Searching \ for '[OT] XY position determination' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=position+determination
Search entire site for: 'XY position determination'.

Exact match. Not showing close matches.
PICList Thread
'[OT] XY position determination'
2000\02\23@050323 by MacBrick

flavicon
face
Hi,

I am looking for suggestions to accomplish the following.
I have a board of about 4 by 3 feet. Subjects in a study are t place objects
on this board. I need to know the location where they pick up an object and
the location where they place it. The resolution is not that important, 16 x
32 squares or similar is sufficient.

Is there a simple way to accomplish this? I have constructed a similar
set-up but on larger scale (50 x 100 feet or so) using a matrix made up of
steel strips covered with a carpet. Something similar, but way smaller would
be nice.
I have also thought of a touch screen as used in PDA's. Does anyone know how
to construct something like that on a larger scale?

Other suggestions/experiences are welcome,


Ronald Leenes

2000\02\23@072542 by Pavel Korensky

flavicon
face
At 11:04 23.2.2000 +0100, you wrote:
>Hi,
>
>I am looking for suggestions to accomplish the following.
>I have a board of about 4 by 3 feet. Subjects in a study are t place objects
>on this board. I need to know the location where they pick up an object and
>the location where they place it. The resolution is not that important, 16 x
>32 squares or similar is sufficient.

You can make a matrix of IR LEDs and receivers. The distance of LEDs around
X and Y edges should be less than the size of the objects. When object is
placed on the board, you can exactly say where it is, because one of X and
Y receivers will be blocked.

Best regards

PavelK
 
**************************************************************************
* Pavel KorenskyÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
* DATOR3 LAN Services spol. s r.o.ÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
* Styblova 13, 140 00, Prague 4, Czech Republic      ÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
*ÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
* PGP Key fingerprint:Ê F3 E1 AE BC 34 18 CB A6Ê CC D0 DA 9E 79 03 41 D4 *
*ÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
* SUMMA SCIENTIA - NIHIL SCIREÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
**************************************************************************

2000\02\23@074833 by Stuart Willcocks

picon face
<x-flowed>How about light dependent resistors or infra-red emitter / sensor pairs
placed in the wooden surface facing upwards?
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com

</x-flowed>

2000\02\23@085540 by adavis

flavicon
face
I can think of a few methods, such as:

Put CDS cells on a grid beneath a table of plexiglass, and shine light from
above.

Use strips of metal cross-wise to each other with conductive foam between them,
and scan it like a keypad with an a/d converter

Shine lasers across the surface(about 2-5mm above the surface) in a grid pattern

If the objects have or can have magnetic properties, place a grid of coils
beneath the surface

Place a video camera above the surface, and make sure the colors contrast
between the surface and the objects

I hope this helps!

-Adam

MacBrick wrote:
{Quote hidden}

2000\02\23@110646 by Chris Eddy

flavicon
face
Do you have any say in  the composition of the objects?  Can they be ferrous?
You could put a hall sensor in each position, and if the objects have no magned,
put a bias magnet behind the hall.  Then the ferrous properties affect the hall
signal in each location.

Can you do optical?  This would presume that each object is centered in the
prospective square.  Do some sort of optical pickup at each location and
threshold a detection.  You could (in software) measure ambient light levels in
each of the squares and look for any squares that fall below this dynamic
ambient value by a certain percentage.

Chris Eddy

MacBrick wrote:

{Quote hidden}

2000\02\23@115846 by Wagner Lipnharski

picon face
Ahaaaa!  LEDS doesn't work.

If the board can be made by some translucid material as for example
white acrylic, and if you have enough ambient light, then you can instal
one cheap LDR below each object position.  Connect the LDR's in a matrix
to form 16 x 32 or two matrix 16 x 16.  You can produce it in such way
that only an object would be sensed, not shadows or things like that.
You probably would need two regular 40W fluorescent lamps above the
board to have enough light.

Your board is 4 x 3 ft, with 16 x 32 cells, so probably the 32 side
would be at the 4 ft side, leaving cells of 1.5" x 2.25".  This row and
column width is small to use LED matrix, since the 1.5" width row is 36
inches long, you would need to use focused lenses LEDs and Receivers,
since the angle opening is less than 2.4¡.

Another problem for LED matrix, is when you have multiple objects, you
can not identify where they are. See the example, LEDs emitting left to
right, top to bottom, sensors at right and at bottom:


Columns   A    B    C    D
         |    |    |    |
Row 1 --------------------- * Light    
         |    |    |    |
Row 2 --------B/2           Dark
         |         |    |
Row 3 -------------C/3      Dark        
         |              |
Row 4 --------------------- * Light
         |              |
         *              *
         ON  OFF  OFF   ON
        Light          Light

When you do it with keys, it is different, since keys "deviate, switch"
electric current, than a scanning system works..

