View Full Version : Team 135 Autonomous
Team 135
19-02-2008, 20:49
Team 135 has been working for the past 4 weeks on our autonomous. It has come a long way to where it is now. It is using cutting edge technology to find the ball and to tell our bot where to get it. We are not sure whether we will need to even need a robo-coach. You should check it out for your self, you will be amazed, lol ;)
We are going to
St. Louis
Purdue
National :D
arpitshah
20-02-2008, 02:27
wowwww...that is sooooooooo fascinating...can u guys upload some videos or something so that we can have a look at it.....m dying to see that amezing machine....and the technology
CyberWolf_22
20-02-2008, 03:16
wowwww...that is sooooooooo fascinating...can u guys upload some videos or something so that we can have a look at it.....m dying to see that amezing machine....and the technology
I agree I think a lot of team would love to see how you are doing this. At least tell everyone what sensors you are using to do this. This sounds like you have put a lot of work in to it and I hope it is very successful during the game.
Team 135
22-02-2008, 16:25
We are using a Logitec USB webcam, interfaced into a gumstix computer. The gumstix computer is interfaced with all of our encoders and IR sensors via eithernet. The webcam is able to move vertically to see where the ball is. So as we are driving up to the rack it will move up to see the ball. The gumstix computer is running Python linux, with C++ and spin programing running on it. The bot is able to distiguish between red and blue as pre programmed before the match. We are also able to tell it where the starting position is. All of this is done with an onboard screen. The other great thing is we can see the output of the camera directly on a screen on the bot. The on board computer records video of the match so we can replay the entire match when it is over.
how are you powering the gumstix computer??? and are you allowed onboard screens??? sounds really cool though would love to see video!
Dan Petrovic
22-02-2008, 18:41
Innovation in Control, anyone?
That sounds awesome! I wish I could see it in person.
Team 135
22-02-2008, 21:40
I am sorry to say that I will not be able to post video now because we did not take any. We will try to take some at St. Louis. The gunstix is powered from the main robot battery. It is also tied into the backup battery to allow it to continue running during minor power fluxuations from the main battery. We also built in a charging circuit for the back up battery.
As I posted before if you would like to see it...see it at one of the competitions that we will be at below
St. Louis
Purdue
Atlanta
Team 135
22-02-2008, 21:43
The main way that we are finding the balls are through blob tracking, color filters, circle recognition, and line recognition.
Uberbots
22-02-2008, 22:30
thats REALLY cool!
is it 100% accurate?
Doug Leppard
23-02-2008, 10:06
How fast is it? Do you hit the home ball and come around and hit the other side ball?
Team 135
23-02-2008, 13:14
Testing has been fairly limited due to the fact that we do not have a full field. We hope to test it fully at St. Louis and get it fully perfected. It should be able to get both balls. Is anyone planning to be at St. Louis?
we found that the CMU cam worked better at tracking the balls than it did the green light.:cool:
Team 135
23-02-2008, 19:56
We did some research on cameras and found that the CMU cam was not configurable enough. We found that using a standard Logitec web cam, we could configure it to work very well. We did not try to use the CMU cam this year so I am not sure how well it might have worked.
Richard Wallace
23-02-2008, 20:39
This sounds really cool and I look forward to seeing it in action.... The gunstix is powered from the main robot battery. It is also tied into the backup battery to allow it to continue running during minor power fluxuations from the main battery. ...(emphasis added)
A direct connection of your custom circuit to the RC backup battery would violate <R73>. This was clarified by an official Q&A response by the GDC here (http://forums.usfirst.org/showthread.php?t=8700).
Per <R55> your custom circuit must be powered either via a 20A circuit from your breaker panel, or by one (or more) of the +5V pin(s) on the Robot Controller. I assume that you meant to say that the Robot Controller's backup battery will allow your custom circuit to continue running during minor power fluctuations.
That sounds fantastic. Please post when you have video of it all running... or, at any rate, I'll see you guys at purdue (as usual, you all were great in '06!)
How long does the cycle take?
How many points can you make in 15sec? (i.e. do you also drive over lines or anything)
Ditto on CMU cam working great for finding trackballs. We didn't try it, but 1501's ability to find our teams shirts... eh... I mean a blue ball... sure convinced me!
Sounds like a great project, please post a video.
