Hooking Multiple Units Together

[Digital_Cowboy]

I was reading in the eTrex Legend that one could interface multiple units together. In doing so would that improve the accuracy? Does Garmin (or anyone) make a cable to link two or more GPSr's together, or two GPSr's to a computer? I don't suppose that there's any sort of bluetooth interface for hooking two or more units together?

 

IF joining two or more GPSr's together improves their accuracy, how far apart do they have to be, and which would be the master unit?

Edited May 6, 2006 by Digital_Cowboy

[+PDOP's]

I think you mean Differential GPS (see here)

 

Commercial survey grade receivers usually work with a base receiver setup over a known point and a rover unit which is used for the actual field work. Corrections from the base can be used in realtime (with a telemetry link) or post processed afterwards.

[+Geo-Cad]

I think you mean Differential GPS (see here)

 

Commercial survey grade receivers usually work with a base receiver setup over a known point and a rover unit which is used for the actual field work. Corrections from the base can be used in realtime (with a telemetry link) or post processed afterwards.

 

We use GPS at work to layout buildings, The rep says it's 1mm in 4,000ft. But then it does cost over 100k

[Neo_Geo]

A man with ONE GPSr always knows where he is.

A man with TWO is never really sure.

[Digital_Cowboy]

I think you mean Differential GPS (see here)

 

Commercial survey grade receivers usually work with a base receiver setup over a known point and a rover unit which is used for the actual field work. Corrections from the base can be used in realtime (with a telemetry link) or post processed afterwards.

 

In Appendix F it talks about the wiring diagram and it says this:

 

Interface formats are selected from teh Setup 'Interface Page' on page 46 of this manual. The input/output lines on your eTrex Legend unit are RS-232 compatible, allowing easy interface to a wide range of external devices, including PC's, differential beacons receivers, marine autopilots and/or a second GPS receiver.

 

The NMEA 0183 version 3.0 interface format is supported by the eTrex Legend and enables the unit to drive up to three NMEA devices.

 

Ok, so how does one go about hooking up two GPS'? And again what would be the benifit of doing so?

[dave and jaime]

while you wont be able to get 'survey' type accuracy, it is common practice to place one unit over a known point and do the roaming with another reciever. the base station display would be recorded at the same time as the roamer and corrections(the corrections being the difference between the base station and the known coordinate) would be applied after the fact to the roamer recordings, this would be the post-processing mentioned earlier.

 

obviously this wont give you accuracy as obtained by commercial gear, as precision of recreational units is limited to 1 meter at best, but is frequently done in many geomatics/cartographic university programs to introduce students to the process. if you are using similar(read same manufacturer) units for the base and roamer you can be reasonably assured that the corrections are valid and would probably provide a better set of coordinates than can be achieved by current waas corrections. using a process like this would be essentially 'tying to a control'.

 

if accuracy is your goal, it would also be worthwhile to checkout one of the mission planning software packages, leica makes a very good one.

[dave and jaime]

I think you mean Differential GPS (see here)

 

Commercial survey grade receivers usually work with a base receiver setup over a known point and a rover unit which is used for the actual field work. Corrections from the base can be used in realtime (with a telemetry link) or post processed afterwards.

 

In Appendix F it talks about the wiring diagram and it says this:

 

Interface formats are selected from teh Setup 'Interface Page' on page 46 of this manual. The input/output lines on your eTrex Legend unit are RS-232 compatible, allowing easy interface to a wide range of external devices, including PC's, differential beacons receivers, marine autopilots and/or a second GPS receiver.

 

The NMEA 0183 version 3.0 interface format is supported by the eTrex Legend and enables the unit to drive up to three NMEA devices.

 

Ok, so how does one go about hooking up two GPS'? And again what would be the benifit of doing so?

 

i think that maybe connecting to a second gps would be similar to connecting 2 graphic calculators together to transfer data from one to the other. i dont think that 2 gps can be connected directly to give improved accuracy.

[peter]

while you wont be able to get 'survey' type accuracy, it is common practice to place one unit over a known point and do the roaming with another reciever. the base station display would be recorded at the same time as the roamer and corrections(the corrections being the difference between the base station and the known coordinate) would be applied after the fact to the roamer recordings, this would be the post-processing mentioned earlier.

The technique you're describing is sometimes called "poor man's DGPS" and generally doesn't work well at all. The problem is that even with two receivers in reasonable proximity there's still no assurance that they'll be using the same set of satellites in performing their position calculation. If different satellites are being seen by the roaming unit then its errors may not be at all correlated with those of the fixed unit. I.e. one may be showing a position too far to the west while the other is off to the east and applying the 'correction' in that case will only make the position worse. Various averaging and similar techniques used by the Kalman filter algorithms in the receivers further uncorrelate their errors. Use of WAAS would be far superior and also more convenient in areas where it's available.

