Dual GPS Rcvr's. Unusual thought...

d_sellers1 · 06-21-2004, 09:26 AM

Just been doing a bit of thinking (hate when that happens).

Logically, the best spot for a GPS receiver would be right in the center of the roof. Now that's not gonna happen. That'll be ugly and require a hold to be drilled in the roof. Naah, I think I'll pass on that.

Picking either the rear deck or on the front dash limits the receiver's view of the sky. Also, depending on direct of travel along with any obstructions.

Do you guys think it's possible to get two receiver's and hardwire them together onto a single USB cable placing one receiver in the front and one in the back. In theory, the software should only see one device but get the information from both perhaps also increasing accuracy because it will be able to "see" more satellites.

What do you guys think...

EDIT: A couple of good points have been brought up in this thread.
1. The software may not like getting multiple data streams.
2. It may be possible to use multiple antennas instead.
3. The GPS won't use all the sats. at once (only 3 are needed for position, 4th for elevation).

Derek

dodgetimes2 · 06-21-2004, 10:17 AM

I was pondering this as well, but it might also "confuse" the software when it's constantly getting 2 sets of different data strings. The 2 sets of data strings will always differ by X amount of feet (X=feet between receiver 1 and 2). Would like to see someone test it though.

[iG] · 06-21-2004, 10:36 AM

Chances are your gps reciever wont even get anything at the back window, as many have people found out the hard way...

d_sellers1 · 06-21-2004, 10:42 AM

Quote: Originally Posted by [iG]

Chances are your gps reciever wont even get anything at the back window, as many have people found out the hard way...

It probably depends on the angle of your back window and the GPS placement. I've got mine in the front window in the top corner. Works pretty good the majority of the time (when testing with laptop).

Derek

Laidback · 06-21-2004, 10:42 AM

IMO, Zero chance with USB and possible with Serial but how do you control when each can send a string? If they both send together you'll end up with garbage.

Yuriy · 06-21-2004, 12:50 PM

Why have two GPS receivers when you can just have two external antennas connected to one receiver?

carpcnoobie · 06-21-2004, 12:58 PM

IMO, I think one GPS antenna on either the deck lid or the front windshield will have enough of a view at all times to get a good signal. I don't think you'll see too many more sats with one verse the other. Furthermore, from my limited knowledge of GPS (I could be wrong), just because you have 10 sats doesn't mean it's using them all. Just the best ones to locate you.

Furthermore, I could just imagine that feeding the software two strings from two diff receivers in two different places will simply serve to upset the software and not perform its task.

Sorry.

rando · 06-21-2004, 01:22 PM

I doubt there would be any benefit to using multip GPS receivers. The typical positional error easily covers the length of most cars/trucks.

Regradless, if you plan to combine the two output streams, you're going to need some form of flow control. Otherewise the receiving software will not see a valid data stream at all. It might make more sense to hook the two receivers up to their own ports and then use a software layer to combine them to a virutal port. This would be quite easy to implement.

kiltjim · 06-21-2004, 06:09 PM

I agree that if you have your reciever in the front window, you should be able to get a decent signal. All you need is 4 satellites to get a 3-D fix. And depending on the software package you are using, you may only need 3. As for combining the signals, you WILL have to have flow control, unless you have 2 serial ports. Once you have the signals in the computer, you could always have a very simple program scanning the incoming data for the $GPRMC strings and combine them and then output them to a virtual port that the GPS program is looking at. As for the other strings, you could just let them pass directly through, maybe with a little editing.

Ricky327 · 06-21-2004, 06:36 PM

If you plan to use 2 GPS one placed at the front and one placed at the back of the car. As been said each one need to be connected to a serial port.

A software need to sample both the $GPRMC sentence coming from the GPSs. Other relevant sentences can also be processed. The position is then averaged out, the new NMEA sentence together with a new checksum is then outputed to a virtual port. You cant just mix them, its pointless otherwise. The software doesnt like to be bombarded with the same sentences too often either...I have tried it before. It must follow the NMEA sentences being fed into the software once every seconds.

All this in theory will give the position of the centre of the car...giving a more accurate position. The software should also know if one of the GPS output is not valid for navigation and therefore switch over to the good GPS...just incase one of the GPS lost the lock.

Can be a little bit tricky but I think it is possible. A PIC MCU should be enough to even do this if a standalone unit is the goal.

You wont be able to pick up more satellite this way, but for accuracy it should work.

Not sure what exactly can be gained as I dont think its economical but I would love to see it working though.

Ricky327 · 06-21-2004, 06:46 PM

Heres a sample of the $GPRMC :

$GPRMC,225446,A,4916.45,N,12311.12,W,000.5,054.7,1 91194,020.3,E*68

225446 Time of fix 22:54:46 UTC
A Navigation receiver warning A = Valid position, V = Warning
4916.45,N Latitude 49 deg. 16.45 min. North
12311.12,W Longitude 123 deg. 11.12 min. West
000.5 Speed over ground, Knots
054.7 Course Made Good, degrees true
191194 UTC Date of fix, 19 November 1994
020.3,E Magnetic variation, 20.3 deg. East
*68 mandatory checksum

The lattitude and longitude can be averaged out with the second GPS sentence to give accurate results. The other info doesnt need processsing...just take a good one from either of the GPS. Some software doesnt look at the checksum but its good to regenerate it anyway

kiltjim · 06-21-2004, 07:03 PM

I was thinking about this a little more, and was thinking that this may be the way that they were using the earlier, mandatorily inaccurate recievers for navigation.

And I think Ricky327 hit something. You should be able to use a simple microprocessor to combine the data. You could flip flop the device ready pin, and selectively turn on and off the recievers. AS far as that goes, you could always have the microprocessor looking at both inputs at the same time, and then average everything together (with exceptions where needed, of course.)

Ricky327 · 06-21-2004, 07:20 PM

You could flip flop the device ready pin, and selectively turn on and off the recievers.

Its best just to ignore the sentences being generated by the "bad" GPS and just spit out the good one...since the $GPRMC sentence will tell if the location is valid or not. Of course once both of the GPSs have the lock then thats when the averaging out will happen. It may even be a good idea to check which GPS have the most lock. It maybe pointless trying to average the GPS with 10 sat locks together GPS with 3 locks.

Theres probably more checking to do before averaging out the locations but its definately workable.

Pokey · 06-21-2004, 08:26 PM

Well, from what i have been told the accuracy of a GPS is around 5 meters. Which i think is great. I dont think it matters too much where on a car the GPS is located.

Some of these satellites are traveling around 20,000-30,000 mile/hr, and for at least 3 of them to figure out where I am within 15 feet 3D on earth and i can see the street I'm looking for. Thats doing good!

Just my opion.

User Name		Remember Me?
Password