How does the Robomower perimeter wire work?
Hans Jakobsen
the area, but they also claim that the Robomower will stop working if it is
taken outside the area. This sounds like the perfect safety system for any
autonomous robot and I would like to use it in the robotic project I am
working with.
How does it sense whether it is inside or outside the area?
What kind of signal is sent through the wires and what kind of sensors
(coils?) are used to sense it?
I'll be happy to get any ideas, suggestions and information you might have.
Regards,
Hans Jakobsen
RedX
robot. When the robot get near the wire, the wire and the coil form a
transfo. The output of the coil is amplified and the frequency is measured.
RedX
"Hans Jakobsen" <please_answer@the_news.group> schreef in bericht
news:ajqddr$2i09$1...@news.cybercity.dk...
Gerry Schneider
Signal is similar to that of an underground dog fence - current thru the wire
sets up a magnetic field. The signal is picked up by a detect coil, and
increases very quickly as the 'bot gets near the wire. This alone won't give
precise perimeter detection - it's good enough for a dog, but not for mowing. To
detect the difference between inside and outside the coil's perimeter, the
current waveform is made non-symmetrical and the polarity flips phase by 180
degrees as the coil passes over the wire from one side to the other. Eg. -
positive ramp becomes negative ramp, or positive narrowish pulse becomes
negative narrowish pulse (or positive widish pulse if you prefer). The flip-over
region is very well defined as long as the wire isn't deeply buried. Some of the
better dog fences do this to allow the dog to re-enter the yard without a shock.
Dog fences use about 8-10kHz - don't know about the 'bot system. I build
proximity detect systems using about 15 kHz, and those telephone pick-up coils
that we used to stick onto the handset work just fine for symmetrical waveforms
- you might need a wider bandwidth coil for non-symmetrical use. The cheaper dog
fence collars resonate the coil with a cap to get enough signal to feed directly
into a (usually PIC) processor, so they can't be hacked for use with
non-symmetrical signals.
Good luck,
Gerry
Ralph
Does the dog need a special collar, implant??
remove n u m b e r s to reply
--
remove n u m b e r s to reply
Gerry Schneider wrote in message <3D61B801...@sympatico.ca>...
chad v
yep and if they get to close to the line shocks the crap out of them haha
"Ralph" <Smith....@team4.telstra6.com> wrote in message
news:ajs94p$9oe$1...@mws-stat-syd.cdn.telstra.com.au...
Kimball Sampson
contains all the actual energy. The gain on the loop is set so that
the collar picks it up, say, 5 feet from the actual wire.
An interesting point to remember here is that signals are vector
additive. If you double back the wire in a certain area, then the
signal is cancelled out. We have a detachaged garage where the unit
is based. The wire goes out, and comes back in along the same path so
that Buck won't get shocked when inside the perimeter.
Zeproo
You make them yourself? Do you have any schematics?
Regards
Hans Jakobsen
If the idea of stopping when the robot is outside the area with the
perimeter wire should work, then it needs to sense an "inside the area"
signal no matter where it is inside the area. then I suppose the perimeter
wire works as a coil with only one winding?
But the signal must to be very weak in the middle of the operational area
and very sensitive to disturbances from other fields. What if a power pylon
is crossing the area? Perhaps a Fourier could solve this?
How about the compass used to navigate the robot, will it be affected?
Regards,
Hans Jakobsen
"Gerry Schneider" <ger...@sympatico.ca> wrote in message
news:3D61B801...@sympatico.ca...
Gerry Schneider
> Thanks for the input Gerry.
> If the idea of stopping when the robot is outside the area with the
> perimeter wire should work, then it needs to sense an "inside the area"
> signal no matter where it is inside the area. then I suppose the perimeter
> wire works as a coil with only one winding?
Yes
>
> But the signal must to be very weak in the middle of the operational area
Surprisingly strong in a typical 2000 sq ft area that I'm familiar with. We
pulse about 2 amps through the perimeter wire. Have no data on the actual system
used with the mowbot.
>
> and very sensitive to disturbances from other fields. What if a power pylon
> is crossing the area? Perhaps a Fourier could solve this?
Large difference in frequency - 10 or so kHz vs 50 or 60 Hz
>
> How about the compass used to navigate the robot, will it be affected?
Possibly a flux-gate would be (don't know) - a DC Hall effect type isn't
bothered. I use the Dinsmore analog sensor for orientation.
Gerry