 

True DGPS determines corrections for each satellite separately so the roaming receiver can apply them properly to the satellites it is seeing at the time. When post-processing is used the roaming receiver records information separately for each satellite so the per-satellite corrections can be properly applied and corrected positions determined.

 

 

One application for sending the NMEA stream from one GPS receiver to another is to have the second receiver display the position of the first on its screen. Of course you need a communication link between the two to make this useful. It's sometimes used in combination with APRS Ham radio networks to display the positions of multiple GPS receivers. I don't know if the Legend model in particular supports that application. (The Garmin Rino models use the FRS/GMRS radio channels to send this information and display the positions of other compatible models.)

Edited May 7, 2006 by peter

[+PDOP's]

i think that maybe connecting to a second gps would be similar to connecting 2 graphic calculators together to transfer data from one to the other. i dont think that 2 gps can be connected directly to give improved accuracy.

IIRC you're right. My old GPS12 could do this and had an interface option for Garmin/Garmin

[peter]

i think that maybe connecting to a second gps would be similar to connecting 2 graphic calculators together to transfer data from one to the other. i dont think that 2 gps can be connected directly to give improved accuracy.

IIRC you're right. My old GPS12 could do this and had an interface option for Garmin/Garmin

Yes, the older Garmins had that option where you could set one unit to be the HOST and transfer waypoints/tracks directly from one to the other with just a cable between them. But I think they dropped that on the eTrex and newer models.

[dave and jaime]

while you wont be able to get 'survey' type accuracy, it is common practice to place one unit over a known point and do the roaming with another reciever. the base station display would be recorded at the same time as the roamer and corrections(the corrections being the difference between the base station and the known coordinate) would be applied after the fact to the roamer recordings, this would be the post-processing mentioned earlier.

The technique you're describing is sometimes called "poor man's DGPS" and generally doesn't work well at all. The problem is that even with two receivers in reasonable proximity there's still no assurance that they'll be using the same set of satellites in performing their position calculation. If different satellites are being seen by the roaming unit then its errors may not be at all correlated with those of the fixed unit. I.e. one may be showing a position too far to the west while the other is off to the east and applying the 'correction' in that case will only make the position worse. Various averaging and similar techniques used by the Kalman filter algorithms in the receivers further uncorrelate their errors. Use of WAAS would be far superior and also more convenient in areas where it's available.

 

True DGPS determines corrections for each satellite separately so the roaming receiver can apply them properly to the satellites it is seeing at the time. When post-processing is used the roaming receiver records information separately for each satellite so the per-satellite corrections can be properly applied and corrected positions determined.

 

 

One application for sending the NMEA stream from one GPS receiver to another is to have the second receiver display the position of the first on its screen. Of course you need a communication link between the two to make this useful. It's sometimes used in combination with APRS Ham radio networks to display the positions of multiple GPS receivers. I don't know if the Legend model in particular supports that application. (The Garmin Rino models use the FRS/GMRS radio channels to send this information and display the positions of other compatible models.)

 

granted, the sat group used for calculations may not be the same but the if using 'similar' gps units the errors introduced by the algorithims would infact be nearly eliminated, not magnified. also by using a time sync, enviromental errors, save for multipath, would largely be eliminated. that leaves errors introduced by using different sats, since the gps system is only generally accurate, in practice use confirms that you dont get any better accuracy using sats 8, 11, 15, 24, 25 as opposed to 13, 17, 6, 28, 30. one thing i never mentioned in my original post is that allowing sufficient averaging time for the gps to gain an accurate fix is critical--whether the unit is a consumer or professional unit.

 

i guess my point is that this system, as used here, was never meant to give meter accuracy and further to that, in situ experience shows that the system is reasonably consistant and using this type of procedure doesnt markedly degrade accuracy.

 

sorry for being ignorant/uniformed, but do the rhinos send their positional data over the gmrs/frs bands to other rhino units?

Edited May 7, 2006 by dave and jaime

[peter]

granted, the sat group used for calculations may not be the same but the if using 'similar' gps units the errors introduced by the algorithims would infact be nearly eliminated, not magnified. also by using a time sync, enviromental errors, save for multipath, would largely be eliminated. that leaves errors introduced by using different sats, since the gps system is only generally accurate, in practice use confirms that you dont get any better accuracy using sats 8, 11, 15, 24, 25 as opposed to 13, 17, 6, 28, 30. one thing i never mentioned in my original post is that allowing sufficient averaging time for the gps to gain an accurate fix is critical--whether the unit is a consumer or professional unit.

