Elevator robots why didn't you..
 

I've been really curious as to why the elevator robots didn't just dead reckon putting a vision tetra on the center goal.. you have a 1/4 chance one will lie in the middle path.. drive (through both possible vis tetras on the center row) then raise then drive alittle more and drop - back away

Re: Elevator robots why didn't you..
 

I am guessing that this is because of the 1 in 4 chance of success. Most teams want an autonomous mode that will perform consistently, every time. Most likely, alliance captains would not want to pick a partner who would only have a one in four chance of having an autonomous mode that actually works. Even though it may work some of the time - and when it would, it would be wonderful - it probably isn't the most appealing characteristic when the time comes for alliance selections. It just isn't reliable.

This actually makes me wonder how it would be for the field attendants who set up the field. If they knew anything about that robot and their strategy, they would potentially be altering the outcome of the match based on how they chose to position the tetras. How are the positions of the vision tetras determined? Is there a set schedule of how the vision tetras are going to be placed match to match? Or do the field attendants make a random decision as to how they are placed? I am not implying that a field attendant would purposely place tetras in a manner which would cause the alliance to win or lose, but I wonder if they would feel a degree of responsibility and what the implications of this would be.

Interesting question.

-- Jaine

Re: Elevator robots why didn't you..
 

I believe the computer randomly picks it. If you notice, the guy behind the computer holds up 2 numbers at the begining of every match.

Re: Elevator robots why didn't you..
 

That is correct,
there is a random number generator that generates 2 numbers when you click a button, the score keeper or their assistant then relays that number to the people responsible for setting the tetras and they place the tetras.

Re: Elevator robots why didn't you..
 

I think they're chosing randomly from like a hat or something...I seem to remember them saying that at the kickoff...

Re: Elevator robots why didn't you..
 

well yes I know this.. but think... the teams that have tried to cap in auto with the cam have done less than 1/4... more like 1/6

Re: Elevator robots why didn't you..
 

I was wondering... in the case of a match replay shouldn't the vision tetra's be placed in the same place as the first match thus creating the exact same feild placement. I only ask because at GLR in the last round of eliminations we had to replay a match and the vision tetras were placed in a different location. this is imporntant because team 66 almost capped the goal during auton in the first match using the camera, but with the different placement of the tetras it was much harder. anybody know the answer to the question: Are the vision tetras supposed to be placed in the same mannor as the first match in the case of a "re-match"?

Re: Elevator robots why didn't you..
 

In the case of a rematch (or restart of an aborted match), the vision tetras are repositioned before the restart. This is to prevent teams from altering the state of their machines (repositioning, reprogramming, realigning, etc) after they know the location of the vision tetras from the first try at the match. To leave them in the preexisting postions would give the teams competing in that match an unfair advantage over all the other alliances, by allowing them access to knowledge that no other alliances would have in other matches (i.e. the vision tetra locations).

-dave

Re: Elevator robots why didn't you..
 

it makes 100% sense to me that they should be placed in the SAME location~
There is no written rule, but I bet if you approach the ref before a match they won't make the mistake

Re: Elevator robots why didn't you..
 

i've been wondering this exact thing. think about this

there are 2 vision tetras

there are 8 locations on the field for those tetras, so at any given match there is a 1/4 chance that they are in a given position.

there are two positions lined up directly in front of the center bot starting position (positions 2 and 7). so multiplying 1/4 by 2 you get 1/2 or a 50% chance of a vision tet being directly in front of the center bot starting position.

what if you have your robot drive forward to where position 2 is and attempt to pick up a tet. a limit switch/sensor/something tells the bot if it picked it up or not. if it did, it drives forward a little more and caps the center goal. BONUS TETS!!! if it doesn't get one in position 2, it drives forward to position 7, same thing happens. if it gets a tet, it goes up to cap, if not, no harm done and you're already to the middle of the field for the start of the match

after watching numerous autons this last weekend in chi-town (65 going to the grocery store, et al) this doesn't seem like an unreasonable request, especially if you have a 50% shot at getting it.

