pic: SimSwerve Omnidirectional Drive System

[CD47-Bot]

loading photo ...

[Tytus Gerrish]

OH... MY... Gosh! I would hate to me the machinest who was handed this drawing.

I love looking at this its an awesome machine featre i would drop to my knees and worship it if i saw it.

i forget, who was it who said Simple is good? not to mention... inexpensive.

OH!, i remember... EVERYBODY!

[ZACH P.]

May I ask why this is nessecary?

[Ian Mackenzie]

Originally Posted by ZACH P.

May I ask why this is nessecary?

We wanted a 'strafing' (omnidirectional) drive system, but also one that would be able to climb the deck. The solution we came up with was to have two driven, independently steerable wheels and two ball casters, each of which could raise and lower independently. The robot then climbed the deck by raising up each wheel as it came in contact with the deck. The main advantages were extreme manoeuvrability on th floor, the ability to climb the deck from any direction, and the ability to move sideways on the upper platform to defend against other robots.

Getting on the deck from any direction was particularly useful - if someone pushed us as we were climbing and turned the robot, we could still climb (usually, in fact, the pushing robot actually helped). The vertical sliding action isn't all that obvious in this photo, although you can see the two sets of linear bearings that the gearbox assembly slid on - check out the discusssion thread for the white paper on the design for CAD models in various formats.

Although admittedly complex, the gearboxes worked very well - we had one shaft snap once, but that was it. The design won Leadership in Control in Pittsburgh (where we were finalists), Delphi Driving Tomorrow's Technology in Long Island (champions), and both Xerox Creativity and Motorola Quality at the Canadian Regional (finalists).

[RogerR]

Originally Posted by Tytus Gerrish

...i forget, who was it who said Simple is good? not to mention... inexpensive.

OH!, i remember... EVERYBODY!

while simple may be good, effective is better.

i love how compact the drive module itself is (what!? they fit two motors in that thing!? no way!!), and how you used only two stages of gear reduction. my only question is, why did you guys use a 'hall magnetic sensor' instead of the more common light sensor?

[Ryan F.]

*wipes drool off keyboard*

Awesome drive system, and a great paper to accompany it. Though I would say our team would never use something like this....I'm still pretty impressed. The paper explained a lot of things to me that I didn't really get about omnidirectional systems. Thanks for posting!

[Karthik]

Originally Posted by Tytus Gerrish

i forget, who was it who said Simple is good? not to mention... inexpensive.

OH!, i remember... EVERYBODY!

Simple is good. There's no denying that. When this season started, we knew had the resources and the talent to build a more complex drive system. If everyone took the simple approach, we never would have seen shift on the fly transmissions or strafing robots in FIRST. These leaps in technology took place because teams were willing to take the risk to innovate.

One the more inspiring parts of FIRST for me, is seeing all sorts of cool innovations. It's partly why I come back every year. So when we had the chance to design and implement something new, I was completely behind the idea. I figured that we had enough experienced people, that if SimSwerve was a failure, we could quickly put together a simple alternative.

This is coming from someone who abhors complexity more than the average bear. I think it's very important for teams to work within their means, and keep things simple. On the other hand, if you've got access to good machine shop, have some great engineers and some sharp students, maybe it's time to try something new.