But the point was not whether using one group of satellites would be better than another (that would depend on their relative geometries), but on whether one could count on the errors of the two receivers being correlated even if they happen to use different satellites. The data I've seen has indicated that such correlation is rather small and therefore 'poor man's GPS' is not a practical way to improve accuracy. Another discussion on the subject is at:

http://gpsinformation.net/main/poordgps.htm

If you have some study results in mind that showed good results using such techniques then I'd be interested in seeing them.

i guess my point is that this system, as used here, was never meant to give meter accuracy and further to that, in situ experience shows that the system is reasonably consistant and using this type of procedure doesnt markedly degrade accuracy.

I wouldn't expect it to degrade accuracy. But the whole point of the exercise is presumably to improve accuracy and I don't think it'll do that significantly either.

do the rhinos send their positional data over the gmrs/frs bands to other rhino units?

Yes, if several people in a group are all using Rino models they can keep track of where the others are located.

Edited May 7, 2006 by peter

[dave and jaime]

peter:

 

i think we ust have crossed wires somehow. i knew that i could send my position on one rhino to another unit, but is their a way for the units to do this automatically and record the data in any meaningful fashion? are there any other recreation units that can automatically track each other?

 

in rereading your previous post, it gives me the impression that you mean that part of the issue was that different sats would in fact make the positioning fix less accurate.

 

in leading a engineering geomatics course, there is an exercise given to the students to introduce them to the idea of dgps. while not scientific the results i found to be interesting. the exercise is setup such that each groups has 2 identical non-waas units. one unit is placed over a horizontal control point and the roamer then cycles several points over a closed survey loop. this is repeated for several averaging times ranging from 1 min to 10 min at each station. generally, the results showed that using a simple shift equal to the error at the base station recording brought the roamer into better alignment with the station. this is an excerise in exploiting the consistancy of the system, and the most important variable appeared to be the averaging time at each station---averaging beyond 5minutes seemed to offer little benefit but averaging of less than a minute provided no worthwhile correlation. the results seemed to suggest a reliable 3-4 meter range of accuracy.

 

ive run this exercise for several years and the results seem to be fairly consistant and at least as reliable as the gps itself.

[+Team Cotati]

Not even if I were to live to be 1,000.

[peter]

i think we ust have crossed wires somehow. i knew that i could send my position on one rhino to another unit, but is their a way for the units to do this automatically and record the data in any meaningful fashion? are there any other recreation units that can automatically track each other?

Not as far as I know. The FRS band was not really intended for data communications (only voice) and my recollection is that Garmin got a special authorization from the FCC to permit data transmission. But it's still a rather restricted use and I don't think it allows for continuous automatic sending of the positional data.

in rereading your previous post, it gives me the impression that you mean that part of the issue was that different sats would in fact make the positioning fix less accurate.

That was certainly not the intention. The errors between the two units will certainly show some degree of correlation so in general I'd expect more cases where there's some improvement than cases where there's degradation. But the results I've seen have indicated that the correlation is pretty weak and therefore the technique isn't of significant practical value.

in leading a engineering geomatics course, there is an exercise given to the students to introduce them to the idea of dgps. while not scientific the results i found to be interesting. the exercise is setup such that each groups has 2 identical non-waas units. one unit is placed over a horizontal control point and the roamer then cycles several points over a closed survey loop. this is repeated for several averaging times ranging from 1 min to 10 min at each station. generally, the results showed that using a simple shift equal to the error at the base station recording brought the roamer into better alignment with the station. this is an excerise in exploiting the consistancy of the system, and the most important variable appeared to be the averaging time at each station---averaging beyond 5minutes seemed to offer little benefit but averaging of less than a minute provided no worthwhile correlation. the results seemed to suggest a reliable 3-4 meter range of accuracy.

But that seems right in line with the level of accuracy obtainable using judicious averaging and only a single receiver as indicated by Wilson's results with non-WAAS data:

http://users.erols.com/dlwilson/gpsacc.htm

Adding WAAS brings the measurement uncertainty down further and when combined with short-term averaging can get you down to the single meter range.

ive run this exercise for several years and the results seem to be fairly consistant and at least as reliable as the gps itself.

Sure, but then you've gone to a lot of trouble setting up a base station and doing extra processing for very little gain. And it doesn't seem to support your initial statement that this method "would probably provide a better set of coordinates than can be achieved by current waas corrections" at least in areas that are properly covered by WAAS.

[dave and jaime]

deleted by user

Edited May 7, 2006 by dave and jaime

[Digital_Cowboy]

WOW, looks like I opened a "can of worms" when I started this thread.