 | | | | | |
I want to make a robot that finds a small object, picks it up, and carries it
somewhere. I don't want to have to tell the robot how to find it (eg with a
line on the floor), and I'd prefer to avoid randomly (or even systematically)
searching the entire floor hoping to just run into the target.
The builtin light sensor seems to be useless at a distance. Even in a darkened
room with the target illuminated by a laser pointer the sensor values barely
fluctuate when it is pointing at the target. I've tried covering the target in
aluminum foil... I'm willing to put a beacon (like a battery powered LED) on
each target (I'm using empty film canisters) but even then I can't seem to make
a dent in the sensor readings. Should I buy an IR LED?
Has anyone tried putting a lens in front of the sensor? SONAR? RADAR?
--Ben
| | | | | | | | | | | | | | | | |
| |
|
Ben Jackson <ben@ben.com> wrote:
> I want to make a robot that finds a small object, picks it up, and carries • it
> somewhere. I don't want to have to tell the robot how to find it (eg with • a
> line on the floor), and I'd prefer to avoid randomly (or even • systematically)
> searching the entire floor hoping to just run into the target.
>
> The builtin light sensor seems to be useless at a distance. Even in a • darkened
> room with the target illuminated by a laser pointer the sensor values • barely
> fluctuate when it is pointing at the target. I've tried covering the • target in
> aluminum foil... I'm willing to put a beacon (like a battery powered LED) • on
> each target (I'm using empty film canisters) but even then I can't seem to • make
> a dent in the sensor readings. Should I buy an IR LED?
>
> Has anyone tried putting a lens in front of the sensor? SONAR? RADAR?
This is a thing I'm interested in too. Those of you who were at the mindfest
might remember the soccer playing robots. The bots were really simple but
performed impressively. The main trick was that the ball was actually
emitting IR (thru' LEDS) and the bots used standard light sensors to
localize it.
Does any of you have suggestion on what kind of IR led is more appropriate
to make such an object? Do I need any additional circuit, or just a LED and
a battery (and the proper resistor) are enough?
Thanks
Mario
Web page: http://www.geocities.com/~marioferrari
LUGNET member page: http://www.lugnet.com/people/members/?m=22
Proud member of ItLUG: http://www.itlug.org
| | | | | | | | | | | | | | | | | | | | | |
| |
|
Mario Ferrari wrote:
> Does any of you have suggestion on what kind of IR led is more appropriate
> to make such an object? Do I need any additional circuit, or just a LED and
> a battery (and the proper resistor) are enough?
I imagine that - even if using such an IR LED as a beacon - roomlight etc. will
still be the major problem.
So, I'm pretty sure that a quite simple means at the detector side would
simplify such a project dramatically: an optical IR filter. This would greatly
enhance the original S/N coming from the sensor. The filter could either be a
bandpass type or a highpass type ("high" refering to wavelength, not to
frequency).
I have no concrete source at hands at the moment, but I'm sure such filters
exist. Probably, a bandpass filter is expensive, but I think a simple highpass
would be good enough. In the best case, it might even not be necessary to buy a
commercial filter, since probably many materials exhibit the required spectral
transmission characteristics. (For instance, a piece of silicon wafer will
transmit IR but block any visible light. But, maybe wafer pieces are not easy to
obtain, too :-) We have to look around for something suitable that is easy to
get and cheap. If you decide to follow this way, I can offer to help locating
such materials.
[I just started playing with RIS, and so I have no experience yet whatsoever
with making own sensors etc. However, my profession is related with optics, so I
know what I'm talking about (well, roughly:-)).]
Further, refering to Ben's other question, yes, I think a kind of lens will be
needed. I think even a very simple lens will be clearly much better than no lens
at all. So, although the lens may not be perfect and may be hard to adjust well
(since the signal is not visible by eye), it will help much. Without any such
kind of "directional filter", things will be hard. (An alternative to a lens
might be a simple arrangement of two aligned pinholes, but that wastes a lot of
signal.)
Uwe
| | | | | | | | | | | | | | | | | | | | | | | |
| |
|
In lugnet.robotics, Uwe Denzer writes:
> Mario Ferrari wrote:
>
> > Does any of you have suggestion on what kind of IR led is more appropriate
Well, I do now! I bought three last night. I got a GaAs IR LED that was
clearly made for a remote control application. The good news is that when it
is pointed at the LEGO light sensor it pegs at 100%! Unfortunately it's higly
directional (it seems to be an LED behind a lens, rather than embedded in a
diffusing material). I chose it because it claimed to be a 3v part, which made
battery selection easier. I'm going to exchange it for one of the 5-6v LEDs
with a standard diffuse plastic body.
