Introducing Scout: the Kitten Bot A Functional Swerve Bot in an Inexpensive (relatively) Tiny Package

[Mr.R_2]

Hello Everyone,

Scoot the kittenbot

Has your team been contemplating attempting swerve, but are not sure whether they have the programming knowledge necessary to make it move let alone go where you want it to go? Have they wanted to try it, but had difficulty justifying the cost of entry for something that may not be a good fit for the team?

Well, we (6762, The Oscats) are in the same boat, and we created a potential solution.

Scoot is a 3d printed swerve chassis that uses only 2 REV swerve modules and has a full FRC control system.

This means you can create a swerve bot for half the price of what a swerve bot would typically cost.

We are learning a lot with this little bot, and we wanted to share in case it could help other teams figure out whether swerve was right for them.

Here is our documentation.

Scoot

Here is our Cadd.

Onshape

And here is our code

GitHub - Oscats/ScootCode

The code uses a modified version of Rev’s Swerve code, but with only two modules.

There is a lot we have yet to figure out (check out the Known issues), but the robot is working.

If anyone has any feedback or words of wisdom, we would be open. As I said, we have a lot to learn.

As I write this, the modules are in stock on Rev’s site, and the motors (Neo and Neo 550s) and controllers are in stock at Andymark.

We have been trying to design a relatively inexpensive and miniature (so mentors can take it home to explore) Swerve solution for a while now, and this is where we landed.

I posted this in programming as this is largely an experiment in programming. If this is in error, mods, please move it to the correct place.

Thank you all for your help. We could not even consider this without all we have learned from and the support of this community.

Edit: I added the CADD link. Our printer is a Prusa MK3+. We chopped the chassis up into pieces to fit on the base. We have a small issue where we did not know how much PLA was going to expand, so we left too much space for the tongue and groove. This is easily remedied with glue or screws, but eventually we will fix the Onshape.

We wanted to publish now in case it could help a team considering swerve for next season.

[BrendanManning]

Comparing the CAD to the video it looks like a few components have moved around a bit. Specifically it looks like the battery and breaker have been moved around a bit. Was this just a case on the things that worked in CAD not doing as well in reality or was there another reason for the change?

[Mr.R_2]

Hello,
Yes. I have not updated the CADD. We used an inexpensive boat PDP instead of the Rev one (the CTRE one may not fit in the robot at all). In doing so, we were able to put the PDP under the chassis. This allowed us to avoid stacking components. Here are the images of the current version.

Another change is that we only used bolts to keep the chassis together and to attach the swerve modules. We used Velcro for everything else. So, many of the bolt holes are not actually used, and we ended up drilling a few more holes to keep everything together.

We did use silicone wire as it is more flexible.

I highly suggest printing with support material for the tongue and groove as well.

[widekarl]

We used an inexpensive boat PDP instead of the Rev one (the CTRE one may not fit in the robot at all).

Do you happen to know the model of that boat PDP or have a link to find it? My team is low on funds, so if we were to do a project like this, we wouldn’t want to have to buy an FRC-legal PDP/PDH. That little version seems like a great option!

[UnofficialForth]

Have you considered a similar version using the new XRP platform? What pitfalls could you see with a team working to integrate something like that.

[Mr.R_2]

Hello,

@widekarl

Do you happen to know the model of that boat PDP or have a link to find it? My team is low on funds, so if we were to do a project like this, we wouldn’t want to have to buy an FRC-legal PDP/PDH. That little version seems like a great option!

We used this one which is going for 19.95 on Amazon. If you use it, you should use a 120 ohm terminating resistor in the CAN bus. That may be optional, but the documentation says it is recommended, so we use one.

@UnofficialForth

Have you considered a similar version using the new XRP platform? What pitfalls could you see with a team working to integrate something like that.

We have experimented with many other mini robot platforms. Many of which I have written about other places on CD

Here is a brief history…
Vmx pi
Romi
Vex Edr
Micro:bit bots
Esp32 bots
2 different Amazon raspberry pi bots
3 or 4 donated tt motor drive bases with no encoders

There are four issues that arise.

1. The sensors are not at the right resolution or easily integrated with WPILib (vex Edr and tt motors)
2. The software is too outdated (VMX-Pi)
3. There are not enough accessible io ports (or power) to accommodate swerve (Romi)
4. There is a websocket layer between WpILib and the robot (Romi and xrp)

All of these issues can be overcome. In fact the last one is arguably negligible. However, to overcone them, we would have to dig deeper into mechanical, software, or electrical engineering or all three. If we were a team with more experienced engineering chops, we would have done that, but seeing our programmers struggles applying the wonderful 0-auto tutorials to our Romis showed me that if we can take the sample code and make as few adjustments as possible, that would be a better starting point for us.

There is an argument, and a good one, to be made that the extra steps will teach us more and make our future robots better. However, the frustration of too many variables is a real thing. We do not have an XRP, and we may have tried that, but when we realized that we can run the robot with 12 volts in a miniaturized format, we decided it was not worth the hassle.

There is an added benefit to doing it this way as well. You can pull the four bolts for the swerve modules, attach Scoot to a full-size robot, modify the wiring a bit (add a better battery connector and wire two more modules), and now you have a full-size swerve bot.

This can probably be modified to work with the XRP platform, but after the others posed hurdles, we did not see the point of investigating that solution (which would also be more expensive for us as we do not currently own one).

Having said all that, we continue to use the Romi, and will also use the VMX-Pi to train our programmers (we want to write companion tutorials to the 0-auto for Romi and Rev hardware), but for this exercise, we decided to go as tournament legal as possible.