I am going to be doing a project where I have 2 semi-permanent GPS base stations setup. The first base will be collecting data and transmitting RTK corrections over UHF. The second base will only be collecting data. On the rover end I will be surveying in control points while collecting data (kinematic mode) for 3-5 minutes on each point. At the end of the day I will have a RTK vector and static vector from base1 and a static vector from base2. Is it possible to combine all the vectors in Starnet to get a least square solution for each point? It has been a while since I used Starnet, I have V8 pro.
Yes, you could do that in StarNet.
But, from Base 1, you should use either the static vector or the RTK vector, but not both. Using the same data twice, which is what using the same satellite data to compute a vector 2 ways would be, won't gain you anything. The second vector from Base 2 will be only a limited benefit.
Thanks Norman. The second base is there as a check. Would it make any difference if I used the static vectors from each base for the adjusted solution or RTK from base1 and static from base2?
The static vectors are likely to be of slightly better quality, particularly if the rover locations are compromised and you engage in vector refinement techniques like excluding some satellites with marginal data in post-processing. If you aren't going to do that there won't be much difference in quality. Which is not to say that the x,y,z components will be identical, but rather just that one is as likely to differ from the theoretical perfect as much as the other.
99% of the time I roll with the RTK vectors because its easier and they are going to be good enough that other concerns, such as centering errors, have greater effect on results.
The second base as a check is checking a receiver above a point on the ground. It checks the receiver, but not the point. To do that will require redundant measurements. A static measurement using a tripod and then later an RTK measurement using a rod would be a great check. But, two static measurements and an RTK measurement with a single set-up is of limited value. The redundancy is critical to control. Same with the semi-permeant bases if they have fixed points the receiver is setting over.
Almost all the "error" in modern Static surveying is bubbles, measure ups and recording errors. Doing it twice with different equipment eliminates almost all of those errors. Using a second location won't improve your error ellipsoids, but it will make them valid.
I can see what you're getting at Moe, but I somewhat disagree regarding redundancy in a single observation with 2 bases- mostly due to me being a pedant. I am a BIG fan of having an RTK vector streaming from one base, then a postprocessed vector streaming simultaneously from another. It's a redundant form of position solution, which is not only a major gut check in my book, but simply another redundant measurement. If they're both static... I'll process full constellation from one base and turn off a constellation (usually GLOSNASS) for the other base. We can argue semantics about independent v. dependent, but when you add different vectors that are fundamentally different due to different processing types or parameters... you're bound to come out with a stronger network in the end.
In the same vein, I would NOT recommend merging all vectors types (static1, RTK1, static2) in a unified adjustment- by having both vector types from what I'll call "base 1" in the adjustment you're incorrectly skewing the adjustment. Use a combo of RTK from base 1 and static from base 2, or static from base 1 and base 2 if you're running 2 bases simultaneously.
edit: HARD agree on most error being set-up/measure-up errors nowadays... There's so many birds in the sky now, and RTK engines are getting better and better.
edit2: this all assumes the user is using an RTK-logging style or non-trimble equivalent.