To power it I got 3v lithium watch batteries (biggest I could find, about the
size of a quarter). If I go with 6v LEDs I'll need two of them sandwiched
together. I plan to use 1/2" heat-shrink tubing to hold the batteries together
with electrodes (aka wires soldered to a loop of large gauge copper wire).
> > to make such an object? Do I need any additional circuit, or just a LED and
> > a battery (and the proper resistor) are enough?
They make some LEDs with builtin resistors now. You should make sure you know
what kind you've got.
> I imagine that - even if using such an IR LED as a beacon - roomlight etc. • will
> still be the major problem.
The LEGO sensor is far more sensitive to IR already, even without a filter.
That's why I was choosing it (although being invisibile to humans is a nice
side-effect).
> So, I'm pretty sure that a quite simple means at the detector side would
> simplify such a project dramatically: an optical IR filter.
You can get these (as well as so-called Hot Mirror filters to block IR) at any
camera store.
--Ben
| | | | | | | | | | | | | | | | | | | | | | | |
Ben Jackson wrote:
>
> You can get these (as well as so-called Hot Mirror filters to block IR) at any
> camera store.
>
> --Ben
Well, I knew that camera stores sell IR blocking filters but IR transmission
filters? Are you sure? (What would be the purpose for normal photography?)
Uwe
| | | | | | | | | | | | | | | | | | | | | | | | | | |
In lugnet.robotics, Uwe Denzer writes:
> Ben Jackson wrote
> > You can get these (as well as so-called Hot Mirror filters to block IR)
> > at any camera store.
> Well, I knew that camera stores sell IR blocking filters but IR transmission
> filters? Are you sure? (What would be the purpose for normal photography?)
IR Photography, of course. Lookup Kodak HIE, for example. The filters you are
looking for are Wratten #87 and 87C. If you look at them they appear to be
black because they pass no visible light. I believe the difference between the
two is exactly which IR wavelengths they pass, but I don't remember the
technical details. I'm sure they're on the Kodak website.
--Ben
| | | | | | | | | | | | | | | | | | | | | | | | | | | | |
Ben Jackson wrote:
> IR Photography, of course. Lookup Kodak HIE, for example. The filters you are
> looking for are Wratten #87 and 87C. If you look at them they appear to be
> black because they pass no visible light. I believe the difference between the
> two is exactly which IR wavelengths they pass, but I don't remember the
> technical details. I'm sure they're on the Kodak website.
>
> --Ben
OK, thanks for the info. That's interesting. I'll look for it.
Uwe
| | | | | | | | | | | | | | | | | | | | | | | | | |
| |
|
Thanks a lot Ben!
I'm going to look for some of those IR LEDs and make some experiments.
Mario
Web page: http://www.geocities.com/~marioferrari
LUGNET member page: http://www.lugnet.com/people/members/?m=22
Proud member of ItLUG: http://www.itlug.org
Ben Jackson <ben@ben.com> wrote:
> In lugnet.robotics, Uwe Denzer writes:
> > Mario Ferrari wrote:
> >
> > > Does any of you have suggestion on what kind of IR led is more • appropriate
>
> Well, I do now! I bought three last night. I got a GaAs IR LED that was
> clearly made for a remote control application. The good news is that when • it
> is pointed at the LEGO light sensor it pegs at 100%! Unfortunately it's • higly
> directional (it seems to be an LED behind a lens, rather than embedded in • a
> diffusing material). I chose it because it claimed to be a 3v part, which • made
> battery selection easier. I'm going to exchange it for one of the 5-6v • LEDs
> with a standard diffuse plastic body.
>
> To power it I got 3v lithium watch batteries (biggest I could find, about • the
> size of a quarter). If I go with 6v LEDs I'll need two of them sandwiched
> together. I plan to use 1/2" heat-shrink tubing to hold the batteries • together
> with electrodes (aka wires soldered to a loop of large gauge copper wire).
>
> > > to make such an object? Do I need any additional circuit, or just a LED • and
> > > a battery (and the proper resistor) are enough?
>
> They make some LEDs with builtin resistors now. You should make sure you • know
> what kind you've got.
