Is it possible to build GPS "like" system for high-accuracy local position finding?

[Richard H.]

It would incorporate 4 GPS like transmitters and a number of receiver
units. This would allow me to read the receiver unit's relative
position in area surrounded by these transmitters.

Explore beyond the GPS model. Who says the mobile devices need to be
the receivers (which requires some hefty horsepower to decode &
triangulate the signals).

* Use a simple transmitter fob on the mobile object.
* Have the object transmit occasionally (as frequently as you desire).
* Have N towers listen for the signal (however many it takes to cover
the site)
* Towers timestamp the signal with very high (nS?) accuracy. Distance
from each tower is calculated based on propagation delay.
* Calculate (x,y) from the 3 closest towers, or (x,y,z) from the closest 4

Sounds a lot like cell tower triangulation, eh?

Options:
* Device transmits very infrequently when stationary
* More frequently when accelerometer detects movement

Using such a system, the trick would be keeping the time sync'd very
accurately between the towers. (Separate GPS receivers, or calibrated
inter-tower cabling.) To really nail the (x,y) accuracy, the field
could be calibrated by walking around with a transmitter to correct for
multipath effects.

This approach lets you keep the mobile units very cheap and put the
power-sucking number-crunching in the towers. If the mobile fobs need
to know their location data, address it to them on a return channel from
one of the bases.

Incidentally, at least one asset tracking product works very similarly,
but the (x,y) resolution is something like 10m. Their fobs run for
years on a coin cell. (And no, I can't find their card anywhere here...)

Cheers,
Richard

 

[Mac]

I thought more about this!

How about:

No walls, large wharehouse, all transmitters hardwired together (to
sync clocks).

Direct line of site from transmitter to receiver.

Would it be possible then?

I don't know. I don't have high hopes. I would still want to use
ultrasound for this. If you correct for air-temperature effects, the
possible precision is pretty good. I thought of another ultrasound
website: massa.com.

Now, if you were to get rid of the roof on the warehouse, or make it out
of an RF transparent material, then you could just use differential GPS.
Honestly, this might be cheaper. ;-)

If you must use RF, and design it yourself, then I think you will have to
keep it simple. Designing a GPS-like system will be tricky. Maybe you can
use something simpler, like some kind of UWB ranging system. Take a look
at time domain's website (www.timedomain.com) for ideas. Maybe even give
them a call?

If you do use RF, then anything you can do to avoid reflections from the
walls, ceiling, and floor will help tremendously. But covering the entire
warehouse with RF absorber might get expensive and ridiculous.

Good luck.

--Mac

 

[Iwo Mergler]

Pretend I was a seasoned EE professional. Could I build this on my
own? Without huge budget and lots of expensive equipment!

The one thing I have an excess of is....time!!

I see you have a few years spare... ;^)

Here is an open source GPS design based on
a rather obsolete RF frontend and correlator
IC:

http://home.earthlink.net/~cwkelley/

That's a starting point. You may be able
to adapt it to a different frequency band
and modulation scheme to improve accuracy.

Also read the GPS documentation. The civilian
stuff is freely available.

http://www.navcen.uscg.gov/gps/geninfo/default.htm

This one practically tells a seasoned EE professional
how to build a receiver:

http://www.navcen.uscg.gov/gps/geninfo/IS-GPS-200D.pdf

Kind regards,

Iwo

 

[blarggstar]

FOUND SOLUTION!!!!

I found what I was looking for. "Indoor position tracking" was the key
search in google. Duh!

http://cricket.csail.mit.edu/

It is pretty accurate 1 - 2 cm. This will have to do!! Uses
ultrasound and RF.

This is opensource software and hardware, as they give the circuit
diagrams along with parts list.

My only question now is, can you assemble this with a saudering iron?
http://cricket.csail.mit.edu/software2.0/Cricket_Mote_Schematic_6310-0335-02_b.pdf

Parts are around $50 bucks, I'm going to try!! Wish me luck.

Crossbow technologies sells them assembled, but at $225 a peice, I'm
going to try building one myself.
http://www.xbow.com/Products/productsdetails.aspx?sid=116

The price is very cheap, but $50 is a whole lot cheaper!!

I'm pretty excited about this, thankyou for all the advice.

 

[blarggstar]

EVERYONE I FOUND A SOLUTION!!!!

I found what I was looking for and the price is right!!!
"Indoor position tracking" was the key search in google. Duh!

http://cricket.csail.mit.edu/

It is pretty accurate 1 - 2 cm. This will have to do!! Uses
ultrasound and RF.

This is opensource software and hardware, as they give the circuit
diagrams along with parts list.

My only question now is, can you assemble this with a saudering iron?
http://cricket.csail.mit.edu/s oftware2.0/Cricket_Mote_Schema
tic_6310-...

