Subject:
|
Re: IR collision avoidance
|
Newsgroups:
|
lugnet.robotics.handyboard
|
Date:
|
Sun, 2 Mar 1997 08:26:36 GMT
|
Original-From:
|
Adam <adam@*avoidspam*ring.zenox.com>
|
Viewed:
|
1496 times
|
| |
|
|
Riachard,
Could you please explain to me how the IR emitter angled at
45 degrees is able to pick up its beem. I'm unsure of the
IR emmission dynamics. Wouldn't an IR angled perpidicular
to the wall being travelled along suffice as the only sensor?
Is the 45 degree angled IR pickup on the upcoming walls?
It sounds like a good system. I'd like to learn more on how
it works. Please explain the dynamics in more detail.
Thank you.
> > On Wed, 26 Feb 1997, EDWIN SALGADO wrote:
> >
> > >
> > >
> > > On Mon, 24 Feb 1997, Richard Vannoy wrote:
> > >
> > > > We have "forward" and 45 degrees right". We wall hug to the right while
> > > > you hug to the left. We ignore IR to left and use only the left contact
> > > > switch for collisions (rare).
> > > >
> > > > We had way too much power and noise with "together" mounting. We
> > > > finally mounted xmitter 2 inches above with a small copper plate to
> > > > sheild the transmitter from direct transmission to the receiver. And we
> > > > had to use sheilded cable for whatever left the circuit boards.
> > > >
> > > > Ours is similar but does not allow ahead straight. We do...
> > > >
> > > > front blocked right blocked action
> > > >
> > > > YES Don't care Turn left 90 degrees.
> > > >
> > > > NO YES Turn wheel 10 degrees left
> > > >
> > > > NO NO Turn wheel 5 degrees right
> > > >
> > > > This puts the robot in a long slow right arc till a wall is detected.
> > > > Then, untill a corner comes up the shift between 10 left and 5 right
> > > > keeps the bot a nice 18 inches from the wall. People who don't know
> > > > that the wheel is almost constantly "correcting" don't even notice it.
> > > > To them it looks like a straight line.
> > > >
> > > >
> > >
> > >
> > > ** My only question is how do you keep the bot at a constant 18 inches
> > > ** away from the wall with a collision detection algorithm? Is it
> > > ** possible to get analog outputs from a sensor indicating the
> > > ** distance from a wall? Or is there is better way to do this?
> > >
> > >
> > > > Good algorithm!! I didn't use a loop because the response seems great
> > > > when we react to any "blocked". So far I've seen no noticable
> > > > transient/false signals, but your way seems much more elegant.
> > >
> > >
> > >
> >
>
>
|
|
Message has 1 Reply:
 | | Re: IR collision avoidance
|
| (...) Someone could probably give some very good reasons why 45 degrees is a good choice. For me it was common sense and experiment. Less than 45 degrees from ahead and the angle the IR is hitting the wall is too "shallow" and I don't get a return. (...) (28 years ago, 5-Mar-97, to lugnet.robotics.handyboard)
|
3 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