>
> > I imagine that - even if using such an IR LED as a beacon - roomlight • etc.
> will
> > still be the major problem.
>
> The LEGO sensor is far more sensitive to IR already, even without a • filter.
> That's why I was choosing it (although being invisibile to humans is a • nice
> side-effect).
>
> > So, I'm pretty sure that a quite simple means at the detector side would
> > simplify such a project dramatically: an optical IR filter.
>
> You can get these (as well as so-called Hot Mirror filters to block IR) at • any
> camera store.
>
> --Ben
| | | | | | | | | | | | | | | | | | | | | | |
Subject:
|
RCX Web Server
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Mon, 3 Jan 2000 12:24:19 GMT
|
Original-From:
|
David Pfeffer <davidp@bytenikSTOPSPAMMERS.com>
|
Viewed:
|
1618 times
|
| |
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi all,
I've been working on a RCX web server, and it seems my project is
almost done. When I finish it, the website running off of the RCX
will be available at http://rcxonline.gbdev.org. Its not up yet, but
it will be within the next few weeks. I wrote it in LegOS (which I
can't get to work anymore... my harddrive crashed and now it won't
install (windows)... thats whats holding up my project... I hook the
IR unit and a modem up to a Basic Stamp IIx from ParalaxInc.com, to
retransmit the data in an IR unit friendly fashion. The RCX has a 5
to 4 compression routine I wrote and you can expect images on the RCX
website because of it (not linked from a real server, but actually
served directly from the RCX.) I've never been much of a hardware
guy, but if someone could design an interface to one of the sensor
and motor ports to read an EEProm, I could hold more data on it. If
my invention is sucessful, I might consider registering
www.rcx_on_the_web.com as a place for my RCX to have a perminant
location. (Just better tell the cleaning lady not to move the RCX
away from the IR beacon...)
David
-----BEGIN PGP SIGNATURE-----
Version: PGPfreeware 6.5.2 for non-commercial use <http://www.pgp.com>
iQA/AwUBOHCU8mU1FANHPcR9EQL7PQCeILf1PTbPR9i1rd/tsEFhZKz/iEoAoJbm
a7gLhKHntrFwzmr0LikOTcD4
=d89I
-----END PGP SIGNATURE-----
| | | | | | | | | | | | | | | | | | |
| |
|
Ben Jackson <ben@ben.com> wrote in message news:FnJpCI.JGo@lugnet.com...
> I want to make a robot that finds a small object, picks it up, and carries • it
> somewhere. I don't want to have to tell the robot how to find it (eg with • a
> line on the floor), and I'd prefer to avoid randomly (or even • systematically)
> searching the entire floor hoping to just run into the target.
>
> The builtin light sensor seems to be useless at a distance. Even in a • darkened
> room with the target illuminated by a laser pointer the sensor values • barely
> fluctuate when it is pointing at the target. I've tried covering the • target in
> aluminum foil... I'm willing to put a beacon (like a battery powered LED) • on
> each target (I'm using empty film canisters) but even then I can't seem to • make
> a dent in the sensor readings. Should I buy an IR LED?
>
> Has anyone tried putting a lens in front of the sensor? SONAR? RADAR?
Consider blocking the light emitter on the sensor, either completely or
partially
(the side of the emitter facing the sensor) to reduce the saturation of the
sensor
by the emitter, then use the IR sender on the RCX to emit pulses of IR and
follow
the high values from the sensor, they'll be from things reflecting the IR.
If you have
the luxury of putting an emitter on the objects themselves, even better.
Or, perhaps,
put a laser pointer on the robot, pointing out over the sensor.
Sounds like fun!
-Peter
--
Peter A. Vogel
Release Engineering Manager
iReady Corporation
http://www.iready.com http://www.iready.net
| | | | | | | | | | | | | | | | |
In lugnet.robotics, Peter Vogel writes:
> Consider blocking the light emitter on the sensor, either completely or
> partially
> (the side of the emitter facing the sensor) to reduce the saturation of the
> sensor
> by the emitter,
I've read this tip several times. I experimented with it last night, and the
difference between blocking and not blocking the sensor was minimal. My light
sensor must not bleed much light directly across. Perhaps I'm doing something
wrong?
Re: directionality, I've found that a black 1x2 beam works well to localize the
sensor, and if you put a black connector peg in there it can look right past
bright light sources and see dark objects. This would probably work well in
conjunction with a laser pointer.
--Ben
| | | | | | |
