View Full Version : Video: Team 294 driving telescope with Kinect
Peter Johnson
16-10-2011, 19:24
We mapped the left and right arm controls to our 2011 telescope and wrist today. Left arm controls the telescope height, right arm controls the wrist angle. Definitely stopped work in the lab for the interactive demonstration! Obviously the gains could use some additional tweaking, but it was pretty fun to see working.
Video is on YouTube: http://www.youtube.com/watch?v=iFsYMiGbDss
We used RobotPy for 2011 competition, so the robot code is Python for this demonstration as well.
Andrew Lawrence
16-10-2011, 19:27
What's the orange banana costume for?
Peter Johnson
16-10-2011, 19:44
What's the orange banana costume for?
It's our team captain wearing our surfboard mascot. Per the i.am.first video (http://www.youtube.com/watch?v=vLMOKM4QKcE, about 7 minute mark), our mascot could be interpreted as a banana or sweet potato, but is actually a surfboard.
The Kinect control does work a bit better without the costume (as the costume makes the head look really big to the Kinect), but we thought it'd be fun for the video.
Peyton Yeung
16-10-2011, 19:45
Banana? I thought that was a surf board.
Andrew Lawrence
16-10-2011, 19:47
It's our team captain wearing our surfboard mascot. Per the i.am.first video (http://www.youtube.com/watch?v=vLMOKM4QKcE, about 7 minute mark), our mascot could be interpreted as a banana or sweet potato, but is actually a surfboard.
The Kinect control does work a bit better without the costume (as the costume makes the head look really big to the Kinect), but we thought it'd be fun for the video.
Sorry, my bad! I forgot you guys were 294, and your mascot was a surfboard. So when I watched it, my first thought was, "Why is the driver wearing and orange banana?"
Great video, and great maneuvering! As a driver, I could never control a robot like that.
Bwahaha that part at ~28secs makes me laugh. The height factoring in is going to play a role I feel. And looks like to keep with the game we're going to have to go ahead and buy a Kinects and start messing around with it.
Very Cool!!
Nice work people!
Tommy F.
16-10-2011, 21:43
And looks like to keep with the game we're going to have to go ahead and buy a Kinects and start messing around with it.
I don't know if you're aware or not, but every Kit of Parts for the 2012 season will have a Kinect. (http://www.youtube.com/watch?v=hlgA2nlXUF8)
True, but I don't believe six weeks is long enough to get, code, and mess around with one.
msimon785
16-10-2011, 22:12
Bwahaha that part at ~28secs makes me laugh. The height factoring in is going to play a role I feel. And looks like to keep with the game we're going to have to go ahead and buy a Kinects and start messing around with it.
Coming from a mechanical background and not a programming one, I wonder whether scaling the heights will solve the height issues. In other words, you can set it such that when the a button is pressed (this can be a physical button, such as a joystick button or something on the Classmate DI dashboard), the Kinect will calibrate the highest hight of the hand as the top of the mast. To be clear, create a calibration system that registers the height of the hands at any given time the button is pressed to be the highest point on the mast.
1515 did this, while not to interface with a Kinect, in the 2011 season. Each time our robot's mast hit the base-pan, thus triggering a vex limit switch, the position was registered as 0. I can put you in contact with our programmers (Java) if you're interested in more detailed information.
Nonetheless, it looks and works fantastic. I am really impressed by the consistency in the Kinect's interpretation of joint positioning. Great job.
True, but I don't believe six weeks is long enough to get, code, and mess around with one.
The mission of the Beta Test teams is to REPORT ALL FINDINGS and also hold public seminars for all teams. I know that is not the same as playing with a Kinect yourself, but I am sure that all of the teams will try to make it so that you can get something working very early in the build season.
Peter Johnson
16-10-2011, 23:01
Coming from a mechanical background and not a programming one, I wonder whether scaling the heights will solve the height issues. In other words, you can set it such that when the a button is pressed (this can be a physical button, such as a joystick button or something on the Classmate DI dashboard), the Kinect will calibrate the highest hight of the hand as the top of the mast. To be clear, create a calibration system that registers the height of the hands at any given time the button is pressed to be the highest point on the mast.
Actually the issue in this case was the costume; the "head" is interpreted as the top of the surfboard, so the arms seem pretty short in comparison. It scales properly (and generally works much better) without the costume, but the costume made for a much better video :cool:
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.