set up it would have been easy to decouple the drive gearboxes and swap the Omni wheels to mecanum. We used a different pneumatic cylinder arrangement from what I’ve traditionally seen done for this type of drive train, it worked well for us.

[TyJac42]

If interested in further examples of a perpendicular drop drive 118 made an awesome video of it being used in competition. Perpendicular Drop Drive - Robonauts 118 2014 - YouTube

[Andrew_L]

You’re gonna drive it like a tank drive, I guarantee it.

[KJARVIS]

Our of curiosity, why are there two cylinders per gearbox? 3-speed? PTO?

[KJARVIS]

Yeah h is great but we found a couple flaws with it-

Really low sideways defencability, since its all omnis if youre hit at an angle on the side you seem to get spun or even just pushed really easily- looks like it happened to 118 a fair bit.

Also, we believed we could improve packaging/leave more space in the middle of the robot by putting the wheels in the rear.

[Tim708]

It was 2 speed with PTO. Seemed like a good idea when we started down that path but in hindsight we would have been better off without the PTO. Still a fun design exercise for the kids.
Our 2018 butterfly was essentially a 3 speed, we had a gear reduction between the omni wheels and traction wheels and had a 2 speed shifter. Low omni and high traction were geared to about the same speed.

[TyJac42]

Did you look at the video (timestamps 0:15-0:30)? It is not an H drive, it is exactly the same thing you put on your robot. They drop an omni-wheel in the rear, tilting a portion of the robot, then they turn the tank drive and turn the omni-wheel for super-fast spinning without stopping. In the competition, they only deploy the omni-wheel for a small period of time when need to do a quick 180-degree maneuver to point at the goal without waiting for defenders to keep up. If you look at it you can see they make a super fast turn, while still keeping that forward momentum. H drive (in my humble opinion) sucks most of the time, and in 2014 it would be terrible since that game had a lot of defense to push you around.