Pegon
Posts: 1
Joined: Wed Jul 18, 2018 4:50 pm

GPS recommendations

Wed Jul 18, 2018 4:58 pm

Hello all,
I am fairly new to the raspberry Pi world and i would like to buy a GPS chip set which tracks my raspberry Pi.
The GPS has to be pretty precise (<1 m) if possible put the speed at which it measures has no requirements ( could be slower than once per second). Also obviously i would prefer a reasonable price.

Thanks in advance

W. H. Heydt
Posts: 8655
Joined: Fri Mar 09, 2012 7:36 pm
Location: Vallejo, CA (US)

Re: GPS recommendations

Wed Jul 18, 2018 6:17 pm

To get 1 m accuracy (and least in less than hours to days) probably requires military grade equipment, which would be both expensive and--most likely--have restrictions placed on purchase.

What I would do is to settle on what you can find within your budget and then replace it in a few years with something that will receive US (GPS), Russian (GLONASS), and EU (Galileo/GNSS, not yet in service) global positioning signals. The specified accuracy for GPS is 5m. GNSS is supposed to have an unencrypted accuracy of 1m. Using two or three systems together *ought* to enhance accuracy.

mosespi
Posts: 499
Joined: Mon May 12, 2014 3:35 pm
Location: 34,-118
Contact: Website

Re: GPS recommendations

Wed Jul 18, 2018 8:02 pm

The easy way for the Pi is to buy a USB GPS 'puck' like this one..

http://www.allspectrum.com/store/gps-na ... -9136.html

I've used that particular one with good results. There are plenty of others available, especially with just a plain serial interface sold for use with drones. A search for "GPS module" should bring up plenty.

You get close to 1m (some of the time at least) with the more modern, double-tracking GNSS chipsets like a ublox 8. They do costs a bit more.

Regards,
-Moses
Power problems? MoPower UPS for the Pi
http://www.allspectrum.com/mopower/

il_diavolo
Posts: 123
Joined: Mon Dec 02, 2013 7:56 pm

Re: GPS recommendations

Wed Jul 18, 2018 9:35 pm

Has anybody tried to make a local differential GPS system using two Pi's?

It would need both Pi's to be fitted with similar GPS modules and both be on the same network so that they can communicate with each other. One of the Pi's (static Pi) will be in a fixed position and readings taken over a length time (minutes, hours?) until a good average position can be established which should be accurate to < 1m.

Having established its position the static Pi can now compare that accurate position with the current "live" position it is receiving and generate a "difference" error. This error is then passed to the mobile Pi and used to correct the "live" position it is receiving. As long as the two Pi's are in the same general locality then the mobile Pi's position will also be accurate to less than 1 metre. The initial averaging should only need to be carried out once as long as the static Pi is always placed in the same location.

I believe that land surveying instruments may use a similar system (I could be wrong!). It was certainly used for marine navigation years ago when selective availability was turned on and accuracy could be tens or even hundreds of metres out.

If anyone does take this further I would be grateful for copies of your code :D

KanoMaster22
Posts: 168
Joined: Wed Oct 05, 2016 6:06 pm

Re: GPS recommendations

Thu Jul 19, 2018 12:00 am

I am not of much help but there was recently an article in the Magpi about local differential GPS.

Edit - I found the project I am talking about (https://projects-raspberry.com/raspberr ... ntial-gps/)

W. H. Heydt
Posts: 8655
Joined: Fri Mar 09, 2012 7:36 pm
Location: Vallejo, CA (US)

Re: GPS recommendations

Thu Jul 19, 2018 12:33 am

il_diavolo wrote:
Wed Jul 18, 2018 9:35 pm
Has anybody tried to make a local differential GPS system using two Pi's?

It would need both Pi's to be fitted with similar GPS modules and both be on the same network so that they can communicate with each other. One of the Pi's (static Pi) will be in a fixed position and readings taken over a length time (minutes, hours?) until a good average position can be established which should be accurate to < 1m.

Having established its position the static Pi can now compare that accurate position with the current "live" position it is receiving and generate a "difference" error. This error is then passed to the mobile Pi and used to correct the "live" position it is receiving. As long as the two Pi's are in the same general locality then the mobile Pi's position will also be accurate to less than 1 metre. The initial averaging should only need to be carried out once as long as the static Pi is always placed in the same location.

I believe that land surveying instruments may use a similar system (I could be wrong!). It was certainly used for marine navigation years ago when selective availability was turned on and accuracy could be tens or even hundreds of metres out.

If anyone does take this further I would be grateful for copies of your code :D
Or find out what the requirements are to use encrypted GNSS. It's supposed to be accurate to 1cm.

Return to “General discussion”

Who is online

Users browsing this forum: Baidu [Spider], hommar, mmaharebi and 51 guests