Discussion on RTK GPS Systems

Keep in mind that the base unit must still be reference to a know point ie a surveyed reference point to achieve geo-accuracy.

I’ve played with the original Pikis which referenced to its self was just ok (inches of error), but referenced to my survey test point was bad (feet of error) in static test.

We have some pro surveyors here that may straighten me out, I’m just a hack, so don’t take me seriously.

At closer examination the Piksi Multi is only using GPS L1/L2 and slightly miss leading with this comment if you had just glanced over these specs and missed it that itx only  “ Hardware-ready for GLONASS G1/G2 BeiDou B1/B2 Galileo E1/E5b QZSS L1/L2 SBAS”  Meaning only GPS-L1/L2 are working. So if and when is it planned to implement to use the others isn’t indicated.  Where as Tersus has all these constellation already included. Also I see in the Tersus GUI it has the ability to smooth the base position to improve its accuracy if there are no accurate survey marker to work of for the base. So now my sentiment is leaning more towards the Tersus. Lets see what feed back I get from my inquiry to Piksi and decide from there.

Raun Benson,

If you are only using 1 GPS ( a Network Rover ), then Yes you will need an internet connection while collecting GPS data.  The Rover will connect to a NTRIP server to receive Real-Time Corrections from a nearby public Base Station.  

But as Gary pointed out, even having 2 units yourself (a Base and a Rover), you still need the accurate coordinates of the Base’s location.  Any RTK survey is dependent on the accuracy of the Base Station coordinates.   So you either must survey a LONG Static session, deferentially correct the data in your office, then return to the project site with known base station coordinates and use the Rover to locate the GCP’s with RTK and radio communication between Base-Rover… OR have cellular service and establish the Base’s location as a Network Rover.   So if you don’t have cellular service and a nearby (22 miles) Base Station, it’s going to take 2 trips to the site.  

I just use 1 unit as a Network Rover since I have decent Coverage for Public Base Stations and Cellular.   If you are in the USA, you can check your Base Station Coverage here http://with radio comms 

Having 2 units (Base and Rover) does help when your Cellular Coverage is poor at a project site.  You can establish the Base Location as a Network Rover.  Once that happens, you wont be depending on Cellular.  It’s a straight RTK survey with accurate Base coordinates and a Rover with radio comms.  

I have not tried the Tersus Board.  They mentioned sending out evaluation units for testing on another forum, but I don’t think that ever happened.  

I didn’t know about the L1/L2 Piksi.  It’s good to have some competition in the DIY crowd.  I don’t see any mention of BlueTooth though.  

If I was starting out in low-cost RTK GPS right now, I would focus on the data collector and it’s software first, probably an Android Tablet.

 

Ryan, I am in SE Asia, and so far Ive not been able to find any public base stations to use NTRIP, but I’m still looking.

Can you please expand more on this comment ? " I would focus on the data collector and it’s software first, probably an Android Tablet "

Its also what I find is missing the most with many of these low cost RTK-GPS providers is the lacking in good docs, end to end suitable software and the explanation of work flow to collect and process the field data. Mostly their docs focus on setup, and using the GUI/interface. But then nearly nothing for best practices in how to carry out the field work for the data collection and then the work flow how to process that data into real usable coordinates.

John,

The GPS (GNSS is the correct term, sorry I’m in the USA)  unit outputs a calculated position.  

This is done using any of the NMEA 0183,  RTCM, or NTRIP protocols.  Or the Receiver can store the Raw Data in RINEX format for post processing but that’s not Real-Time. 

You still need to connect to the GNSS unit with a data collector to store and average the real-time corrected positions, setup your parameters, etc.   That’s the same for the DIY boards (Tersus, Piksi, etc) as it is with Survey Grade expensive receivers from Topcon or Trimble.  It’s just easy to forget about the data collector when we are focused on selecting a DIY GNSS receiver.  

I think the Lefubure NTRIP client for Android is a good starting point, if you have Public Base Stations.

https://play.google.com/store/apps/details?id=com.lefebure.ntripclient&hl=en 

You will still need software for a tablet/phone to receive the positions from the GNSS unit (normally over BlueTooth), store them, and output the positions to a text file to use with Pix4D.  

