RTK Precision

Share:

# RTK Precision

Page 1 / 4
Posts: 432
Member
(@field-dog)
250+ posts
Joined: 5 years ago

We set a pair of points today using RTK. Each point was shot twice consecutively for three minutes. To begin with, the residuals were terrible. Something like 0.04' H and 0.09' V. The satellite geometry was good. I tweaked the connection using different combinations of single baseline and MRS. I also switched from primary network to secondary and back again. No significant improvement in the residuals. I was told to average the coordinates of each point. I noticed the second shot on each point was off mostly to the north. Why?

Topic Tags
25 Replies
Posts: 5750
Member
(@paul-in-pa)
5,000+ posts
Joined: 9 years ago

One satellite difference can be all it takes to get a large N-S or E-W difference. If you want precision you do at least 3 shots with different satellite geometry. No matter what some people say a pair of points is useless for precision. It is much, much better to set 3 points that at intervisible and can be confirmed with a set of total station observations. Your residuals were a valid prediction of what you were going to end up with. Time for a do over.

Paul in PA

2 Replies
Member
(@field-dog)
Joined: 5 years ago

250+ posts
Posts: 432

@ Paul in PA

Thanks for replying! We have another day left on the job. It's a boundary survey of a single family home in a suburban subdivision. The pair of points are situated so that half the found property corners can be observed from one point and half from the other point. If we do a third RTK session in the afternoon after observing the property corners in the morning, can the office re-calculate the property corners? I assume they can.

Member
(@paul-in-pa)
Joined: 9 years ago

5,000+ posts
Posts: 5750

You had 2 3 minute observations, what and where was your base? Can you use your base for occupying one of them? Since you say these two points are for total station observations are they intervisible? 2 15 minute observations could get you independent OPUS-RS solutions. If you do not have RTK units capable of logging field data for post processing there is little you can do later to recover from poor observations.

As to your question; "I have no idea how long the vectors were. How do I find out?" I suggest thorough education.

Your residuals suggest you may have had too few satellites visible, the only solution to too few satellites is a lot more time. You have to have enough understanding of your results in the field to know when it is necessary to invest more time on the job. GPS works in the dark and I have often stayed well past Dark:thirty so that I was not having to make a return trip.

Paul in PA

Posts: 6005
Member
(@jim-frame)
5,000+ posts
Joined: 9 years ago

Why?

Because RTK.

How long were the vectors to the network stations?

Member
(@field-dog)
Joined: 5 years ago

250+ posts
Posts: 432

@ Jim Frame

Thanks for replying! I have no idea how long the vectors were. How do I find out?

Posts: 6095
Member
(@mightymoe)
5,000+ posts
Joined: 9 years ago

That's not a good result, .04 x .09 translstes to .12 x .3.  I would say you don't have control points yet and you need to resolve it, no one can tell you why with the info provided, call the network provider or do a base rover to get it tight.

Member
(@field-dog)
Joined: 5 years ago

250+ posts
Posts: 432

@ MightyMoe

Thanks for replying! You think it might be a network problem? I normally see residuals tighten up over a 3 minute session. The residuals stayed the same.

Posts: 4468
Member
(@norman-oklahoma)
2,500+ posts
Joined: 4 years ago
Posted by: Field Dog

We set a pair of points today using RTK. Each point was shot twice consecutively for three minutes. To begin with, the residuals were terrible. Something like 0.04' H and 0.09' V. The satellite geometry was good. I tweaked the connection using different combinations of single baseline and MRS. I also switched from primary network to secondary and back again. No significant improvement in the residuals. I was told to average the coordinates of each point. I noticed the second shot on each point was off mostly to the north. Why?

Why? In a word, multipath. You haven't told us anything about the site conditions, PDOP, etc. You have measured the points in various ways which seems to prove that there is nothing wrong with the network or your equipment. The thing that is left is the observing conditions.

So, you averaged multiple shots. That tactic is much more effective when the shots are spaced hours apart in time, so that the satellite geometry is different.  You are still going to get huge residuals if the multipath conditions are bad, but the average (might be?) should be closer to true. No guarantees, though.

Page 1 / 4