RE: [SeattleRobotics] Re: Steampunk Mouse
- How about this reflective IR sensor for a foot sensor?
I've got GP2Y0A51SK0F Sharp IR sensors currently on my table top 'bots, but
I can't find a source for them!
Mine are at an angle, and I have a pair riding below the forward-looking IR
Can be seen in this pix.
> -----Original Message-----
> On Behalf Of Peter Balch
> Brilliant. An excellent answer to a real need. That will go on my To Do
> Did your foot sensors look straight down or forward?
- > If you're moving in a straight line, and the ball is at aBut the whole point of LOT is that the catcher doesn't move in a straight line.Fig 2Ball Catching: An Example of Psychologically-based Behavioural Animation
M F P Gillies and N A Dodgson
Cambridge University Computer Laboratory
February 3, 1999and the (shaggy) dog storybhattacharjee nyt 1-7-03 fly ball or frisbee.pdfhttp://digitalunion.osu.edu/r2/summer06/maynor/popular%20press/bhattacharjee%20nyt%201-7-03%20fly%20ball%20or%20frisbee.pdfI think one of the problems is they are working from 2d images from one camera (with no depth information so projection is no big deal) whereas animals work from 3d and our depth perception is very fine even at distances where they say we can't judge depth (they are probably doing sums again).DAvid----- Original Message -----From: Randy M. DumseSent: Friday, April 15, 2011 11:57 PMSubject: RE: [SeattleRobotics] dream bots ... [was: Re: Steampunk Mouse
David Buckley said: Friday, April 15, 2011 5:00 PM
> they certainly are not doing sums
I still hold they are using a special solution to a differential
equations, but can't offer one yet.
But let me explain my inspiration for suggesting such. Something
from my Navy days on the ships bridge as a watch officer. As
Officer of the Deck (second only to the Captain himself for
operational control of the ship) we were taught to use parallel
rulers, and a special chart paper, to determine how close
another ship would pass to us, called CPA (closest point of
approach) which is very complex due to relative motion of both
ships through the water. But even the lowliest seaman was taught
how to tell if we were on a colision course.
Constant bearing, decreasing range. You're gonna get hit.
It doesn't matter if two ships are heading toward each other, or
one toward and one away, or at any odd angle and any odd speed.
If you see constant bearing, decreasing range, courses and
speeds remain constant, collision is going to happen.
Here, with relative motion of ships, the problem is simpler than
the baseball, because the action is all in 2D, without the third
vertical dimension with gravity's acceleration component. Non
the less, all the complexities of the algebraic solution
disappear and the constant bearing, decreasing range, the result
is certain. I think the principle is the same for the 3D case.
If you're moving in a straight line, and the ball is at a
constant relative bearing, you're going to intercept it's line
of motion. This doesn't explain how you know if you'll be where
you can reach it, whether it will be above your head, or already
rolling on the ground, but your lines of motion will intersect
So there's part of the key. What line to choose so it is the
right height is the other part. How that is done, I don't know.
But once you know that line speed is the only variable, running
fast enough to keep the relative bearing static will get you