Post URL to your autonomous mode

[Doug Leppard]

Originally Posted by 1629coalition

This is 1629's auto-mode on youtube
http://www.youtube.com/watch?v=nC3bwTTS2lQ

Good video, but I can not download youtube, at least I don't know how. Also be sure to tell me the method (IE camera).

[Kristian Calhoun]

Here is a autonomous compilation video that I recently made, showcasing our successful caps. In order, from the video's sequence, the matches are:

NJ Match 39 (This keeper actually didn't count, as it was supported by the stinger.
NJ Match 11
UTC Match 64 (One bot from each alliance capping, us on blue, 1124 on red.)
Galileo Match 22 (Nice close up shot.)
Galileo Match 82
Galileo Match 90* (1126 positions correctly, but knocks off their keeper as they back away. We cap ours on the same alliance.)
UTC Match 52**
NJ Finals 1*
NJ Finals 2***

Match 52 from Galileo, Match 52, and the NJ Finals are my favorites. (I'm still amazed at how we scored in match 52 after slamming into the rack and pulling up from beneath the spider.) As for our method, we used a single ultrasonic sensor to determine the proper distance from the base of the rack. During the matches at NJ, we used the CMUcam to align left/right towards the light, if it were in the view of the camera, and if not it would proceed to use the sonic sensor code. At UTC, we experienced camera troubles, and decided to just use the sonic sensors, and try our best to determine/guess where the spider leg would be after the rack had been moved. At championships, we solely used this method, and scored during 3 out of our 7 qualification matches.

The asterisks indicate my preference on which are the most entertaining to be used for the video, three stars being the highest.

[Doug Leppard]

Originally Posted by Kristian Calhoun

Here is a autonomous compilation video that I recently made, showcasing our successful caps. In order, from the video's sequence, the matches are:

NJ Match 39 (This keeper actually didn't count, as it was supported by the stinger.
NJ Match 11
UTC Match 64 (One bot from each alliance capping, us on blue, 1124 on red.)
Galileo Match 22 (Nice close up shot.)
Galileo Match 82
Galileo Match 90* (1126 positions correctly, but knocks off their keeper as they back away. We cap ours on the same alliance.)
UTC Match 52**
NJ Finals 1*
NJ Finals 2***

SOAP is currently down, which has the video of UTC Match 64, and the NJ Final Rounds. If it is not back up soon, I could get you the videos by uploading them to a file sharing site or possibly our team website (AIM file transfers also work). Match 52 from Galileo, Match 52, and the NJ Finals are my favorites. (I'm still amazed at how we scored in match 52 after slamming into the rack and pulling up from beneath the spider.) As for our method, we used a single ultrasonic sensor to determine the proper distance from the base of the rack. During the matches at NJ, we used the CMUcam to align left/right towards the light, if it were in the view of the camera, and if not it would proceed to use the sonic sensor code. At UTC, we experienced camera troubles, and decided to just use the sonic sensors, and try our best to determine/guess where the spider leg would be after the rack had been moved. At championships, we solely used this method, and scored during 3 out of our 7 qualification matches.

The asterisks indicate my preference on which are the most entertaining to be used for the video, three stars being the highest.

Great compilation video. Tell me when the others are ready.

[Guy Davidson]

Sorry to bump up an old thread, but I think this is the best place to ask this question. All of you who had butt-kicking autonomous modes last years, what did you use? Which sensors, custom circuits, and special logic? What pointers would you have that generally lead to a quality autonomous mode?

Thanks

[Doug Leppard]

#1 best secret. Hardware done by end of week five, that gave us a week to practice driving and auto mode. Auto mode just takes time. So dependable hardware and done with enough time to spend lots of time perfecting it.

#2 we had a 3/4 size field to practice auto mode.

#3 we used easyproC for fast developing and more people could do it.

#4 we proto typed lots of sensors using vex to when we go the hardware we didn't start from scratch but knew what we could do and not do with sensors and had routines working.

For 2007 we used the camera. 2006 we used gyro mainly. Previous years we use IR sensors and wheel encoders. Really depends on the contest. We have also tested sonic sensors.

Suggest practice using vex with all the sensors to get a feel for them.

[Guy Davidson]

Very interesting. Thanks Doug. Obviously we'd love to (and hopefully will) have the hardware done by the end of week five. We're planning to do our own thing with a coprocessor this year, so we're starting it early to give us time to work out kinks. We'll definitely try to get some time on a practice field and test as many sensors as we think we might use, so we can use them if we need to.

[Doug Leppard]

Originally Posted by sumadin

Very interesting. Thanks Doug. Obviously we'd love to (and hopefully will) have the hardware done by the end of week five. We're planning to do our own thing with a coprocessor this year, so we're starting it early to give us time to work out kinks. We'll definitely try to get some time on a practice field and test as many sensors as we think we might use, so we can use them if we need to.

In 2006 the hardware guys saw how important the auto mode was. So they decided that it was not an option to finish by week 5. They kept the pressure on it and cut out things to make it done by week 5.

