Vision Tetra Alignment Issues

[phrontist]

I have been unable to find anything in the rules that states that the vision tetras will be aligned with the feild in a predefined matter, so we can assume that they could be facing any direction, right? If this is the case, does anyone have any ideas how to align a grabbing mechanism that is dependent on being at a certain angle when grabbing? I was thinking of an algorithm that uses the size of the green blob to determine what angle the tet is relative to the camera...

Too bad it couldn't have been nice and simple, with the bottem bar of the triangular face being parallel with the feild.

[Lil' Lavery]

Simpler solution, dont have it dependant on the orientation.

[phrontist]

Originally Posted by Lil' Lavery

Simpler solution, dont have it dependant on the orientation.

Well, there are various trade offs if you do that. For instance, it seems to me that most gripper designs that allow for any angle of attack also require the arm to lift higher. For this thread, we'll assume that we want to handle this problem in code, as oppose to alteration of gripper design.

[jgannon]

All of the illustrations I've seen have depicted the bar closest to your driver station being parallel to the driver station wall. However, I would ask the FIRST Q&A to verify this.

[phrontist]

Originally Posted by jgannon

All of the illustrations I've seen have depicted the bar closest to your driver station being parallel to the driver station wall. However, I would ask the FIRST Q&A to verify this.

Oh, that would be too bloody fantastic!

How do you do such a Q&A?

[jgannon]

Originally Posted by phrontist

Oh, that would be too bloody fantastic!

Well, let us all know when you find out. I'd like to not have to worry about that in code.

FIRST Q&A:
http://www.usfirst.org/robotics/2005/qa.htm

[Caleb Fulton]

You know the tilt angle of the camera and you know the (x, y) coordinates of the centroid of the blob, so you can make a useful estimate of the distance given the proper transformation which, for simple (FIRST robot) purposes, could be as simple as calling the "distance" to the target the difference between some reference point's y-coordinate and the y-coordinate of the centroid while always keeping the x-coordinate of the centroid at the center of the screen by steering the robot. Of course, this is not an accurate measurement in "feet" (nor is it even linear), but you will get a qualitative estimate that can surely suit your purposes.

[RonStoppable102]

Originally Posted by phrontist

Too bad it couldn't have been nice and simple, with the bottem bar of the triangular face being parallel with the feild.

As stated, I do recall seeing them place the vision tetras like that during the kickoff, I would check in the manual, but its 2:24am.... I'm also guessing that it can be set to track different colors (i'm guessing it interprets color as a function of frequency, but I haven't had a chance to mess with it), we have a kid everyone hates that always wears a red hoodie, for example. Any ideas?

[phrontist]

Originally Posted by RonStoppable102

As stated, I do recall seeing them place the vision tetras like that during the kickoff, I would check in the manual, but its 2:24am.... I'm also guessing that it can be set to track different colors (i'm guessing it interprets color as a function of frequency, but I haven't had a chance to mess with it), we have a kid everyone hates that always wears a red hoodie, for example. Any ideas?

Yeah, here is an idea. Tranform that guy everyone hates into a functioning member of soceity through peer pressure and FIRST. Hurting people with robots doesn't do much for the image of first, espesically with the crappy media we have in the US...

Tonight 60 minutes investigates the tragic death of Tommy Redhood, who was killed by a group of students in a nationwide robotics gang. This gang is a looming threat, and determined recruiters may bring it to YOUR CHILDS SCHOOL. *dramatic reverb*