Texas Vex observations

[jgraber]

Here are some random observations from the FVC competition at UT Arlington Texas.
1. A team from Newton Kansas drove down. One school from Florida flew.
2. My first thought on walking into the room: "We are way outclassed", then later, "We'll do ok with our forklift bot."
3. 21 teams, ~7 complex bots (~4 of which actually work), ~7 RC carbots.
4. Berkner Robotics (the team I mentor) won every qualifying round and took number 1 seed, through luck of good partners and less challenging or sandbagging opponents, then lost every match in semi-finals(to tie for 4th?)
5. Plano West had two teams, seeded 2 and 4. Berkner picked #2 as alliance partner, they declined, so we picked #3. Then #2 picked #4.
I keep wondering what might have happened if we had skipped #2,#3
and went directly to #4. If they declined, they would not have been able
to be picked by #2.
6. Through similar mergers of high seeded alliance captains, the original #12? seed became the #8 alliance captain. (be prepared!)
7. By third round pick, every alliance had a simple starter kit carbot for third partner (be prepared with your favorite of the bottom 8!).
8. Check-in procedure did not include checking the template.
9. Some teams did not bring a computer, and didn't know how to load
a program.
10. Competition staff were primarily college students unfamiliar with the rules.
11. At one point the Atlas ball was ejected, and a ref tried to kick it back in.
12. There were no calls against agressive behavior, out-of-size starting positions, or preloaded balls touching the ground.
13. Scoring was similarly imprecise: Atlas ball still on the platform at end of match did not count, even if off center. Credit give for low goal balls even if robot was still touching them.
14. Few hanging attempts, and fewer still (3?) successful ones.
15. Traction is always a problem. Need to do tractor pull experiments next time, with various wheels, wheel covers(traction mat), etc.
16. Our team bought an aluminum kit to use for the arm, very light and good.
17. To make it easier to hang, we converted the wheel frame to aluminum as well, so we got pushed around and had poor traction (bad). A volumetric/cost analysis of Vex parts to use as weights would be a good idea.
18. Our meet set a new national high score for FVC, 160pts, pre FVC rep.
19. Once or twice, the control system apparently re-trigger autonomous mode during Operator Control time.
20. A big box bot was once disabled while scoring, by having a ball shoved under it.
21. One of the high-seeded big box bots had a successful hang; It can climb on the platform even though it has small omni wheels in front, by jerking forward and back until the front wheels can jump up on the platform.
22. Of course, the next time it tried, it tipped over onto its back.
23. Tactics used during qualifying rounds were mostly different from tactics used in finals, which were much more agressive.
24. Some teams have big budgets, some from parents, some from corporate sponsors.
25. Chain drive on wheels is unreliable.
26. A viable hang method is to hang from the corner elbow, lift up past the lower coupler, then sag down to catch on the coupler.

[gblake]

jg,

To try to get to the root of some of the problems you described, my current hunch is that teams dramatically under-estimate the time required to carry out even one iteration of a few design->build/program->integrate->test->repeat cycle; and that going through that process at least twice is necessary to put a decent non-trivial robot on the field.

Did you get a sense that the teams that struggled either got started too late or didn't invest enough time each week?

Or was it more a case of needing more sophisticated advice from coaches/mentors?

Or was it more a case of the Vex kits and software being too complicated (needs better manuals/instructions)? or too hard for students who need to continue their educations a bit more before attempting FVC?

Or something else?

[gblake]

jg,

Can we expect to see some videos or still pictures? Did you see many cameras in action there?

What robot designs contributed to that very high score? Did one or two excellent machines hit their stride that match?

Blake

[jgraber]

I'm not very sociable in person, so I didn't talk much with the other teams.
Our team did well for the amount of time they put in, by copying the basic idea of a robot off a picture from the web. That helps save some iteration. We had time for several gear train revs, and I had prepared and guided them along a path of known pitfalls (small gear teeth stripping, axle holes slipping) that I picked up from reading here and on chiefdelphi. I dont see any other Texas people writing here, so my guess is that the non-web-research-enabled teams are missing a good resource.
I was more perturbed by the game management and variation in rule enforcement, which has nothing to do with the above. I thought I'd comment on it to get comparisons to other remote tournaments, or to prep others about potential pitfalls, and ways they can help.
To explicitly count your questionmarks and reply:
Iteration cycles + time? Maybe (as above, I didn't poll other teams)
Need better advice? Probably, coaches should read more here and misc web.
Vex too complicated? Maybe. The inventors guide has nothing but squarebot. The old erector set guides at least showed pictures of misc stuff that could be done, even if they were more "looks like a plane" type than "working model of mechanism X" type.
Something Else? Things I struggled with our team: Keeping kids focused, getting them to write in the journal (the way I would do it is different than the way the other coach got from the web), failure to sub-componentize so that both people could work at one time (Should have had removeable arm-frame, so that gear work and chassis work could be independent. Scheduling time amoung other activites, and during weekends when school was open. We missed 10% of our scheduled days near the end due to snowdays (in Texas!)

[jgraber]

I'm not good with processing posting pictures,
but I'll see if I have anything suitable tonight.
You might be able to find a web site for Plano Texas, West Plano High School Robotics, or whatever their two teams were, to see if they have documentation for their robots ("Hammered Steel" and "Riveted Steel").
There was a brief segment (I was told) on Channel 8 news.

Contributors to high score: The high score robot had a large square basket that could scoop up a whole stack of balls at once, by backhoe-type action.
Then drive to goal and relatively quickly lift and tip out the stack.
The basket was independently rotation powered from the arm.
I think it could do one stack in autonomous mode as well, which gives some idea of the speed (two dumps in 20 seconds) 160 point breakdown was something like: 6b x 3pt = 18 + 10bonus = 28 points in autonomous mode.
Remaining points from pickup 5 more stacks to the high goal, then 2x ball.
Against a better opponent, they might be limited to a total of 4 stacks plus control of the 2x ball to win. I didn't actually see that match, so I dont know how much their alliance partner helped, if it was scoring or defence, or offence or 2x ball control.
A few at a time strategy seemed to work better in general than "all or nothing", since time goes fast, and opponents get aggressive late in the match.

[Brandon Martus]

Here are some I took with my (low quality) camera phone.