Subject:
|
Re[2]: IR concept
|
Newsgroups:
|
lugnet.robotics.handyboard
|
Date:
|
Tue, 1 Apr 1997 08:17:41 GMT
|
Original-From:
|
Richard Whitehead <Richard.Whitehead@ukos^stopspam^.varian.com>
|
Viewed:
|
1693 times
|
| |
|
|
Chuck,
I love the wire whisker! Have you got any more tricks and tips I
could steal???
Richard
______________________________ Reply Separator
_________________________________
Subject: Re: IR concept
Author: cmcmanis@freegate.net at Internet-Mail
Date: 29/03/97 00:58
Adam wrote:
> An alternate IR system....
[Digitize the output from an IR LED with and without
a signal present to achieve some common mode noise
rejection.]
This type of system is susceptible to being 'blinded' by
high intensity IR sources (try it outside :-) I used a
similar technique on a Sumo robot to detect the white
rim of the Arena, only the real contest was held under a
skylight and I discovered this weakness the hard way :-)
> I've been scanning through some of the IR stuff to see if
> I can't pick out a good solid system for collision avoidance
> on a short range (approx. 6-10cm) scale, but everything seems
> like alot of trouble.
If you're maximum detection range is 10 cm, (< 4 inches) then I
strongly recommend that you consider a high compliance whisker.
Put a piece of #80 piano wire (spring steel wire) with a ball
on the end (a glob of hot glue works well) and mount it in a
deflection detection chamber (ie a piece of brass tubing.)
Write if you want manufacturing details (the hot glue gun is
your friend :-)
> I'm curious to know what others have to say about this approach.
As in all sensor designs it has its strong points and its weak
points. Easily implemented in hardware, susceptible to some common
interference sources.
Another common way to implement this sensor is to hook the LED
up to a modulating source (such as a square wave from a 555),
then feed that input, and the output of a detector into an
opamp. Since one input is inverting, and the other non-inverting,
if you have the same signal on both pins (the detector is following
the transmitter) the opamp will produce zero output, when the
signal on the detector fades, the output of the opamp will increase.
You can rectify the output and put it through a level detector to
get a binary output. This sensor gets even more interesting if you
put a variable feedback stage into the opamp so that the gain of
the detector input is programmatically controlled. If you can
recover the signal by increasing the gain then the input to the
gain control gives you a rough approximation of the distance to the
target (remember that light intensity is 1/d^2 so the output curve is
exponential!) At that point you've pretty much built one of the
Hammatsu sensors by hand. You can increase selectivity by putting
a bandpass filter on the output of the IR detector.
The bottom line on sensors is that if you can get three different
kinds of sensors that can all detect the same thing, then you
have a very reliable way of detecting it. So the more the merrier.
--Chuck
--
cmcmanis@netcom.com http://www.professionals.com/~cmcmanis
All opinions in the non-included text above are the sole opinions of
the author.
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
