Autonomous: CMUCam or dead reckoning?

[EricH]

Originally Posted by Greg Marra

Wow, did you guys really manage to judge distance based on the size of what the camera is seeing? I'm impressed.

We did it by mounting the camera about three feet up and using trig to calculate the distance to the tetra based on the angle of depression to it.

As far as I know, it isn't the size. It's the position that counts. Size may have something to do with it, but I don't think so.

[miketwalker]

Originally Posted by Greg Marra

Wow, did you guys really manage to judge distance based on the size of what the camera is seeing? I'm impressed.

We did it by mounting the camera about three feet up and using trig to calculate the distance to the tetra based on the angle of depression to it.

If you think bout it... from any angle the tetra (or goal for that matter) is turned... the width is about the same, not enough off to make any significant difference. Thus, when lookin at it and having the camera straight on at the same angle every time... you can say "once the thing is this wide, stop and lift your arm". Both the trig way and this way work similar from what I can tell, I just figured A) without servos theres one less thing to mechanically go wrong and B) dont need to waste precious variable and programming space on trig functions.

[Greg Marra]

Originally Posted by miketwalker

If you think bout it... from any angle the tetra (or goal for that matter) is turned... the width is about the same, not enough off to make any significant difference. Thus, when lookin at it and having the camera straight on at the same angle every time... you can say "once the thing is this wide, stop and lift your arm". Both the trig way and this way work similar from what I can tell, I just figured A) without servos theres one less thing to mechanically go wrong and B) dont need to waste precious variable and programming space on trig functions.

No servos?

What do you do if the tetra isn't in the field of vision covered by the camera? Turn the robot to look for it? We found that our camera only has about a 55 degree field of vision, which isn't enough to cover all the tetra locations in any of the starting positions (although, to be honest, I never really thought of having the robot start turned so that the two leftmost tetras would go in the leftmost part of the field of view if you started in the leftmost starting position; but then again wouldn't your alliance partners get in the way a little bit?

[miketwalker]

Originally Posted by Greg Marra

What do you do if the tetra isn't in the field of vision covered by the camera? Turn the robot to look for it?

That's exactly what I'm doing. A simple timer saying to turn left for so many loops, then right, then back, and whenever you see something swap over to following it.