Robot Arm C

[mrsteven]

Mhm interesting so it came down to documentation

What do we think theyre going to change for next year in the rules? Im guessing a height requirement and limiting the voltage of batteries slightly..

[harryk]

Mhm interesting so it came down to documentation

What do we think theyre going to change for next year in the rules? Im guessing a height requirement and limiting the voltage of batteries slightly..

I don't think there will be any change in battery voltage, most people are using servos which run 7.2v or less, and I doubt they would make the requirement less than 9v given that that is what most controller use. Though a height requirement would make things interesting, and really difficult, but interesting

[twototwenty]

Other than having four motors, what were the winning designs like? Was there anything particularly innovative/unexpected, or did it really come down t how well the arms were controlled?

[jander14indoor]

Nationals was great, many good robots, and great behavior on the part of all teams!

There were actually 7 perfect runs, one team lost points due to incomplete documentation. One team had more than 4 motors. The remaining 5 had 4 motors and I had to use the documentation tiebreaker. At this level all the documents were very good, but the rubric I used allowed me to seperate them and assign places.

Next year. Hmm, thought about this a lot, I'm not going to say much because this is a committee effort that then goes to the central NSO management and the rules can get changed along the way. I'd hate to mislead you and get you spending money or time on that basis. Robot Arm is still the intended event and I expect you will recognize it as such, but there will be significant changes you'll need to respond to to do well next year. So look for the books in Sept/Oct, anything you hear till then treat as VERY tentative.

As to what the designs were like, they were all over the map.
There was one tower crane concept with a very inovative pickup. Basically, it had a dozer blade on a metal tape ruler that extended and very safely scraped the pipes and pencils away from the edge into a box. The box then metered the objects into the goal or bonus box as needed. Metal objects handled by a magnet.
There was one mostly autonomous design. Manually controlled to pick up the bonus items and batteries. Then it picked up the rest of the goal box stuff completely autonomously and put it in the goal boxes. JUST missed perfect because it missed one pencil that was a little off location vs program. They went back to pick up manually and ran out of time.
The rest were mostly standard single or double arm designs with a joystick controller or master slave systems. A couple used pneuamatics to varying degrees of success. Fortunately no pressure bottles.

A comment on cost.
The master slave systems turned out to have very cheap solutions and at least two of these got perfect scores.
The cheap ones used masters that were crude models of the slaves with simple, cheap radioshack potentiometers to indicate position. One used legos for the arms and probably cost less than $5!!! Great out of box thinking.
Inexpensive arduino controllers translated the potentiometer info into servo commands. I heard numbers around $40 or so.
They also figured out a simple way to use smaller/cheaper servos. Don't reach heavy bojects to the far boxes, drag the far boxes to you! Two teams had a simple hook they grabbed with their gripper, Use hook to drag the far goal and bonus box in close, drop hook, pick up objects and never reach more than 40 cm or so, if that. Arms were shorter, stiffer and lighter. Servos much, MUCH smaller. And short distances cut down on time and need for great speed.

Of course there were expensive solutions, but in no way did they dominate the top scores.

Jeff Anderson
Livonia, MI

[RAntonello]

mrsteven wrote:
Mhm interesting so it came down to documentation

What do we think theyre going to change for next year in the rules? Im guessing a height requirement and limiting the voltage of batteries slightly..

Might I suggest an idea for a different second tie breaker besides documentation for next year? You could have a time-based time breaker, where a team that is complete can say "Stop" and the time recorded can serve to differentiate perfect scores. Technical Documentation would still be a third tie-breaker for teams who run out of time and have the same score. This would practically eliminate the problem of subjectivity due to documentation, as ties in both motor number and non-perfect score would be extremely rare.

[fr00tl00p]

Is it just me or does this seem like an especially expensive event? I payed for my Arm all by myself, and, yup, I'm the fruit who was using pots & lego's in my controller ^

And I agree with RAntonello, I think time should be the primary tie-breaker. It makes a lot more sense then the other ones. In fact, I think the number of motors shouldn't be used at all as a tie-breaker. (Cooler arms have more motors and can do more stuff, IMO )

And I'm also curious to know how detailed other people's technical documentation was. I myself modeled mine in SolidWorks and wrote a 1 page operating description and my parts list wasn't anything special, either.

Either way, this was definitely my favorite event. Robots rock!

