View Full Version : Can you score in autonomous yet?
JBotAlan
17-02-2007, 22:18
Just out of curiosity:
Don't lie. Don't bend the truth.
What is your 'bot capable of doing during autonomous right now? Select what it does best if more than one option fits:
hang a tube under a vision goal
hang a tube under a non-vision goal with vision aid
hang a tube regardless of vision (using other sensors besides cmucam)
defend the goal blindly
sit and look pretty
deploy the ramp in the home zone
position the arm/whatever scoring mechanism ready to start the match
something I haven't listed...please, do tell!
Just thought it would be fun to ask.
Our 'bot can't do much yet, but it should be able to at least hang under a vision goal and defend tomorrow.
JBot
pheadxdll
17-02-2007, 22:22
Our robot is pretty jerky now but it can hang a tube under a vision rack almost most of the time. We just need to tune our PID routine. Well back to work... :)
Greg Marra
17-02-2007, 22:22
Team 230 capped in autonomous mode at the Winter War Zone scrimmage today. The Blue Alliance will be posting footage of the event later.
nope cant score jack in auto . . but neither will any one on the opposing alliance. For better or worse we've written the most ruthless defensive auto code Ive seen in 4 years of watching/participating in first events.
benhulett
17-02-2007, 23:57
Our bot sits there looking for a green light....and keeps looking for a green light.....so unfortunately, we can't test our autonomous at the scrimmage we're going to tommorow..unless I can code it correctly (on the way to D.C.) and convince the mechanical team that the autonomous mode won't make the bot go into full foward and destroy our arm :)
Tom Bottiglieri
18-02-2007, 00:01
Right now our autonomous mode sits and looks pretty. When I mount the encoder to the arm tomorrow, we should be wheelin' and dealin' within a few hours.:cool:
Rick-906
18-02-2007, 02:16
i seem to be one of the few people on my team to consider good auto play an important feature in a robot... it's what got us through last year.
im suggesting it as a last minute "side project"
kitscuzz
18-02-2007, 02:23
We just managed to get the robot up and running today :/ And by "today" I mean at 1:15AM
We should be able to get some decent autonomous going before regionals, but we really just need time to figure out how we're going to get it on at all, let alone during autonomous.
That field-oriented holonomic is looking better and better...
paulcd2000
18-02-2007, 08:37
Wow, right now our robot can't even get to the light accurately. Encoder troubles, mostly. But it will hang a ring under a vision goal, i hope.
So far our robot can try to block the other alliance's robots, but we still need to work out many, many bugs. Mainly the fact that it only makes it halfway to the rack right now...
We're also trying to get it to score on the rack, but we first need to get our camera to stop showing us the "Red Light of Doom" (thinks it sees a light no matter where it's pointing).
Doug Leppard
18-02-2007, 12:10
What we can do today:
In practice match we capped several times yesterday.
We can cap low and middle spider legs.
We can do it from left, right or middle position.
Can distinguish lights when seeing two and decide to go to left light or right light.
Decided late last night not to go for non-lit leg.
Now working on what robot does after capping.
We can do this much because hardware guys gave us a bot to work on with over a week to go. Thank you hardware guys!!!!!!!!!!!!!
Ours not only sets up the mast but also sits and looks pretty... ;)
We sat and looked pretty at the CT Scrimmage yesterday, since we were primarily tuning our drive systems and training drivers. But our autonomous code is actually working fairly well (just too much risk with everything else yesterday to run it for real, although I would've liked to).
Kendal Reed
18-02-2007, 14:00
We deploy our huge arm and then kind of sit and look pretty. No CMUCam for us.
*looks at above posts* Holy **** it's Paul! Hi Paul!
jesusescobar
18-02-2007, 17:10
mine follows and drives to the light thats what it does best
Elgin Clock
19-02-2007, 14:39
A preview of one of our auto modes so far:
http://www.chiefdelphi.com/forums/showthread.php?t=54483&highlight=237+auto
Btw, the poll needs to allow for multiple options.
Our autonomous mode works great here at our work shop. When we went to the scrimmage it was not so good. We are looking at the Noise setting on the camera
Kingofl337
19-02-2007, 22:38
You should add to the poll
Mechnical team is still working on the robot. :eek:
I just got the robot today to program. :D
The robot breaks faster then I can program it. :p
Our autonomous is working very, very well on our practice field. Here's to hoping the gremlins don't strike in Manchester.
d.courtney
20-02-2007, 00:53
our robot can score in 6-10 sec 90% of the time... when placing the tube it will pull back with enough force that it shakes the rack (hoping to effect other teams success) our wait time is variable as to allow enough time if an ally needs to score.it will score under the rack, and on any spider height.
our robot can score in 6-10 sec 90% of the time...