I have been in seasons that it wasn’t even finished at the 1st competition and took all of the practice day, then we had times we finished just before time to put it in the crate, then last year finishing at end of week 5. I tell you the team was so much more ready for competition by finishing by week 5.

[Guy Davidson]

Originally Posted by Doug Leppard

In 2006 the hardware guys saw how important the auto mode was. So they decided that it was not an option to finish by week 5. They kept the pressure on it and cut out things to make it done by week 5.

I have been in seasons that it wasn’t even finished at the 1st competition and took all of the practice day, then we had times we finished just before time to put it in the crate, then last year finishing at end of week 5. I tell you the team was so much more ready for competition by finishing by week 5.

Thanks Doug. As a programmer, I hope they keep making autonomous more interesting and important, as that's the most interesting part of the competition for us (that, and designing good control algorithms for our drivers). I guess I'll have to try and ask our build guys to do their best to finish (heh, as if they don't already do), so we'll get as much robot time as we can.

[Doug Leppard]

Originally Posted by sumadin

Thanks Doug. As a programmer, I hope they keep making autonomous more interesting and important, as that's the most interesting part of the competition for us (that, and designing good control algorithms for our drivers). I guess I'll have to try and ask our build guys to do their best to finish (heh, as if they don't already do), so we'll get as much robot time as we can.

Can you take the 2007 robot and do some of the programming? Assuming you have the same base. Try to do as much as possible before the week 6. Our programmers worked the whole time, but especially week 6.

[Guy Davidson]

Originally Posted by Doug Leppard

Can you take the 2007 robot and do some of the programming? Assuming you have the same base. Try to do as much as possible before the week 6. Our programmers worked the whole time, but especially week 6.

We can. Better yet, we're working on a practice drive train right now, that will hopefully closely resemble the competition one. It will have an entire IFI control system with all the electronics, and we will be using it to program before the real one is done, and probably do some work on our co-processor solution before the season starts.

[Doug Leppard]

Originally Posted by sumadin

....probably do some work on our co-processor solution before the season starts.

What are you using the co-processor for?

[Guy Davidson]

Originally Posted by Doug Leppard

What are you using the co-processor for?

We're looking into using it as a learning experience and something big we (myself and a few others) can do in my (and some of theirs') senior year to learn and leave for the team to use. We have to first get the system up and running, probably building on what the AdamBots did and using a gumstix. Eventually, our goal is for the coprocessor to read and handle sensor inputs, as well as receieve the user inputs from the OI (via the RC), and do all computation and logic. We might end up leaving some sensors connected to the RC if it ends up being easier. We think it will be a great learning experience, but for game use, it will let us run as many sensors as we want, do high-level logic with them, and as much floating point and trig as we need to without worrying about the RC. If you have any comments, ideas, or suggestions, we'll be more than happy to hear them, as the project is still in its very early stages right now.

[Doug Leppard]

Originally Posted by Kristian Calhoun

Here is a autonomous compilation video that I recently made, showcasing our successful caps. In order, from the video's sequence, the matches are:

NJ Match 39 (This keeper actually didn't count, as it was supported by the stinger.
NJ Match 11
UTC Match 64 (One bot from each alliance capping, us on blue, 1124 on red.)
Galileo Match 22 (Nice close up shot.)
Galileo Match 82
Galileo Match 90* (1126 positions correctly, but knocks off their keeper as they back away. We cap ours on the same alliance.)
UTC Match 52**
NJ Finals 1*
NJ Finals 2***

Match 52 from Galileo, Match 52, and the NJ Finals are my favorites. (I'm still amazed at how we scored in match 52 after slamming into the rack and pulling up from beneath the spider.) As for our method, we used a single ultrasonic sensor to determine the proper distance from the base of the rack. During the matches at NJ, we used the CMUcam to align left/right towards the light, if it were in the view of the camera, and if not it would proceed to use the sonic sensor code. At UTC, we experienced camera troubles, and decided to just use the sonic sensors, and try our best to determine/guess where the spider leg would be after the rack had been moved. At championships, we solely used this method, and scored during 3 out of our 7 qualification matches.

The asterisks indicate my preference on which are the most entertaining to be used for the video, three stars being the highest.

Got it, great job.

[PhilBot]

Originally Posted by Doug Leppard

Good video, but I can not download youtube, at least I don't know how. Also be sure to tell me the method (IE camera).

Here is a direct link to Team 1629's Autonomous video (wmv) file.

http://www.gearsinc.org/images/auto_1629.wmv

These are from the elimination matches at Chesapeake.

Navigation is performed by camera and gyro.
The camera is hard-mounted to the robot (no pan/tilt).
It uses the Gyro to get withing viewing range, and then locks in with the camera. Both heading and range are provided by the camera.
The vertical car and rotational arm (two axes of motion) are positioned based on feedback from potentiometers.

Once we got the camera so it could see the lights, we mised one, and scored two.

In addition to a direct near-side score, we are working on scoring on the opposite side of the rack (for off season games).