-q
Uberbots
23-02-2008, 23:43
We did some research on cameras and found that the CMU cam was not configurable enough. We found that using a standard Logitec web cam, we could configure it to work very well. We did not try to use the CMU cam this year so I am not sure how well it might have worked.
The cmucam3 is... ive made some pretty fancy image recognition algorithms on it
Team 135
24-02-2008, 09:14
Let me correct what I wrote before...The back up battery is tied into the custom board in order to charge it.
The main part of our autonomous is finding the balls but we did put major thought into how to knock off the ball without having to stop. We have figured it out with minor modification to our end effector. (Since I am not at school today, I do not have any videos..but will upload some soon).
Glad to hear that we will see 1024 at Purdue :cool:
This sounds really cool and I look forward to seeing it in action.(emphasis added)
A direct connection of your custom circuit to the RC backup battery would violate <R73>. This was clarified by an official Q&A response by the GDC here (http://forums.usfirst.org/showthread.php?t=8700).
Per <R55> your custom circuit must be powered either via a 20A circuit from your breaker panel, or by one (or more) of the +5V pin(s) on the Robot Controller. I assume that you meant to say that the Robot Controller's backup battery will allow your custom circuit to continue running during minor power fluctuations.
Richard,
Team 358 posted a clarification to their question (http://forums.usfirst.org/showthread.php?t=8741), and the GDC allowed teams to use the +7.2V "PWM pins" for power.
Richard Wallace
24-02-2008, 11:47
Team 358 posted a clarification to their question (http://forums.usfirst.org/showthread.php?t=8741), and the GDC allowed teams to use the +7.2V "PWM pins" for power.Thanks, Manoel.
Kingsmen
25-02-2008, 09:18
This is a picture of the screen and on the left a little bit of the webcam.
http://i64.photobucket.com/albums/h192/Buddylee263/DSC01196.jpg
Looks nice, wish I had stuff like that to play with.
Wayne Doenges
25-02-2008, 20:04
IMHO the CMU cam is just as configurable.
We can find both balls (no matter where they are on the overpass) during autonoumous (we don't use Hybrid) and knock them down in 15 seconds.
We are using the CMU cam (to find the balls)and acoustic sensors (to find the walls).
Good luck Team 135. See you at BMR.
Doug Leppard
25-02-2008, 20:48
IMHO the CMU cam is just as configurable.
We can find both balls (no matter where they are on the overpass) during autonoumous (we don't use Hybrid) and knock them down in 15 seconds.
We are using the CMU cam (to find the balls)and acoustic sensors (to find the walls).
Good luck Team 135. See you at BMR.
That's great, hard thing to do. I thought of using a camera but didn't want to deal with it.
How long does it take to see the ball and track it?
We get the first ball reliably but had trouble with 2nd and tus decided to go for points by driving around.
Team 135
18-03-2008, 10:52
As asked for I have posted some videos. When you load the youtube page you will see that there are about 7 videos from the boilermaker regional. In the videos you will see what the robot sees after being processed by our software. It will put blue circle like things around what it thinks is the ball. It will the put little sun looking things in the middle of the blue circles. If the sun is green it has determined that what it is looking at is the ball. The yellow suns are what the robot has determined are the next best thing to go after.
these videos should answer the questions previously posted.
http://www.youtube.com/profile_videos?user=cageorge76
if you have any questions please post them here!!!
Anthony
Lead Controls/Wiring
Team 135
Doug Leppard
18-03-2008, 16:11
Anthony,
Very impressive. How did you do? How many times did you hit a ball and go around?
BTW I recognized crashes like ours. Our machine started to mysteriously start out and make a rapid turn to the left, other times file off a relay it didn’t have code in there to fire off.
Tom Bottiglieri
18-03-2008, 16:26
We just drive the robot with the remote. Always hit about 4 lines and the opposite side's ball in CT. Took about 5 minutes to write the code for, too.
Doug Leppard
18-03-2008, 16:39
We just drive the robot with the remote. Always hit about 4 lines and the opposite side's ball in CT. Took about 5 minutes to write the code for, too.
Do you have any videos?
Tom Bottiglieri
18-03-2008, 16:48
Do you have any videos?
Yes, but they are not online or accessible right now. TBA should have the CT matches archived soon, so I'll post in this thread with whatever comes first.
wt200999
18-03-2008, 17:50
Our scouts had you guys down as having amazing auton, I didn't realize you guys did that all with just the remote! Thats awesome! :ahh:
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.