That's slightly faster than us, although some more practice time could speed things up. And we have the adjustable delay as well, but haven't really tested it.
Ours does everything BUT correct towards the light. We can see the light, get the keeper, position the arm, and go toward it the light, but if its not dead on and the robot tries to correct, it freaks out. Our programmer said that that part of the code is whats breaking autonomous. I'm assuming he'll have it sorted out by the first day of GLR.
You should add to the poll
Mechnical team is still working on the robot. :eek:
I just got the robot today to program. :D
The robot breaks faster then I can program it. :p
This would be our option as well.
We currently have 9 different autonomous modes written, and we've used a test robot to prove that our code for turning/driving to a light and using a gyro works correctly. However, we haven't been given any time to tune the PIDs on the new robot or test the rest of the code, so we won't know how much we can do until the end of our regional, the way things look now.
Well...we've got a code mostly written, and we have the calculations done. Now all we need is for that darn camera to actually find the light...
bjimster1
21-02-2007, 00:03
We've done it once. im not counting on it, it'll be more like gravy. I think we will be able to get about 50% of the time
Well, our team is theoretically capable of hanging, but on the last day we had an exclamation of "Ah!!!! Sparks on the camera board!" It now refuses to work in any usable fashion (power light comes on, but it either chooses to continuously "see" the light or never see it.)
Assuming we steal/borrow a camera from someone, it should work.
We did not get our autonomous mode working... Can anyone tell me if we need to move out of the home zone or can we just move so many feet and then drop our keeper?
We did not get our autonomous mode working... Can anyone tell me if we need to move out of the home zone or can we just move so many feet and then drop our keeper?
You can just sit there and shake the keeper off after autonomous is over.
mormannoob
22-02-2007, 18:23
we never got to test it but heres is the plan if everything happens perfectly
1) hang keeper
2)go to end with ringers and pick one up
3) score said ringer
we never got to test it but heres is the plan if everything happens perfectly
1) hang keeper
2)go to end with ringers and pick one up
3) score said ringerGood golly. Can you really move that fast?
(Sorry if that's a little off topic. :))
Eldarion
22-02-2007, 18:31
We should be able to score on a non-lit arm...hopefully. That hasn't been written yet. :)
Dave Scheck
22-02-2007, 18:42
We should be able to score on a non-lit arm...hopefully. That hasn't been written yet. :)Be careful of when you write this so you're not in violation of the Fix-It window rules!
Eldarion
22-02-2007, 18:50
Yup, I know.
To be more specific, it hasn't been tested yet. (We all know how well that works :) )
katiyeh07
22-02-2007, 19:17
we never got to test it but heres is the plan if everything happens perfectly
1) hang keeper
2)go to end with ringers and pick one up
3) score said ringer
all in 15 seconds? are you serious??
good luck with that:rolleyes:
We can score on the legs under the lit column, and add a 1-5 second delay at the beginning of it whenever we want, but we decided that unlit there were just too many variables to do it reliably without rewriting the entire tracking.c.
Anyway... we do have another that we had working some of the time, run the the side of the goal DR, and then track to under a light and score on the mid or lower leg on either side of the rack, so we're not bad off.:)
mormannoob
22-02-2007, 20:04
all in 15 seconds? are you serious??
good luck with that:rolleyes:
yea it was all donut's idea
he thinks its possible
lukevanoort
22-02-2007, 20:17
Right now, nothing. The code has been mostly written, but not tested to score a keeper on the top, lit spider. If the camera fails to get a lock, the robot should drop its keeper and position itself and the arm to receive a ringer at the beginning of the match. Whether or not the camera part works, the positioning bit should be operational.
Our bot can read camera data, but does nothing with it.
Our bot has code for a gyro to go straight, but no gyro.
Our bot has the ability to choose up to 16 autonomous modes, but only the Do Nothing one is programmed.
You can just sit there and shake the keeper off after autonomous is over.
Do we have to move out of the home zone before shaking the keeper off?
Do we have to move out of the home zone before shaking the keeper off?
Definitely not.
Are there any rules about autonomous mode and if so where can I find them?
