Measuring Distance via Radio Waves

[Ztrack]

Does anyone know of a way to use radio to radio transmission to determine the distance between the two radios with an error of about a meter? These radios would need to be small and economical - preferably transmitting on a free frequency with the longest range possible.

 

[IRstuff]

That's done with two differential GPS units and radio modems ;-)

TTFN

FAQ731-376

 

[VisiGoth]

If you still want to create your own system and you can afford the cost of free spectrum then do the math. SNR and BW determine distance variance per the CRB. Search those terms or find them in the books. You need bandwidth to overcome noise. You will find that the open spectrum using the rules for those channels does not meet one meter as the root of the variance for the CRB for any reasonable distance. You should define your systems engineering a little better, like what distance is your threshold requirement and what is your goal?

Doing your own design may be interesting, and sounds like a schoolboy project. But it could be a legit hobby question.

IRStuff gave you the answer for a commercial quality product. You don't even have to put up your own satellite constellation.

 

[MacGyverS2000]

Sounds like a senior design project...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Ztrack]

I suppose I should explain a bit further. I will have 100s or even 1000s deployed in a city - each in motion (on automobiles). The ultimate goal is to approximate location without needing a pricy GPS in each one. I can do the math with only a couple known locations and extrapolate based on distances and trilateration. Also, I'd like to be able to determine location without a clear sky view. I had thought about an FM receiver to set an internal clock via the atomic clock broadcast and then use time differentials. Is there a free frequency that can be used to send data over 1km with a small inexpensive unit? Or any other ideas?

 

[IRstuff]

If GPS can't work, neither will your radio.

I'm unclear what you mean by "pricey." Do you really think that your custom design will be cheaper than a mass-produced GPS, particularly since you need as good as, if not better, timing performance, coupled with no simple means of synchronizing the time? Without something like a cell tower network, the amount of RF you need to generate will not be in the free regime. Without a GPS-like modulation scheme, you're going to get garbage for time synchronization in a typical city with tons of multipath.

You seem to be ignoring the overall problem of multipath and seem to think that this is a trivial problem. If it were that easy, GPS would have been supplanted by now. Have you even considered how accurate your timebases need to be?

I suggest that you look that one of the following as a more realistic alternative:

> GPS-equipped cell phones

> a cell carrier that will give you triangulated data from their cell towers for non-GPS cell phones

> Earthmate LT-20 only costs about $60. If wired to a simple USB-equipped microcontroller, you could do the hardware for less than $180 per installation.



TTFN

FAQ731-376

 

[VisiGoth]

You may be thinking about toy GPS like the ones built for cars that need nearly line of sight. The higher sensitivity used in cell phones can probably receive as well as you think you can do at your own frequency.
Contact the telematics group at the major cell phone companies. They specialize in making just this sort of device for people like you. You are their target market. They take sensitivity and reliability seriously.

In your own design making ad-hoc networks is not trivial. Also if you have a unit that has visibility to only other units in a plane or even worse co-linear, then doing the math to get location has issues.

 

[VEBill]

One leading vendor of 'hot' GPS chips is

http://www.sirf.com

.

 

[Comcokid]

GPS modules can be found for under $30 each in low volumes A quick random search from Digikey shows part number GM-205610-000TR-ND shows it to be $26 at around 350 pieces. Send the GPS data out over the "radio", or add a radio or transceiver module, and you have the foundation of a system. Would require a little development work. Actually, the radio modules and antenna would be the most expensive component, not the GPS.

 

[Higgler]

Phase switching of a CW signal is the method used by one company to detect range of an aircraft from a ground station.
If you use pulses, that works but is often messy. This company in San Diego (forgot the name) would set up a CW, or nearly CW signal and change the phase of the signal on transmit, the receiver would pick it up and transmit on a different frequency the same received signal. They knew the inherent time delay thru the repeater and the original transmitter measures delay in the received phase shifted signal.
The primary reason this technique is good is that ground bounces don't change much in the time it takes for the energy to fly out and return back to you. (about 10.5 microseconds per mile distance between the two points).

I'm not sure if there are any transceiver modules readily available to use this technique, but maybe you can build your own hardware. Video and RF/IF bandwidth considerations will affect the actual rise time of the 180 phase shift and hence the distance measuring accuracy of the system. Not certain of the details.

kch

 

[BrianR]

The key to this is to measure propogation time to 3ns or better and about the only practical way of doing this
is by using a carrier modulated by a pseudo random sequence (which as VisiGoth points out results in a spread spectrum signal). Eg a 10 bit PRS at a 5MHz chip rate will do it provided you have a good correlator in the receiver.

The frequency must be UHF or higher to allow the bandwidth needed so range will be limited limited.

You'll need a good Kalman filter in each RX to eliminate the effects of multi-path and the variations in propogation delay you'll encounter.

Using DGPS is going to be a lot cheaper although even that will only give the position of each unit.

 

[BOMBOVA]

Ihave seen this done in 1966 by Cubic Corp. withmicrowave eqipment.

http://cubic.com/cts/

takes me back to my youth. fun stuff. i have pictures of the artic from then also... lots of ice... GerwingR

 

[Higgler]

Thanks, Bombova, that's the company I worked with in the mid 80's, Cubic from San Diego. They let the multi path bounces settle out in each pulse, then did a 180 degree phase switching. Pretty simple concept.

kch

 

[VEBill]

I worked on project to modify a Medium Wave navigation system so that the phase of the groundwave would have already been measured before the skywave arrived and messed it all up. The proof-of-concept prototype worked great. But this was just about the same time that GPS was coming on stream.

 

[SamuraiMike]

What about omega or loran and a lot of averaging? A meter is really tight. I think omega is completely shutdown, but loran still exists. Both are pre-GPS nav systems and how they worked might provide your answers. Omega was LF and based on phase.

 

[IRstuff]

Omega is dead:

http://en.wikipedia.org/wiki/OMEGA_Navigation_System

Standard Loran C accuracy was 460 m:

http://www.navcen.uscg.gov/eLORAN/overview.htm

TTFN

FAQ731-376

 

[sreid]

Below, interesting high resolution time to digital IC's.

http://www.acam-usa.com/Content/English/acam.html