I have no idea how people survey in Asia.  We can always fall back on OPUS in the USA, with a static session if all else fails.

Surely you have Public and Private RTK Networks.   Hopefully someone from Asia will chime in.  

 

This may not help the conversation, but you can normally find older GNSS receivers on Ebay that will work fine, w/ data collector and software.  

Interesting discussion here.

I am using Emlid Reach so far and my goal is to have it onboard the UAV. I do not go for RTK, PPK is my prefered option.

For me (I am in Africa, no Cors or even permanent station near by) the base must be dual frequency. I am using a friend Trimble 5700 for the moment but I need to have my own solution. Budget is really an issue as the activity is slowly starting, so best option now seems to be the Tersus.

I’ll add a little more info for those readers that may not be familiar with the concepts.

PPK is Post Processed Kinematic (OPUS in my original post in this thread)

RTK is Real Time Kinematic

Both are similar.  PPK has the benefit of using forward and reverse solutions for each instance in time, since it’s processed after the survey is complete.  Obviously RTK doesn’t know the “future” satellite data during the real-time calculations, so no forward solutions.  

RTK has the benefit of giving the user a good idea of the precision (and accuracy w/ proper workflow) during the field work, to reduce the chance of having to return to that project site for more field survey.  

But for this discussion (highly accurate GCP’s) both RTK and PPK require accurate Base Station Coordinates, if it’s your local Base or a continuously operated remote station.  

Many users wont require the accuracy that we shoot for when performing Civil Engineering Surveys, so most of this discussion may be overkill.   For instance, if I publish the elevation of a Man Hole or Drop Inlet on a set of plans, I need that elevation to be repeatable in future surveys and during construction.  That demands accuracy, not just precision.  

The most important thing to understand is that both PPK and RTK solutions can only (at best) achieve the accuracy of the coordinates of the Base Station.  If your Base Station has 3 cm of error, then all of your shots will contain this error, in addition to the random errors of the Rover and collection procedure.   

I chose RTN (Real Time Network) which is a RTK to survey GCP’s for my Pix4D projects.  That only requires 1 receiver and a cellular connection. The correction source is automatically chosen for the closest Networked Base Station.  Each Continuously Operated Reference Station has coordinates that were previously established by long term PPK surveys.  RTN removes half the problems for me.  

If you plan to have 2 receivers and setup a Local Base Station at each project site, you still need a way to establish the coordinates of the Local Base.  This still requires a correction source (another previously surveyed Base Station) for both RTK or PPK , or unrealistic occupation times for Aerial Survey Projects. That’s why RTN was the best choice for me.  

If I don’t have a Continuously Operated Reference Station within 22 miles of a project site, then I will PPK a monument by several long static sessions and then return at a later date to operate a local Base Station at the project site.  

 

 

Hi Ryan,

 

I can not thank you enough for all your assistance and guidance and gave you me so much information. I have done some more research and I have decided that the RTN solution seems really good for me getting GCP coordinates.

I will purchase the Precis BX305 and a bluetooth module. What is your opinion on a cheaper GNSS antenna’s and is there a specific cheaper antenna that you recommend? I will use the NTRIP Lefebure client on my android phone. Once I am connect and have everything up and running, with the NTRIP client, can a view the UTM coordinates on the andriod device so that I can record the reading for my GCP’s or do I need something else to view the coordinates or do some conversion?

I am hoping to have everything up and running in mid Jan. I will post some pictures soon.

I am excited to see what the GCP’s will do to my 3D modelling accuracy.

Your very welcome Ruan Benson,

Tersus sells an antenna for $160.  But you may want to contact them about splitting the 2-unit kit, so you can get the enclosure for the board.  Then add the $30 Bluetooth module.  

I tried Lefubure NTRIP client on my $100 Acer 840 tablet and it worked fine.  It did require me to turn-on Debugging in the Android settings.  

Lefubure NTRIP does spoof the RTN corrected positions (in Lat/Long) so the Android device can use the corrected positions just as it would it’s internal GPS device.  Now you need an app or method to average the positions for each GCP for 1-2 minutes.  And I always shoot each GCP before and after the flight as a check, with as much time between the shots as possible. 

Please let us know your results.

 

http://eng.topodrone.ru/reviews/blog-9-2/