I experimented with a solution that might help.

imagine a square arena or room.
Two servos are placed. One in each of the lower left and lower right
hand corners.
Each servo has an IR TX LED assembly that transmits the direction that
that servo is pointing (0 to 90 degs).
(Since each servo is located in a corner, it only needs to sweep
backwards and forwards 90 degs. Mine were set to complete a 90deg sweep
in about 1 second.
Along with the IR telemetry, a servo ID is also sent so that the robot
knows which IR signal is being received at any time.
If the sweep rate is fast enough the robot should be able to get a very
accurate fix on it's position within the area.
The trick was to make sure that both IR LED's aren't transmitting at the
same time which just confused the robot CPU.
You will need to experiment with a good IR collector for the robot.
Also, it works a lot better in open space than a room sine the IR
bounces all over the place.
I did try building a laser version using a couple of those cheap laser
pointers you can get, and it did look really cool with red laser beams
sweeping across the room, but that put the customer off a bit - funny
really !

Does that make sense.



-----Original Message-----
From: Madhu Annapragada [mailto:mapr@COMCAST.NET]
Sent: 22 May 2002 13:58
To: PICLIST@MITVMA.MIT.EDU
Subject: Re: [EE]:Positioning


Not a simple answer to this I am afraid; the problem with absolute
positioning for mobile robots, as you have no doubt discovered in the
myriad
of literature out there, is that you need more than one positioning
device
to even out the errors one accumulates. Typically, one uses wheel
encoders
and some other sort of triangulation methods (IR, RF etc) on a periodic
basis to correct the errors accumulated by the wheel encoder measurement
(due to wheel slippage). Since you have a walking and not a rolling
base,
one option is to use IR triangulation. This works on the principle of
three
or more IR emitters (each distinct from the other say in the frequency
at
which they turn on and off) and a receiver that can be rotated on the
base
to receive these emissions. One could also use RF along the same lines
and
that will give you greater freedom at the cost of increased complexity.
Some
groups have used barcodes placed at strategic points in the environment
and
a scanner on the mobile base. Some others use fixed markers that emit IR
at
strategic locations. The key is to get multiple inputs and then use
software
to obtain the best estimate.
Madhu

> I wonder if there's anyone that have any good? ideas about small-scale
> positioning?
> Scenario: You have a small robot walking around finding obstacles.
When
you
> find
> these you want to store the position.
> As this is just a hobby thing it can't cost too much so things like
gps is
> not in
> question, if even applicable.
>
> TIA
> Niklas
>
> --
> Niklas Lovgren

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads