Steve Douty
Well-Known Member
I am looking for an article, class, youtube post, or other learning resource that I can read, or study so that I have a better feel for the differences between PPK and RTK. Any suggestions?
RTK v. PPK
- Thread starter Steve Douty
- Start date
Nate The Surveyor
Well-Known Member
No suggestions.
I just offer a few thoughts "from the trenches of experience" not "from the exhalted pinicals of academia".
I started surveying with an old 1923 adolph leitz transit, and a 200' steel tape. I did it this way for years.
Moved up to theodolite, and edm.
Then total station.
Then data collector.
Then L1 gps.
I really learned alot with L1 gps.
Then a reflectorless leica, used with the L1 units.
Then Dual freq. Rtk.
Then Javad LS.
Then he added ppk.
The ppk he has added to rtk, there's probably not a book for. It runs at 1 second rates, and highly automates "data cleaning".
In L1, or dual freq, you can "clean" the raw data, manually. I have done a good bit of it. In a general sense, nobody else is going as far as javad has, so, at this point, I think we are in uncharted teritory, to some extent.
For obstructed places, using ppk only, (depending on how densely obstructed)
Heavily obstructed, 3 -15 min shots.
Medium obstructions 3 -10 min shots.
Lightly obstructed, 3-5 or 6 min shots.
No obstructions, 1 or 2, 5 to 7 min shots. Longer, if far from base.
This is my generally 100% safe method.
I used to do 1, 30 min ppk, but then you cannot see if something addled it for a few mins.
Of course, you will want to play with it, to determine your comfort levels, and your definition of obstruction, and if you can afford to come back later, for a second obs. Session.
For instance, i did one in c/l, medium obstructed, some traffic, (resume was great), where i did 3- 10 min obs.
2 agreed by some 0.03', and one was 0.22' away. I had enough data, to just hold the 2 that agreed.
I feel that we are "doing what they say can't be done", and therefore we are "writing the book" on how to do this.
This is strictly my opinion, not endorsed at all by anybody else.
Use my ideas, but prove them, and improve them.
Nate
I just offer a few thoughts "from the trenches of experience" not "from the exhalted pinicals of academia".
I started surveying with an old 1923 adolph leitz transit, and a 200' steel tape. I did it this way for years.
Moved up to theodolite, and edm.
Then total station.
Then data collector.
Then L1 gps.
I really learned alot with L1 gps.
Then a reflectorless leica, used with the L1 units.
Then Dual freq. Rtk.
Then Javad LS.
Then he added ppk.
The ppk he has added to rtk, there's probably not a book for. It runs at 1 second rates, and highly automates "data cleaning".
In L1, or dual freq, you can "clean" the raw data, manually. I have done a good bit of it. In a general sense, nobody else is going as far as javad has, so, at this point, I think we are in uncharted teritory, to some extent.
For obstructed places, using ppk only, (depending on how densely obstructed)
Heavily obstructed, 3 -15 min shots.
Medium obstructions 3 -10 min shots.
Lightly obstructed, 3-5 or 6 min shots.
No obstructions, 1 or 2, 5 to 7 min shots. Longer, if far from base.
This is my generally 100% safe method.
I used to do 1, 30 min ppk, but then you cannot see if something addled it for a few mins.
Of course, you will want to play with it, to determine your comfort levels, and your definition of obstruction, and if you can afford to come back later, for a second obs. Session.
For instance, i did one in c/l, medium obstructed, some traffic, (resume was great), where i did 3- 10 min obs.
2 agreed by some 0.03', and one was 0.22' away. I had enough data, to just hold the 2 that agreed.
I feel that we are "doing what they say can't be done", and therefore we are "writing the book" on how to do this.
This is strictly my opinion, not endorsed at all by anybody else.
Use my ideas, but prove them, and improve them.
Nate
Ill chime in and agree with Nate. If you are going to do PPK only you need to take multiple shots. I actually recommend multiple shots on RTK as well as PPK shots. Multiple observations increase your relative accuracy if you cluster average them afterwards.
Steve Douty
Well-Known Member
It becomes clearer as we test our equipment, and ourselves, that longer is better, up to some practical limit, and redundant shots always help both accuracy and confidence. That is the constant battle between profit and ethics. The point of this thread, however, is not how to take PPK or RTK observations. The question I am attempting to ask looks like the old test questions we always hated. Compare and contrast RTK and PPK observations, include in your answer explanations that explain the advantages, strengths, weaknesses, and methodology of both observation systems.
I think I understand what RTK means and at least have a feel for what I think is happening. I don't have the same understanding of PPK.
I sent a job to DPOS yesterday and one of the points had a fixed solution for both the RTK and PPK solution.
Epochs/s RTK = 331/110 PPK = 864/940
I am not sure how to evaluate and compare the solutions because I am not sure I have a feel for what PPK really means.
I am looking for a simple (or if necessary, complex) narrative that explains the solution process of both systems.
I think I understand what RTK means and at least have a feel for what I think is happening. I don't have the same understanding of PPK.
I sent a job to DPOS yesterday and one of the points had a fixed solution for both the RTK and PPK solution.
Epochs/s RTK = 331/110 PPK = 864/940
I am not sure how to evaluate and compare the solutions because I am not sure I have a feel for what PPK really means.
I am looking for a simple (or if necessary, complex) narrative that explains the solution process of both systems.
Last edited:
Nate The Surveyor
Well-Known Member
PPK = Post Processed Kinematic
When sitting still, it is acting as PPS = Post Processed Static
Essentially, it is a Static Observation, that is cleaned by some automated process.
I do not know why the call it "Kinematic", except, that there is probably something up somebody's sleeve, to use this while in motion, as well as stopped.
Epochs RTK 331/110 PPK = 864/940.
The PPK does it once per second.
The RTK CAN do it 5x per second, but it can also do it with little or no radio, which may let only 100 Epochs, via rtk, and have 1000 epochs with PPK.
Which one to hold to, and why.... I don't know. I have seen RTK be better than PPK, because I was using a Total Station Coord, to compare against.
I cannot definitively answer your good question. I don't know.
But, I am exploring like crazy, with this fine gear, to narrow down a few hundredths. Ain't it fun?
I have not had this much fun, since me and my wife's honeymoon! (Grin)
N
When sitting still, it is acting as PPS = Post Processed Static
Essentially, it is a Static Observation, that is cleaned by some automated process.
I do not know why the call it "Kinematic", except, that there is probably something up somebody's sleeve, to use this while in motion, as well as stopped.
Epochs RTK 331/110 PPK = 864/940.
The PPK does it once per second.
The RTK CAN do it 5x per second, but it can also do it with little or no radio, which may let only 100 Epochs, via rtk, and have 1000 epochs with PPK.
Which one to hold to, and why.... I don't know. I have seen RTK be better than PPK, because I was using a Total Station Coord, to compare against.
I cannot definitively answer your good question. I don't know.
But, I am exploring like crazy, with this fine gear, to narrow down a few hundredths. Ain't it fun?
I have not had this much fun, since me and my wife's honeymoon! (Grin)
N
One difference is that RTK looks at the signals coming in to get an instantaneous solution. PPK looks at all of the data forward and back to try and mitigate the multipath.
Duane Frymire
Active Member
My understanding (which might not be correct) is that in both situations (using dual frequency GPS+GLO) the error we are concerned with stems from multipath and in band interference. RTK is using certain algorithms designed by the manufacturer and so is PPK, but they are different methods. RTK is solving integer ambiguity based on a subset of data it has found to be most reliable over a short period of time. While PPK is using all the data to do the same thing in a different way. Theoretically, PPK should catch and reject a spike in DOP or cycle slip during the session that an RTK solution might not catch. Longer time on point with either method samples more geometry and should improve results. But ultimately which solution is better has to do with how good the RTK v. Post processed mathematical procedures are. One may give better results under certain conditions than another. You can look at the pdop, time on point, overall rms, of the RTK v. the PPK, but those don't necessarily indicate the "best" solution.
So the value really is in being able to see if the two are comparable, not necessarily how to contrast one with the other (the software developers could explain the contrast, but I doubt I would understand it). If they don't agree within a set error budget for your project, then you know there is too much multipath and you should try a static session or use a total station to locate that point. If all you were using were RTK you would only be using one solution method or algorithm, no matter how many times you shoot the point. You would get different satellite geometry, but still only the one method. So, redundancy is always good. In addition, if you can't get a rtk solution you can let it cook and get the ppk. Quality indicators are less than 3cm overall rms, and dop of less than 2 (some sources say 3 or 4).