GPS for range testing radio devices?

Un bel giorno larwe digitò:

IIRC one feet is about 0.3 m. Common GPS have absolute precisions around

2-3 m CEP. Therefore you need a *very* precise GPS, most probably a DGPS. 5 k$ are more likely. Something like this:

--
emboliaschizoide.splinder.com

I just tried my Garmin GPSmap 60Cx, and even though it shows the distance in miles, it can show 2 digits after the decimal point corresponding to 0.01 mile resolution (52.8 ft). If you don't mind metric, you can get 0.01 km resolution (32.8 ft)

The resolution here is tighter than I need. 25ft accuracy would be acceptable.

Not sure about off-the-shelf but if you're up for a bit of programming a PocketPC (iPAQ, of course) and a Bluetooth GPS would give you a platform for just about any custom GPS app. With the .NET Compact Framework you can use the CSerialPort class for fairly painless interfacing to the GPS.

Andrew

Oh no, I don't have time for custom-developing something. Basically I just want to order it for the guys that mostly do the range testing (it's usually not me, just my products :). Besides, I don't do Microsoft.

Ahh, I thought that might be the case. Well, there might be a niche there for someone then. I was going to mention that you can install Linux on iPAQs and other PocketPCs but it's even fiddlier but ultimately more rewarding.

Andrew

About any GPS has a serial output in NMEA 183 format at 4800 baud. Feed that serial stream to your link and receive it at the other end. using another identical GPS unit to compare with, you have DGPS! Take the difference in position to get range. Quite accurate over the range you are talking about.

Don

Except that "taking the difference in position" isn't trivial. There might be a way to use maps.google.com or some other bit cartography package, but writing and debugging a program to compute line-of-site distances between two points on the surface of the geoid is icky.

--
Grant Edwards                   grante             Yow! Awright, which one of
                                  at               you hid my PENIS ENVY?
                               visi.com

Calculating the "great circle" distance is fairly easy and possibly accurate enough for your needs. See for instance:

(Look for the spherical law)

If this is not accurate enough, click through to the Vincenty page, but that indeed gets more tricky. ;-)

--
Stef    (remove caps, dashes and .invalid from e-mail address to reply by mail)

For the distances involved, you can consider the earth flat. It will be close enough.

The calculations are quite easy. One minute of latitude is one nautical mile (~6076 feet). The longitude is the same except that must be multiplied by cosine of latitude. Bring both readings into a spreadsheet. Let it do the calculating. Dist = SQRT(X^2 + Y^2).

If you often do the range testing, set up a series of points at known distances.

Don

Why not just record the transmitter position to as many digits precision as the GPS will report, then record the receiver position as you move away? Plug the numbers into a couple of equations and voila!

NS distance = 6000 * latitude difference in minutes. EW distance = cos(lat) * 6000 * longitude difference in minutes

range = sqrt(NS * NS + EW *EW)

Range should be good to +/25 feet out to 5 miles.

If you plan to move away from the transmitter along a known path, you could pre-compute a few dozen ranges, along with coordinates, and simply move from one set of coordinates to the next for the tests.

Mark Borgerson

To the accuracy he's talking about, at the distances he's talking about (5 miles), he can ignore the geoid and assume the world is spherical. I doubt if it matters whether he gets 25000 feet or 25010 feet.

Un bel giorno Donald Harris digitò:

I wish it could be that simple! You can't assume that two identical modules "drift" in the same way, so that you increase the precision just by calculating the distance between the coordinates. In fact, identical modules drift in completely different ways even if they are at one meter from each other! If you use two modules in this way, you will get the exact same precision that you would get by using only one module.

DGPS works in another, more indirect way. The GPS satellites send some "correction parameters" that take account of several disturbances (ionospheric delay, slight changes in the trajectory of satellites, etc). However these parameters have some intrinsic precision limits, because you can't predict in realtime which will be the exact status of the disturbances at a given moment and at a given point on the Earth. Therefore, with DGPS you need to know *for sure* the exact position of one of your two modules (with a long term surveying or by using a more precise instrument). In this way you can apply the reverse formulas and precisely determine the correction parameters, so that you can send them to your mobile station.

--
emboliaschizoide.splinder.com

Un bel giorno larwe digitò:

You can buy for 200 or 300$ a GPS-equipped palm device (for example I own the Asus A636 which is quite good), then use some of the several low cost (or free) GIS related programs designed for Windows Mobile.

P.S. If you don't like M$, probably you will find an equivalent or better alternative with PalmOS or Linux or whathever, by spending twice the time and the money. But you will remain pure! ;-)

--
emboliaschizoide.splinder.com

Because this has to be a turnkey "it's working at 2000 feet" solution, something handheld and standalone and capable of being used by a technician. They explicitly do not want to carry a laptop around.

Could be. Somebody provided a link to a page that does calculations for both a spherical model and for what's-his-name's geoid. Before going with the spherical assumption, the OP could verify the accuracy of the spherical model at the location in question.

--
Grant Edwards                   grante             Yow! I once decorated my
                                  at               apartment entirely in ten
                               visi.com            foot salad forks!!

Quite true, but the interfaces are the same!

Don

Strangely I was doing some similar range testing with someone last week. They had a handheld GPS, think it was a Garmin Etrek or some similar name.

He just marked the transmitter as a waypoint and drove away getting a continuous reading of bearing and distance (in meters) to the transmitter waypoint.

I thought this is one of the most basic features of hand held GPS units.

Yes, the OP pointed out that you can do that. The problem described by the OP is that when range increases to a certain point, the units of the displayed distance switches fro feet (or meters) to miles (or km) with too few decimal places displayed for the resolution required by the OP.

--
Grant Edwards                   grante             Yow! Make me look like
                                  at               LINDA RONSTADT again!!
                               visi.com