GNSS Network Planning

How many receivers?

Many least squares packages include the ability to run simulations. You can explore different observing scenarios and obtain estimates of results.

Reviewing the literature would also be useful.

https://www.degruyter.com/view/j/jogs.2014.4.issue-1/jogs-2014-0005/jogs-2014-0005.xml

https://www.amazon.com/Geodetic-Network-Analysis-Optimal-Design/dp/1575040441

?ÿ

?ÿ

Having the opportunity to use 3 receivers and place them in good earth geometry with one another will allow you to obtain best results and not worry about correcting after multiple occupations in an open sky.

Related to the issue is Robustness Analysis described here:?ÿ http://www2.unb.ca/gge/Pubs/TR156.pdf

?ÿ

3 control sites? Are they within or surrounding the project?

GNSS implies you are using multi-constellation Receivers? Again the question , how many?

Will his be RTK, short or long static? Will you use OPUS or OPUS-RS?

Are you intending to include elevations as an important part of this project?

With 39 points in your network the possible vectors run into the many thousands. You would be better off to just analysis the actual few simultaneous network points (quality) you would get with the number of receivers you have available.

In my opinion, the best software is in your head full of years experience.

Considering the small areal size of your project do not ignore the use of GPS or GPS/Glonass L1 only receivers in the mix. You can easily let a few of them run for longer periods as you jump around other points with the GNSS receivers.?ÿ

Let me make an assumption of what you may have available in an office with long GNSS? experience, 2 GNSS receivers, 4 GPS receivers in the form of 2 RTK setups, and possibly some L1 (GPS or GPS/Glonass) recievers. You can use them all in the mix, but again the how depends on how many and what you have.

I have done a good number of network GPS projects using 3 L1/L2 receivers (OPUS) and 7 L1 receivers, for a total of 10-15 points an one observation period. Typically with 3 men and 3 vehicles. Generally a pair of L1s would be set up within site of an L1/L2 receivers for quick traverse check before heading off to?ÿsurvey point with less favorable sky.?ÿI have personally setup?ÿ3 L1/L2 and 4 L1 receivers at a time on an eventually 14 point network, which I just added 3 points to last week. I have also found out that filling a network with short static that I am unable to setup, run and move 5 receivers fast enough to stay at a minimum observation time. Today I would not exceed 2 L1/L2 receivers and 3 L1 in the filed at one time. Because I am blessed by being in an excellent area for OPUS-RS?ÿfor thw work I do I would no consider having to upgrade to GNSS receivers. I would do it because I want to and not that I need to.?ÿ

I have been very successful in combining GPS OPUS, OPUS-RS long and short term L1/L2 and L1 static in my post-processing solutions. Understanding that you could be combing OPUS and OPUS-RS, GPS and GNSS long and short static and RTK into one network solution you should be very successful.

The only real planning you need to do is in planning the different times for the Geometry of the Constellations to change. You have a multi day project and you should be operating in long days and geometry of the constellations will change enough not to worry about filling in any holes until near the end of your planned observations.

Paul in PA

The most important thing is to correctly and precisely record your antenna heights. An incorrect height on a GNSS point that is observed points can warp a network and create unbelievable horizontal positions variances. That lesson I learned the hard way on my first big network.

I have (3) NGS monuments on site, and the other 36 monuments are existing monuments. The goal of this project is to create NAD83 positions on these points from an old control system that has been in place on this site for a long time. As the facility moves into the GIS era, the need for the NAD83 values are becoming a pressing issue.

My plan is to run digital levels to these monuments for elevation values.

I have (2) OPUS X90 receivers, and (4) Topcon Hiper V units that I will use to run this network. With a total of (6) receivers, I should be able to create a fairly strong network with redundancy.

Thanks for the input and advice.

EDIT: I plan on post processing these occupations myself. My plan is to set the two OPUS X90 units on the NGS control points, and use the Topcon Hipers to move around the network using 30 minute occupations. This is a controlled access site, so I only have two vehicles that I can access the site with, so I'll need to plan the occupations so that I can maximize the ties between points. Many of the points are inter-visible, and many are not, so I'll definitely want good vectors between the ones you can occupy and backsight.

Jimmy

Are all 6 GNSS enabled? For this project I would consider GPS only as sufficient.

I would begin by occupying the NGS and 3 other monuments around the project perimeter for 3 sessions, making sure I had the X90s on each point once. that becomes your primary network with 15 vectors. For the next round occupy any 3 of the 6 as bases and start occupying the other 33 points and you will be very busy. Redo the next day occupying the other 3 bases. You are now done occupying those 6 bases. Pick the best 3 of the other 33 for bases for the third occupation on the remaining 30. To simplify you may want to hold the first 6 bases fixed in the second step and exclude the subsequent?ÿbase to base vectors in the following steps. When you are satisfied with your field work, include all vectors and see if you crash your PP program?

I over exaggerated brfore, using my cacluator your 36 points would have a maximum of 630 vectors.

Paul in PA

Although all of the large networks projects I have planned or been directly involved with (up to 330 stations using 10 receivers) have used overlapping observation "clusters" in accordance with NGS-58 and NGS-59, cases have been made for what I call the "hub and spoke" method. An excellent resource can be found here:

http://csrc.ucsd.edu/docs/CLSA_CSRC_GNSS_Standards_and_Specifications_v1.1.pdf

For the large networks I have planned, the requirement was to get a minimum of 30 minutes of quality data between all 10 observers for each observation session. We eventually planned these sessions for 60 minutes to ensure the 30 minute minimum. Invariably we would have delays with observers getting set up and started late (from getting stuck at train crossings, to making a wrong turn, to you name it). Once the observation schedule, with adequate move times between sessions, was finalized, there was no changing it "on the fly" because of a delay in the field. Doing so created a domino effect for the remainder of the day. Once we were done with the planned sessions, we usually had between 5-8% of the baselines that required reobservation, whether from being out of specification, or having been missed entirely for whatever reason, like a flat tire.

Because you are not going to be establishing the vertical with the GNSS observations, you may want to consider the "hub and spoke" approach, which gives your observers much more latitude from not all having to occupy at the same time to get the "loop closures".?ÿ

Good luck. Sounds like a fun project.