Dynamic Chassis with Holonomic Drive

Howdy Partinoids,

I was thinking about wallbots, and murdomeek’s use of the term “split” and viola.

Normally, in hononomic or mecanum wheelbases, internal forces aren’t very useful. But if you split the left/right sides of the chassis apart and connect them with slides instead of beams, you can have either side of your drive strafe in opposite directions to expand your wheelbase wider, or vice versa to fit through small spaces. You can still maneuver normally in all directions, but now you have the choice of expanding to act as a wall, or contracting to easily maneuver somewhere.

Diagram

:o!!! Oh my gosh… That’s like genius… hahaha. I can’t believe nobody has thought of that before!!! That design would fit my de-scorer on it.

Out team tried this for clean sweep, and it was a pretty good design. Unfortunately, it didn’t have too much application in the game, so we changed it. This year it might be a great idea. You just have to make sure that your sliding mechanism can take the torque and uneven stress that will be caused by turning.

I actually came up with a semi sorta design like this… but it expanded with rubber bands. I like this design a lot more.

Also, you’ll need a sensor on the sliding mechanism so you know exactly how much it’s pulled out. Otherwise the random stresses will very much affect your drive in autonomous. The driving algorithm will also become more complicated. It’s not really a problem, but worth mentioning.

Couldn’t you just have the sliders stop like 3/4 the way out using screws on each to stop themselves?

Well, changing the space the drive pods are from each other definitely influences the turning moment, but that only matters if your autonomous has parametric turning, which this year, with manual robot manipulation, isn’t really needed.

Awesome! That was the chassis design originally planned for 1492Z’s Honeybadger 3.0. It’s currently on the 4th iteration though after some design re-thinks. One major problem we had with a holonomic of this nature is that turning is extremely slow, because the drive turns into a rectangular holonomic instead of the ideal square. Similar problem was seen in the 2.x build with the fold out.

Yes, for wallbots, there’s a big tradeoff between turning speed and pushing effectiveness. If your wheels are near the center of your wall, you might be able to turn easier, but the other alliance can rotate you easier, and vice versa.

It’s not about stopping before falling out. That’s quite easy mechanically. I’m talking about preventing unintentional sliding when you’re driving. Certain maneuvers will pull apart your robot without you wanting that to happen. You have to code around this.

What about for programming skills?

Well, code can stop your drive from pushing together through normal driving, but there is still the chance of someone else forcing the drive together. You could have an adaptable program that depends on the distance between the two halves.

It’d be interesting to see an adaptable wall bot doing skills challenge well, because you don’t normally expect robots to diversify instead of specializing (in being a wall). But in programming skills, there’s still no real turning. Also, you could put pins in, locking the drive together because you don’t need the wall capabilities.

Hm, if you had a tower, where would you mount it? On the sliders? Would the drive have a chance of extending on one side before the other, leading to your robot being asymmetrical? Very interesting with the possibilities though.
How much time do you spend thinking about these things?

This was mainly for a wallbot/wall enveloper, so programming skills isn’t the main point of the design.

However, you could probably use pneumatics as a sort of locking system to keep unintended chassis changes to a minimum.

How about adding 2 more powered omnis in the middle? Each with its own motor. If each holonomic wheel gets 2 motors, 1 393, and 1 269, then the middle ones each get a 269. Then program them to turn with the outer wheels (Yes the inner wheels would turn faster, but it would help). Also a piston lock could easily be introduced, making it so you couldn’t retract while driving. Darn, I was a little too late.

Just an update, we built a robot similar to this for a scrimmage yesterday which competed as 4886C. The design worked very well, at least after we were able to get a power expander on it. It could almost always trap both opposing robots in their isolation zone. It also had the added feature of being able to hoard opponents objects’ between the sliders as it drove around. There were no awards, but one of the people running the event said that if there were, they would have given us the innovate award.:rolleyes: Anyway, I hope you don’t mind we used your idea-- it works very well. (It wont be going to worlds so you don’t have to worry about it).

Yes, and sliding while driving is not a huge problem. You just need a good driver (which I am not a great one). But if you have a good driver and the robot fully working, then yes, the robot is even better than Wallie V2.