[Robotgroup] Triangulation

[Andre Lamothe]

The problem is really a two part problem. So the 1st thing you need to do is 
simply figure out a good way to compute energy of the radio signal, this is 
not easy my any means and you will probably have to buy some pre-made things 
to do it. But, many radio link chips have "signal strength" outputs that can 
be used to guage power very accurately. Since you are on a low budget, what 
you will probably have to make things yourself, this is not easy to do. You 
have to make a transmitter (easy enough to do with design on internet), then 
a receiver (same there), then you have to either calculate (hard), or use 
measurements to create a baseline for your calculations my measuring the 
signal strength at the receiver by putting it thru a detector circuit much 
like that of RFID to convert the signal to power or an average voltage, this 
will be related to the distance, not proportional because of the inverse 
square law. Then once you have your receiver putting out a voltage based on 
distance, but this isn't what you care about, you are just going to find the 
maximum of the energy in a 360 degree field. Each radio beacon will show up 
on your scan as a "hump" in your data, the center of each hump is the center 
of the beacon. Then you know the positions of the beacons, so you can draw 
lines to find the location of the receiver.

 So, what this means is that you need a revolving scanner that makes a 360 
turn very fast, as it does, it measures the radio energy at each angle 
(stepper motor needed), then you take the max energy and that's the center 
of the source, record it, and do the others.

Now, once you have a radio beacon working, the next step is to put 3 of them 
or 4 at corners of your field.

Then we have a vector problem with 3 vectors and the the position is the 
vector sum, you can think of it like a force problem, its the EASY part of 
this.

So first, explore making a transmitter, receiver and getting a voltage that 
is proportional to the signal, then the distance is a function of this 
voltage, the location of the beacon is the high or maxium point of the scan, 
which you draw a mathematical line to relative to your own coordiate system.

Lastly, 1-3cm, good luck with that :) I am working on a laser rangefinder, 
and even with that kind of state of the art technology 1cm out of 100 feet 
would be a miracle. So if you can get a few inches that would surprise me.

This problem can also be done "maybe" with IR lasers, then you could get rid 
of all the RF. As a test, one thing you might start with is get a IR 
emitter, an IR transistor, put the receiver on a stepper motor, then scan 
the stepper around, find the center of the energy, then put two emitters, 
then 3, find the center of each with one revolution scan, do this on your 
desk. This is a model for a bigger one.

Andre'



----- Original Message ----- 
From: "Robo Al" <sp-18561884 at ggsys.net>
To: "Austin Robotgroup Mailing List" <robotgroup at puremagic.com>
Sent: Friday, February 01, 2008 4:20 PM
Subject: [Robotgroup] Triangulation


Does anybody know any links or reference material
about non GPS triangulation?

I would like to play with something within these
requirements:

* open and closed spaces around 150'x100' max size (ie.
  a yard or inside a house)
* accurate to within 1 to 3 centimeters
* inexpensive. Hopefully sub $40 for parts.
* A computer or fast uC will be available to do the
  calculations so the $40 is just for sensors and I/O
  interfacing to the computer.

I would imagine setting 3 or 4 radio beacons within my
fence posts, but I'm afraid I just don't know enough
about radio triangulation.

Thanks,

Alberto


_______________________________________________
Robotgroup mailing list
Robotgroup at puremagic.com
http://lists.puremagic.com/cgi-bin/mailman/listinfo/robotgroup