Can we simply shake off the keeper and set our robot up and drive over to where we will pick up the ringer before the autonomous mode is done?
Dave Scheck
22-02-2007, 23:39
Yup, I know.
To be more specific, it hasn't been tested yet. (We all know how well that works :) )I figured you'd be careful...I just wanted to send out a reminder to all teams....Your post just happened to be the one that I responded to :cool:
Depending on your interpretation of the rules, even testing needs to abide by the fix-it windows. I'd be interested in finding out how many teams are actually going to follow this.
Eldarion
23-02-2007, 00:48
I figured you'd be careful...I just wanted to send out a reminder to all teams....Your post just happened to be the one that I responded to :cool:
Depending on your interpretation of the rules, even testing needs to abide by the fix-it windows. I'd be interested in finding out how many teams are actually going to follow this.
So would I...I know we won't be testing until our first regional anyway, as our robot is down there.
we never got to test it but heres is the plan if everything happens perfectly
1) hang keeper
2)go to end with ringers and pick one up
3) score said ringer
Good golly. Can you really move that fast?
(Sorry if that's a little off topic. :))
all in 15 seconds? are you serious??
good luck with that:rolleyes:
We have the sensors and code to pull it off, and I think we have the speed to do it (drive wise at least, 12 fps lets you get across the field in 5 seconds or less). I doubt it'll happen anytime soon, as we're going to be hard pressed to get enough testing time just to get hanging the keeper working, but we can always dream.
Besides, a mentor has offered to cook the programming team dinner if we can do it, there's hard motivation in that.
Bharat Nain
05-03-2007, 15:38
So, week 1, not too many capped in autonomous, even fewer using the green light entirely(including us). Any observations, explanations etc?
Lil' Lavery
05-03-2007, 15:46
So, week 1, not too many capped in autonomous, even fewer using the green light entirely(including us). Any observations, explanations etc?
At VCU, 384 and 1086 both scored keepers using the camera, and 540 came close several times. 116 and 1610 both scored "blind". I believe that those were the only 4 teams to score keepers.
116 attempted to score blind 10 total times, only succeeding once (would have been twice if 2068 didn't land a perfect autonomous block in QF 3.1). 384 scored 2 keepers iirc, and 1610 and 1086 scored 3 iirc.
116 will be upgrading and fine tuning our autonomous (and hopefully finishing incorporating our camera) for Championship.
Eldarion
05-03-2007, 16:20
1625 tried many times at the St. Louis regional, scoring only one "virtual tube". :D (The tube fell out of our gripper at the start of autonomous, and the robot proceeded to perfectly maneuver and move the arm such that it would have scored if the tube hadn't fallen out. FRUSTRATING!) We were having lots of mechanical difficulties with our arm, though, so we couldn't really even score in manual mode very well.
375 scored 3 keepers in New Jersey, with 1 more being blocked by an alliance partner, and 2 more positioned perfectly, with less than half a second needed to drop.
This was all using the camera, and I hope to retune, making it better.
Samuel H.
05-03-2007, 17:19
We were 4" off in our final match (using camera). We plan to have the final bugs out at Boilermaker (of course :) ).
Good luck,
Sam
Vexkingof716
05-03-2007, 18:02
We got one keeper on the bottom rack using dead reckoning. we were one of only a few teams at new Jersey that was able to score in autonomus.
In the name of honest accounting...
Our autonomous never worked in competition, since something was causing the infrared sensors to trigger prematurely. We've since adjusted it, but it will be Atlanta before we see if it works for real.
so far our bot will ram the rack, we have been brainstorming to make the camera work and it might come competition time but as it stands it will run into the rack hopefully stopping all other autos from working.
lukevanoort
05-03-2007, 19:16
In the name of honest accounting...
Our autonomous never worked in competition, since something was causing the infrared sensors to trigger prematurely. We've since adjusted it, but it will be Atlanta before we see if it works for real.
On a similar note, ours didn't work either. We also had sensor trouble all through the first couple of days (our arm actually ripped through one of its pots during inspection, afterwards the inspector looked a bit wary) which got in the way of getting the PID to work. We will be working at Palmetto on Practice Day to try to get something happening, probably just rotating our arm into playing position. (its on a turret) I still really want that keeper on the rack though...
987HighRoller
05-03-2007, 22:02
:rolleyes:
cziggy343
05-03-2007, 22:08
our robot is under the "sit and look pretty" option for now. we may change that though. but i no that 1610, 384, and 1086 all hung keepers at VCU. if i left any robot that did it, let me no. in fact, 116 may have done it once.
We burned one of our fix it windows reviewing stuff today, and we're happy to say all our auto-command table code is working and the auto arm positioning was doing just what it should (first try too! Since when does code EVER work on the first try!?). It wasn't tuned perfectly, but we were able to reduce most of the overshoot with about 2 minutes of tuning, wasn't much point in doing more than that since we'll have to retune for the real robot (and not a crappy hacked '06 bot mockup) at Arizona.
I didn't see alot of keepers scored week 1, but I was happy to see at least a few got it working (or tried anyway), I hope we can continue the tradition and if things continue to fall our way like this we may even get to try some of that crazy double ring score we wanted.
Doug Leppard
06-03-2007, 07:16
We compete in a few days. During practice we did it 10 times in row and did it at a practice match twice no problem.
See it in action:
http://youtube.com/watch?v=iBQoGrs1XKQ
We will see how it really works at competition.
We had hardware delivered a week and half before ship date. We spent 60 hours perfecting the auto mode. It is not easy.
So why didn't many succeed? Many because software guys did not have enough time with the bot, thank you again hardware guys for delivering a bot to us early!!!!!
So, week 1, not too many capped in autonomous, even fewer using the green light entirely(including us). Any observations, explanations etc?
Our robot has an autonomous routine, fairly simple and it worked in our test rack at home. Maybe not 100% but it would get to the rack at least. It was a hit-or-miss chance it would hook the tube. Go to the green light, if no green light stop. (Okay, a little more complicated than that, but we have to keep some secrets ;) )
After watching our robot in Manchester go loopy and run to the side wall, as well as a couple of other robots doing similar antics, I have a possible answer. The camera picks up stray green lights, particularly reflections off the side walls. As it gets closer to the rack, the (real) green light gets smaller and the ghost light becomes the bigger target.
We finally worked on Kevin's code and a couple of other adjustments and here is the result (thanks to our finals alliance team mates at 172 "Think Pink" for uploading the video!):
http://www.youtube.com/watch?v=P-mKZxMb3bc&NR
(hope I typed that in right.)
We talked to the lead field crew, who says she tested the field for just such a problem and did not find the ghost lights. To be fair I don't remember the results in the practice room, which she suggested. We also had the default color -- no adjustments there.
There may be other issues -- the green timer bar on the big screen, for example -- that neither one of us thought of, and we will continue to track it down. But when I saw other robots doing the same thing, veering off to the sidewalls just like our robot, I feel it is a problem, something we obviously have to overcome.
I don't know why more robots weren't doing autonomous. When I saw the Kickoff I thought it would be difficult, but it ended up being simple. In a way, I'm thankful there aren't more robots doing autonomous -- it would make the field awfully crowded! And yes, those two points were well worth the many sleepless nights!
ChrisMcK2186
06-03-2007, 08:23
The autonomous mode for our bot involved dropping a lance like arm and driving full speed forward untill we hit something. We changed it to half speed after while.
Chris
Snappel328
07-03-2007, 20:58
Scoring in Autonomous?
Pff no problem at all ;)
Uberbots
07-03-2007, 22:05
we just did another test tonight, and got an 18/23 consistency. part of the problem is that our defense robot smashed into the camera frame, therefore rendering our positioning system useless...
we hope to bring it up to 24/24 consistency by the weekend (=
meatmanek
07-03-2007, 22:32
Autonomous has always been the last thing we do (teleoperated mode comes first - autonomous is 15 seconds, teleop is 2 minutes, and most scoring happens in teleop)
We're only halfway through writing our autonomous, but we're not sure it'll help all that much anyway. This may change, but at St. Louis, hardly any spoilers were used ever.
That means the keepers are worth just about as much as a ringer. There are probably optimizations I can (and will, as soon as I discover them) make to my sweet state-machine based teleoperated mode that would help us score one more ringer in a match.
Now, we'll be trying to get autonomous working whenever we have free time at Purdue (and it's a good possibility that we actually get it working), but it's not our first priority.
Well, after a day of trying, we haven't done anything in autonomous.
Because someone accidently turned the camera off when connecting it.
Maybe things will work today?
One word... Low-Frequency SDAR :D
After this weekend should have the occasional anomalies out of it... but it does work... just a matter of getting the occasional screw ups out of it and not get spooked!
-q
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.