At the matrix above you can "see" where the objects are, at B2 and C3,
but if the objects were at C2 and B3, the situation would be the same at
the sensors... right?  or wrong?

How can you tell where are the *only 2* objects at the matrix below,
based on the light/dark sensors, C3 and D4 or C4 and D3? (you can use
scanning, turn on one LED at time, etc). Of course there is an object in
Column C and another in column D, intersection with rows 3 and 4, but
where?:

Columns   A    B    C    D
         |    |
Row 1->-------------------------* Light      
         |    |
Row 2->-------------------------* Light
         |    |
Row 3               ??   ??      Dark        
         |    |
Row 4               ??   ??      Dark
         |    |
         v    v
      Light Light Dark Dark

The LDR solution (or even photodiodes below the board) would work, since
they would switch electric signals in the matrix.

Wagner

2000\02\23@121548 by Wagner Lipnharski

picon face
Again, obstructing light in a matrix beam system doesn't work for
multiple objects. is just like to use two switches per key in a matrix
system, one switch interrupts the rows line another interrupt the
columns, you would never know where are two or more objects.  Except if
all objects are in the same row or the same column.

Probably if you use a third axis (Z), horizontal as the others are,  but
in angle between X and Y you could isolate the multiple hits.


"M. Adam Davis" wrote:
[snip]
> Shine lasers across the surface(about 2-5mm above the surface) in a grid pattern
[snip]

2000\02\23@140730 by Alice Campbell

flavicon
face
How about drawing a thick pencil line around the perimeter, with
one dot in one corner, two dots in next corner, three dots in third
corner, and ticks along the sides for a grid.  use a whisker that
detects resitance of pencil line.  locate wall, feel around till you find
a corner, count dots to identify corner, and go back to where you
were when you got lost, and pick up what you were doing.

alice

{Quote hidden}

2000\02\23@142635 by andy howard

flavicon
face
> I am looking for suggestions to accomplish the following.
> I have a board of about 4 by 3 feet. Subjects in a study are t place
objects
> on this board. I need to know the location where they pick up an
object and
> the location where they place it. The resolution is not that
important, 16 x
> 32 squares or similar is sufficient.
>
> Is there a simple way to accomplish this? I have constructed a similar
> set-up but on larger scale (50 x 100 feet or so) using a matrix made
up of
> steel strips covered with a carpet. Something similar, but way smaller
would
> be nice.
> I have also thought of a touch screen as used in PDA's. Does anyone
know how
> to construct something like that on a larger scale?
>
> Other suggestions/experiences are welcome,

Someone posted this URL the other day. It seems to be pretty much what
you want.

http://qprox.com

Works through up to 2 inches of non-conducting material (they claim).









.

2000\02\24@065922 by Pavel Korensky

flavicon
face
At 11:54 23.2.2000 -0500, you wrote:
>Ahaaaa!  LEDS doesn't work.

:-)))) You are right. I forgot about the multiple objects.

PavelK

**************************************************************************
* Pavel KorenskyÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
* DATOR3 LAN Services spol. s r.o.ÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
* Styblova 13, 140 00, Prague 4, Czech Republic      ÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
*ÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
* PGP Key fingerprint:Ê F3 E1 AE BC 34 18 CB A6Ê CC D0 DA 9E 79 03 41 D4 *
*ÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
* SUMMA SCIENTIA - NIHIL SCIREÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊÊ *
**************************************************************************

2000\02\24@082129 by finevlad

flavicon
face
Maybe use ultrasonic coming from object at random time
and in parallel RF ID shows who it is
We can measure sound delay f.e. from RF TAG?

Vladimir


{Quote hidden}

pattern
>[snip]
>

2000\02\25@034105 by MacBrick

flavicon
face
Gee, I made some tongues loose.

I forgot to mention we're dealing with some 160 objects that have to be
placed on the board (and remain there). And yes, that means opto detection
is out.
Thanks for your input.
There are two workable options:
- some sort of keyboard made of a wire matrix covered with something. This
option relies on the subjects pressing down the object to place on the
board.
- the touch screen  (http://qprox.com/). I will look into this.

thanks

2000\02\25@112046 by Wagner Lipnharski

picon face
A switch wire grid works, but you will have flexibility problems for so
small cell sizes (1.5 x 2.25"). Except of course, if you create a
mechanical barriers around each cell, to avoid the object pressing the
wires of one cell to accidentally make wires contact at the adjacents
empty cells.