Re: Elevator robots why didn't you..
 

The probability of at least one vision tetra being only the center line is actually slightly lower than 50% because of the case where both vision tetras are on the center line.

There are 56 possible permutations of tetra combinations. 26 of these combinations include cases where at least one tetra lies on the center line. 30 combinations include cases where no tetras lie on the center line.

So, there is a 46.4% chance that a vision tetra lies on the center line at the start of a match (either do the brute force division or subtract the 100*(2/56) percent chance of two tetras starting on the center line from 50%).

Personally, the 3.6% discrepancy doesn't matter much to me, strategy-wise. If a robot can cap a vision tetra on the center goal at the start of a round roughly half of the time, it is probably worth it to try - on one condition. That condition is that no matter what, your robot puts the starting tetra for the alliance on the center goal regardless of whether or not a vision tetra lies on the center line. If a vision tetra is on the line, the robot puts both up. If a vision tetra is not on the line, then it puts only your starting tetra up. Either way, you start the round with at tetra on the center goal, so you really can't go wrong.

Re: Elevator robots why didn't you..
 

Going for the tetra that is lined up in the center is a very good strategy, but there are also a few problems with this.

1. What happens if there are two vision tetras lined up in positions 2 and 7? You would pick up the tetra closest to your starting position and head for the center goal. The second tetra would then block your path resulting in not capping the goal in autonomous.
2. Do your teamates have reliable autonomous programs that use the camera? If they do, and your robot picks up the vision tetra placed in front of you, your teamate will either track the tetra you pick up, or possibly run into your robot while trying to pick up a different vision tetra. In either case, this would more than likely delay your teamate (and your own robot???) from reaching a goal in time to cap it.

Some robots that would be able to pick up the vision tetra lift too slow to pick it up and place it within 15 seconds so they have decided to not waste their time and either do something else that will benefit them in the long run or do nothing at all.

Re: Elevator robots why didn't you..
 

Keep in mind that only a few teams have been able to cap with vision at all. Not to mention reliably. A program that can work even a third of the time would be way ahead of most teams.

And on the issue of repositioning tetras at the beginning of a rematch, it would seem highly unlikely to me that teams would have a way of selecting what position the tetra is in considering that they knew before hand that they wouldnt have that ability. I can see why it is the way it is, but it would make more sense to me to have a rematch the same as the original match and just rely on any team that does have such a selector being GP enough not to use it.

Re: Elevator robots why didn't you..
 

As a team with an elevator robot--the only reason we didn't is because we were unable to test our autonomous before shipping, and it didn't work once we got to competition. We actually had 9 potential autonomous modes, and 3 of them involved driving forward, picking up whatever tetra may be in our path, and dropping it on either the center or one of the side goals. Too bad it didn't work.

Actually, we're still trying to figure out what went wrong. Since our competition (St. Louis) we've tried several variations, but have yet to figure out why it's not working. At least it gives us a good programming excercise for our offseason.

Re: Elevator robots why didn't you..
 

The thing is that many teams probably had more on their mind and were thinking about how they could use their autonomos to their advantage. If you think about it you have like a 30% chance to have the tetra right in front of you. Now if you have any autonomous to go to a loading station you already know that you can start with a tetra. Instead of dead reconing , it seems that you could instead start off where u know u can get a tetra and have a faster start on the goals.

Re: Elevator robots why didn't you..
 

If you alliance partner has a reliable autonomous program (more so than yours) I'm assuming you would be able to communicate this before the match and switch your autonomous to something that doesn't interfere with the other, and if yours was more reliable, they would switch theres. That seems like the most logical solution to that problem.

Re: Elevator robots why didn't you..
 

I think the reason teams don't cap vision tetras in autonomous is because the chance of successfully capping the tetra would be a less than 1/4.

