NI Week – the Jedi Challenge

Last year at NI Week we had an on-the-spot challenge between several teams, which was a lot of fun. This year they did this again, with some fun twists and turns. First, the table was a huge colorful structure that a group from Ohio brought (driving it overland!). With room to run four robots head-to-head (each in their own separate area), I think this greatly improved the entertainment value for the spectators (it also provided a semi-controlled “robot arena” the rest of the time, which was great fun for all), as well as making it tough to guess in advance just what the challenge would be. The robots were also more complex, with a Hitechnic color sensor facing forward and a downward facing light sensor, they were clearly able to not just line follow (but which line?!?) but identify what color object was in front of them. And perhaps most interesting, they had an elegantly designed “tweezers” on the front using the third motor, one that would close two long axles together, and then tilt back to lift whatever might be caught between them. I’m not sure who designed this mechanism, but it would have made a great example to show a lot of other folks. It not only illustrated studless construction very well, but also how to get two motions out of one motor – when the two long axles can no longer close, the now blocked gear train would then crank up at an angle, lifting whatever was in the “tweezers” (I wish I would have gotten better pictures of this… anyone?).

The actual challenge was to try to grab as many “Jedi starfighters” (barbell-shaped structures of red or blue bricks... kind of a squarish TIE fighter resting on one side) as possible within a short period of time, collecting them from a common area (the red "barbell" in the rear of this picture) and placing the red ones in one area (where the closer red “barbell” is in this photo), and the blue ones (not shown) in a nearby area, thus sorting them into two colors. We even had Leia and Yoda hosting (with a very entertaining running commentary). Teams were pairs of folks, in some cases one of the MCP paired up with a random person pulled from the audience (including kids). We had a short time to try to program our robots to drive down the entrance ramp, pick up a “fighter”, and deposit it in the “rebel base”. The picture shows what's more or less a starting set-up, but with one red "barbell" in the drop-off zone so it shows up clearly.

The results were interesting. Almost no one got line following to work well in the time allotted (my team did use line following, with some success, to find the pick-up point, but never quite got a perfect drop). Most teams ended up using dead-reckoning, which lived up to it’s name – in the short bit of time we had, it was extremely tough to get it reliably right and fast at the same time. Steve Hassenplug had the novel approach of coming down the ramp backwards, which helped some… but in the end even he didn’t quite get the reliability licked in the time allotted (although shortly after the official event, he had very good luck with his programming).

All in all, a really good time, with a lot of the crowd enjoying the event as well as watching the frantic programming.

I just wish I had gotten better documentation of the really nifty little rovers Laura was building for this challenge. I never seem to get enough pictures…

--
Brian Davis

Comments

Eric D. Burdo said…
I'd love to see that lifting mechanism up close. It would be quite useful in other applications too!
Anonymous said…
It's hard to tell from the photo, but the lifting arm appears similar to the arms on these robots (only a studless version): Jonathan Knudsens's Minerva - www.oreilly.com/catalog/lmstorms/building/minerva1.5/StepsM.html, Ben Williamson's Fetchbot - www.ozbricks.com/benw/lego/fetchbot/instr/part1.html, and/or Dean Hystad's ULK robot from the book "10 Cool LEGO MINDSTORM Robotics Invention System 2 Projects".
Brian Davis said…
Thank you! Yes, it is rather similar, and these are all excellent examples of it. I just wish I would have gotten a picture of the actual studless solution, because it was nice and elegant. I've used this sort of thing for grabbing arms before, in some cases a "grab and tilt" application like here, or in at least one case a "grab and pull" version (linear motion) on my Tic-Tac-Toe machine. Even though this is an old solution, it's still a very surprising one to most people.

Popular Posts