There is a high count involved here, there are 32x16 = 512 cells, 160
objects. It is a bunch of sensors to be used. Also thinking about how to
read this system (and the best way is the matrix keyboard), probably the
use of a reed switch buried under each cell, in a matrix formation.  In
quantity, a read switch can cost cents. Magnets could be installed at
the bottom of each object.  There was a small old electronic chess-board
that used this technique.

The use of *any* photo sensor at each cell will not work in matrix
scanning system, since you need the sensor to conduct (low resistance)
when the object is over it.  Reflective systems would cost much.

Qprox system is based on transferred charged energy (human body for
example), and it senses presence for certain amount of time, until it
"get used" with.

The most cheap system I can thing would use coils, with a piece of metal
at the bottom of each object. Imagine two small coils (side by side)
under each object cell, coils C and coils R.  All C coils at the same
COLUMN are in series. All R coils at each ROW are in series.  By sending
a scanning pulse in each ROW, the small piece of metal in each object
will help the magnetic field to be transferred from one coil to another,
so the COLUMNS coils will pick up that signal stronger and present at
the output. 4 op-amp LM324 can read all the 16 COLUMNS output and
generate a strong digital pulse. The pulse generation at the 32 ROWS
would be done by using 4 x 74HCT138, or 4 x 8 shift registers.
Experimental testing could identify the best value for the coils, wire
turns, etc., based on the metal piece at the bottom of each object.
Want to play with persons in large cells?  just attach a small metal
plate under the person shoe.  A single continuous wire that makes some
turns below each cell in each row, and in each column would do the
trick.

Wagner


MacBrick wrote:
> I forgot to mention we're dealing with some 160 objects that have to be
> placed on the board (and remain there). And yes, that means opto detection
> is out.
> Thanks for your input.
> There are two workable options:
> - some sort of keyboard made of a wire matrix covered with something. This
> option relies on the subjects pressing down the object to place on the
> board.

2000\02\25@113145 by Alan Pearce

face picon face
>The most cheap system I can thing would use coils, with a piece of metal
>at the bottom of each object. Imagine two small coils (side by side)
>under each object cell, coils C and coils R.  All C coils at the same
>COLUMN are in series. All R coils at each ROW are in series.  By sending
>a scanning pulse in each ROW, the small piece of metal in each object
>will help the magnetic field to be transferred from one coil to another,
>so the COLUMNS coils will pick up that signal stronger and present at

I remember HP describing this technique for a keyboard matrix in one of their
pieces of equipment in the past. The coils were printed circuit traces. It
evidently worked quite well, but other technologies have overtaken it for cost
effectiveness in keyboards.

2000\02\25@125706 by jamesnewton

face picon face
Optical from above should still be possible and quite possibly cheaper that
160 switches.
see
http://techref.massmind.org/robotics
in the eyes section and
http://www.cyberg8t.com/pendragn/actlite.htm

---
James Newton spam_OUTjamesnewtonTakeThisOuTspamgeocities.com 1-619-652-0593
http://techref.massmind.org NEW! FINALLY A REAL NAME!
Members can add private/public comments/pages ($0 TANSTAAFL web hosting)


{Original Message removed}

2000\02\25@152524 by Sam Littlewood

flavicon
picon face
If the objects are being placed on the board as discreet events - how about
a strain gauge at each corner, and look at the ratio of the change in
pressure as each object is put down?

SamL

2000\02\27@184649 by Peter Wintulich

flavicon
face
You could use a similar method to that of a digitizer.

under the board you could set up a grid of wires, there is a common termination for each axis.
Each wire runs accros the board moves up one grid space and back to the side, this forms a row of long thin loops on one axis. You end up with a common and 'n' connections down one side for one axis.
Do the same for the other axis.

Each object would need a coil of wire and a capacitor minimum. ( better a coil, PIC & battery.  This would allow individual objects to be identified through addressing the object to locate).

The number of wires ( loops) in the grid could be adjusted to suit the resolution required.
With the passive objects, a loop on each axis may need to be used simultaniously to identify an object location.( ie one as a transmitter and the other as an antenna.

With active objects, first broadcast the objects address, then lissen on each loop one at a time to find the strongest signal & therefor an objects location. In a digitizer more than one loop gets the signal & the ratio of the signal levels are used to give a finner resolution than the grid.

More... (looser matching)
- Last day of these posts
- In 2000 , 2001 only
- Today
- New search...