The tetra's are placed in certain locations in the beginning of the match but the Field reset people are only human. Therefore, the tetras may be offset a bit which could throw off a robot.

Furthermore, if a robot does obtain the tetra then the robot still has to make its way over a goal and cap it. Robots starting out with a tetra couldn't even cap every time.

Re: Elevator robots why didn't you..
 

Our robot is not an elevator robot but we are going to work on a blind vision tetra cap. In Sacramento, we were able get our robot to knock off the hanging tetra, and in silicon valley, we were able to cap a non-vision tetra. We were able to do this because our robot was so reliable. We ran out of things to do in the pit between matches so we gave the opportunity for our programmers to shine. Our capping program at silicon valley helped us get the GM Industrial Design Award. At Atlanta, we anticipate, plenty of time to add to our current program and create a blind vision tetra cap. Why would we have one? If all of our robots can knock down the hanging tetra or cap the non-vision one, one robot has to sit out in the middle. So if our robot has a chance of capping a vision tetra, it is still better than doing nothing.

Re: Elevator robots why didn't you..
 

If you'd like to see videos of two *very* close attempts at the center goal with a vision tetra (along with the initial tetra!), download "20050312 Vision Tetras.torrent" from my BitTorrent tracker at http://ldeffenb.dnsalias.net:30049. Remember, this was our first and ONLY regional competition (we're not a rich team).

We have a camera on board, have code to identify all 8 possibilities from each of the three starting points, drive paths to acquire and deliver those tetras to the center (and, in some cases side) goals, only to discover that our drive and lift were slow enough that we only had a chance at delivering the center 2 positions to the center goal.

That's when we decided to use the motion to handle the initial tetra as well (hey, 3 points and owning the center in autonomous is better than 0 points looking like a sitting duck). (Want to watch painful replays? Download "20050310-12 386 Auto SloMo.mpg.torrent" for slow motion reruns of all of our automous runs. They'll make any software team cry from the effort wasted when software meets hard(ware) reality!)

We now know that we're tilting too far towards the goal and that our lift motor backdrives (watch closely in QF4-2) and those problems will be corrected in Atlanta. If no one does it before that, it'll be a race to the finish as far as Team Voltage 386 is concerned!

We *will* do it (vision tetra onto center goal) in Atlanta!

Lynn (D) - Team Voltage 386 Coach and Software Mentor

PS. Download a BitTorrent client at www.BitTorrent.com or www.BitTornado.com (Windoze only).

PPS. If you download 20050222 Tetras.torrent, you'll see us actually do a 2 or 7 vision tetra in autonomous just before shipping.

PPPS. We *will* succeed in Atlanta! Look for us. We'll be the team waving the Tiki Trophy!

PPPPS. Our team statistics expert had calculated the 46.4% probabilty just before we shipped to help the software team get out of the dumps based on the slowness of the robot after we worked hard at programming all possible combinations of starting and tetra positions and target center-line goals. We had hope! And with an updated tilt gear to be installed in Atlanta, we *will* do it!

Re: Elevator robots why didn't you..
 

Even with the 46.4% chance that there will be a 2 or 7 vision tetra, it is *extremely* hard to get a robot to drive straight, even with encoders and gyros, for 27 (actually about 20+) feet.

You do the math, if your 9" diameter wheels are off by 0.1 inches, how much horizontal difference will there be between the left and right sides after 20 feet of matching wheel rotations? What position will your robot be in when it drives through the positions 2 and 7 and what will the orientation be when you arrive at the center? What tolerance do you have to actually get the tetra on the goal in a scoring (meaning you can't be touching it) position?

And can you do it all in 15 seconds?!?!?!?

Contrary to Team 25 (offering the Tiki trophy), we *know* it *can* be done, but a bunch of semi-controllable variable have to align *after* the 46.4% has been achieved!

Lynn (D) - Team Voltage 386 Software Mentor