Parts are around $50 bucks, I'm going to try!! Wish me luck.

Crossbow technologies sells them assembled, but at $225 a peice, I'm
going to try building one myself.
http://www.xbow.com/Products/p roductsdetails.aspx?sid=116

The price is very cheap, but $50 is a whole lot cheaper!!

I'm pretty excited about this, thankyou for all the advice.

NOTE: This is a double post, and I am sorry.

 

[MSC]

I know you have found a solution, but here's a thought.

Instead of 'n' static transmitters and 1 mobile receiver, how about
using 1 mobile transmitter and 'n' receivers? That should be easier to
calibrate and less prone to reflections.

Have fun )

Mike

 

[Sylvain Munaut]

My only question now is, can you assemble this with a saudering iron?
http://cricket.csail.mit.edu/s oftware2.0/Cricket_Mote_Schema
tic_6310-...

Sure they're pretty big SMD AFAICS.

Parts are around $50 bucks, I'm going to try!! Wish me luck.

Crossbow technologies sells them assembled, but at $225 a peice, I'm
going to try building one myself.
http://www.xbow.com/Products/p roductsdetails.aspx?sid=116

The price is very cheap, but $50 is a whole lot cheaper!!

Well, huh ...
You need to design the PCB first because I didn't see it on they're
site. And it will most likely be a 2 layer PCB so you have to add that
to the price.

And there is the time it will take you to do it.


Sylvain

 

[Spehro Pefhany]

Sure they're pretty big SMD AFAICS.


Well, huh ...
You need to design the PCB first because I didn't see it on they're
site. And it will most likely be a 2 layer PCB so you have to add that
to the price.

And there is the time it will take you to do it.


Sylvain

The ATMega is available either of two 64-pin packages, neither of them
all that easy to work with on a breadboard.


Best regards,
Spehro Pefhany

 

[blarggstar]

Here is my reasoning:

For initial prototyping phase, I'll need five crickets. $5 x 225 =
$1125

For the second phase prototype, I'll need 30 - 50 crickets ~$6750 to
$11250. But I'd imagine getting a discount at that quantity, so this
isn't accurate. My point is, I need a lot of these things.

If I could design the PCB, (which I currently have NEVER done) and
solder them by hand, this would save a HUGE amount of money.

It would take a long long time for me to do this, but I have lots of
time and no money.

Finished product size isn't a huge issue, so I'll make them bigger then
they are when you buy them. This will make it easier for me to create
them.

A couple of ICs have a lot of pins. (20 - 50). Do these guys have
"seats" or "connectors" so I do not have to try to solder the IC
directly?

I'm researching this plenty but any advice is a great help.

 

[blarggstar]

I have played with that idea, but my object in motion, needs to know
his own coordinates in the space. Somehow this would have to be
trasmitted back to him.

I guess that is possible though and the problems I'd face would seem
easier then doing it the other way.

 

[blarggstar]

I'm hoping just to solder everything and skip the breadboard
alltogether. Does it have some kind of "seat"? Are their universal
seats for many pin IC's like this?

I'm still looking...

 

[Sylvain Munaut]

Here is my reasoning:

For initial prototyping phase, I'll need five crickets. $5 x 225 =
$1125

For the second phase prototype, I'll need 30 - 50 crickets ~$6750 to
$11250. But I'd imagine getting a discount at that quantity, so this
isn't accurate. My point is, I need a lot of these things.

If I could design the PCB, (which I currently have NEVER done) and
solder them by hand, this would save a HUGE amount of money.

Well, sure you can. But solder 30-50 units is kinda boring

It would take a long long time for me to do this, but I have lots of
time and no money.

The schematics is fairly easy, shouldn't take too long. Of course if you
don't already know any CAD program, that may take some time.

As an advice : the harder is to place the component so that the
connection are easy to make. Since you have a photo of a finished
product just use their placement

Finished product size isn't a huge issue, so I'll make them bigger then
they are when you buy them. This will make it easier for me to create
them.

The bigger it is, the expansier the PCB will cost. For 50 units that
should be < 7.5$ per PCB.

A couple of ICs have a lot of pins. (20 - 50). Do these guys have
"seats" or "connectors" so I do not have to try to solder the IC
directly?

??? that's not that hard to solder, it has a pretty big pitch, hard to
screw up with a good iron (and if you're planning on doing 50 units, you
have to buy one .

Just look up on the net how to solder SMD's properly. (try 'hand smd
soldering').


Sylvain

 

[Fred Abse]

No walls, large wharehouse

Warehouse?
Whorehouse?

 

[Michael A. Terrell]

Warehouse?
Whorehouse?

A "wharehouse" sounds like a place to keep old politicians after they
have finished plying their trade.

--
Link to my "Computers for disabled Veterans" project website deleted
after threats were telephoned to my church.

Michael A. Terrell
Central Florida