[jander14indoor]

Tech documentation ranged from none to multi-binder presentations. I think 5-10 pages was pretty typical. Longer/more detail wasn't necessarily better if you didn't first give the critical stuff. Like I said, one perfect run had a very nice presentation, except the drawings had no labels. Had to guess what the power supply, motor, etc was. Scored incomplete, 5% penalty. Of the 7 perfect runs, only one was very long. The winner's wasn't longest, just met the rubric better. Oh, and one or two of the top scores had hand drawn tech drawings. Very nicely done and detailed, didn't cost them any points.

As to cost, you certainly could spend a lot. I know some devices exceeded $1000 because the technical documentation listed the costs! I know some others were under $100 and very successful. Like all tech events, they cost more than the testing events. But the low end wasn't crazy expensive and the high end had no better success than the cheap solutions. At least two of the perfect runs were on the low end for cost, possibly more, memory blurs after watching 60 robots!

And that lego master wasn't fruity, it was brilliant. I loved it.

Jeff Anderson
Livonia, MI

[Jdogg]

-- Deleted --

[nelran]

"There was one mostly autonomous design. Manually controlled to pick up the bonus items and batteries. Then it picked up the rest of the goal box stuff completely autonomously and put it in the goal boxes. JUST missed perfect because it missed one pencil that was a little off location vs program. They went back to pick up manually and ran out of time."

Jeff Anderson
Livonia, MI

I watched some of the building events, and the organization was excellent. In particular, the RA event was very well planned with 3 set ups, allowing multiple side by side runs comfortably. Competitors got enough time for evaluation and test runs. Even the watching area was nice: the events could be observed from above or at ground level, very close without interfering with the event.

Just to clarify, the alluded autonomous RA that Mr. Anderson had described had two operation “modes”: manual mode which was used for aiming and releasing objects in the bonus box, and the rest of the items (including batteries) which were picked up autonomously by running a pre-programmed sequence (macro). Sadly, in the official run, one of the pencils was out of the position so the RA missed it during the autonomous sequence. The “macro” sequence was carefully planned and tested tens of times to obtain the most efficient and fastest way to pick up all the objects. The programming involved about 120 “steps” (coordinated movements) and also assumed that objects were at the right locations (coordinates) as per the rules manual, as the official, printed layout was used which can be downloaded from soinc.org. However at Nationals, the layouts were made with blue tape while the objects’ and goal boxes’ positions were not clearly marked (or at least not as clearly as in the official layout), so it could have happened that one or more of the objects/boxes were out of place and neither the proctors nor competitors realized such at the moment of the run. Usually this is overlooked at events because the overwhelming majority of RA are manually operated (if I can recall well, none of the other 60 robots were autonomous); so in the case of the other, manually operated Robot Arms it is not a big problem: the operator can “compensate” or adjust for any difference at object’s/box location; however, for an R.A. using a pre-programmed sequence, it is critical to make sure that objects are placed in the right (expected) positions in the competition area (even if robot’s gripper aperture—about 1.5”—allows for some slight misplacement of the objects, but not of the boxes).
This can be easily addressed if the Robot Arm Official Rules emphasize/implement the use of the official layout as the “Standard” Competition Area to be used for any tournament (whatever future dimensions which are to be selected), which can be easily downloaded and printed from the website. This way, one can assure a consistent competition area; making it easier to set up and “reassemble” the area as exactly as possible after each run. Another suggestion could be to include in the rules that the locations of each object and boxes must be clearly marked (but how that would be defined could be a little difficult).

[jander14indoor]

Hoping that some folks are still following SciOly over the summer. I'd like to leverage the expertise/ideas that exists within the SciOly community.

We're adding an open ended scoring element to the event for next year (I hate technology events where perfect scores are possible!) to minimize need for tiebreakers. Since its still draft, I'm not going to get to specific since it might change, but...

Here's the concept. Pick up one of the objects and raise it as high as you can while executing the rest of the tasks. The higher you go, the more points. Scaled to be significant, but not to swamp the rest of the tasks. My concern is the students will surprise us (as they routinely do) and we'll be seeing heights in the multi story range at nationals.

I'm not concerned with measuring height accurately till you get above reasonable step ladder range. Then I need to start looking at indirect methods. Now I'm not arrogant enough to think I'm the one with the best ideas on how to measure, so...

Question, how would YOU measure the height accurately. In detail. Low to mid range. Really crazy high.
Alternatively, can I make the students responsible for displaying height reached? If so, how do I validate what they report?

Thanks,

Jeff Anderson
Livonia, MI