Old rtpk vs. New rtpk

Alexey Razumovsky

Well-Known Member
Staff member
5PLS
Alexey,

I have attached a link to a project that has point numbers 106, 107 and 108. These points were collected in WAAS Float mode. Point 108 got 1 epoch of a float solution. All 3 points were collected for 10 minutes each. RTPK would not process any of these points after stopping the base and downloading the base file. All 3 points were on the same iron pin. Just an FWI we got 3 good RTK shots on this pin with the VRS in about 5 minutes total or so. I would expect RTPK to give a good result in 10 minutes if RTK can give results like that with the VRS network. Can you take a look at these points and let me know.


Thank you!
No overlapped data for base and rovers
 

nusouthsc

Member
Darren, next time when you get "no solution for the vector" in 15 minutes send a project to support.
We are doing a lot of test in different locations but Ive never seen this for 15 minutes session. I believe it will be helpful. RTPK is not so mature as RTK is. We might to find best approach together.
My point for challenge locations is 2 minutes data split interval for LS+, 6 minutes for standard LS. Am I correct?
@Alexey Razumovsky I have been doing some testing today in an extreme environment and was able to produce a 15 min session that didn't compute a vector. I hope that I stored it correctly that it can be useful. Please let me know if I need to get you anything else. Point 2.3 is the one that didn't work. Link below:

 

nusouthsc

Member
I am a bit confused on the Minimum RTPK Verification Level setting. I was under the impression that was counting how many RTPK fixes were in agreement. I have been trying out the auto-accept features and I cant tell a difference if I am using 1 or 15 for the setting. As long as my epoch count, and time settings are met it will accept with one RTPK Fix every time. Apparently I am missing something. I appreciate any guidance.
 

Adam

Well-Known Member
5PLS
That is how it should be. If variety is required and met and rtpk has 1 fix, it's good. The rtpk min verification level really comes into play when you aren't getting any reliable rtk. You can call me Josh
 

Matthew D. Sibole

Well-Known Member
5PLS
@Alexey Razumovsky

Alexey I just sent a project to support. Bill Osborne 01037 serial number. Multiple (5) RTPK shots with 0.8’ between them but the pole never moved.

212C6401-E773-45A0-B8CA-3A581815DA00.jpeg
 

Matt Johnson

Well-Known Member
5PLS
@Alexey Razumovsky

Alexey I just sent a project to support. Bill Osborne 01037 serial number. Multiple (5) RTPK shots with 0.8’ between them but the pole never moved.

View attachment 11669

What was interval period? The receiver looks like it is right beside a home with aluminum siding which is very reflective. In locations like this I recommend multiple observations with at least 1 hour between them.
 

Matthew D. Sibole

Well-Known Member
5PLS
I ran the RTPK processing at about every 60 seconds. However, the ones that had 5 RTPK shot agreement took almost 10 minutes to get that level of repeatability. I had 2 shots like that. I had a few RTPK observations that agreed with the RTK shots as well.
 

Matt Johnson

Well-Known Member
5PLS
I ran the RTPK processing at about every 60 seconds. However, the ones that had 5 RTPK shot agreement took almost 10 minutes to get that level of repeatability. I had 2 shots like that. I had a few RTPK observations that agreed with the RTK shots as well.

What was the data split period as defined in RTPK Environmental Category screen though? Manually starting the process at different times does not determine when the data is split.

I recommend that there be at least 6 minutes of time separation between the first and last RTPK session that agree. To get 6 minutes in between sessions that agree divide 6 minutes by the session interval and add 2 to get the required Min RTPK Verification Level. For example if you have the session interval set to 60 seconds this means that you need 8 sessions that agree.

Near buildings with large reflective surfaces (metal and glass) I have found that this will provide a good solution about 99% of the time. To guarantee a good solution I think you need to survey the point at least 1 hour later to check it. With the test LS near the factory in San Jose I have seen rare periods where RTPK will return the same bad solution for periods of about 30 minutes. I have also found that it is possible for RTK to agree with a bad RTPK solution. This shows that the issue is from the multipath and not with the way the data is processed.
 

Matthew D. Sibole

Well-Known Member
5PLS
I had it set at the 2 min split.

I had not worked with the standard LS with multi-constellation 6 engines in such a long time. I was trying to figure out a good procedure and settings for success with limited time. I still do not have the procedure down for it and will take additional time.
 

ken larson

Member
What was the data split period as defined in RTPK Environmental Category screen though? Manually starting the process at different times does not determine when the data is split.

I recommend that there be at least 6 minutes of time separation between the first and last RTPK session that agree. To get 6 minutes in between sessions that agree divide 6 minutes by the session interval and add 2 to get the required Min RTPK Verification Level. For example if you have the session interval set to 60 seconds this means that you need 8 sessions that agree.

Near buildings with large reflective surfaces (metal and glass) I have found that this will provide a good solution about 99% of the time. To guarantee a good solution I think you need to survey the point at least 1 hour later to check it. With the test LS near the factory in San Jose I have seen rare periods where RTPK will return the same bad solution for periods of about 30 minutes. I have also found that it is possible for RTK to agree with a bad RTPK solution. This shows that the issue is from the multipath and not with the way the data is processed.
I am using the 6 min rule above for my ls+ and t3...is this still the suggested rule of thumb?
 

Alexey Razumovsky

Well-Known Member
Staff member
5PLS
I had it set at the 2 min split.

I had not worked with the standard LS with multi-constellation 6 engines in such a long time. I was trying to figure out a good procedure and settings for success with limited time. I still do not have the procedure down for it and will take additional time.
pp.106-108 deals 3 files (3:31; 10:32; 5:25). Total time at point 19 minutes 28 seconds. You processed files separetely with split parameter = 120.
I merged files and got 1168 seconds file. I processed the file with different parameters.
1627885251137.png

All RTPK solutions are within 0.013 mm in hor.
1627885351140.png

I still suggest to avoid multiple occupation on a point. Give to RTPK enough data (big file). It can process and adjust correctly.
In the meantime one is able to reRTPK file with updated splittiong interval and compare the results.
1627885725502.png

Repeated occupations would be useful with big time separation (~1 hours)
 
Last edited:

Bill Osborne

New Member
I am learning the new RTPK process with the new upgrade Yesterday I was testing it in a pretty tough spot down a hollow with a lot of canopy. My Standard LS with multi constellations went the first 12 minutes without a RTPK vector fix and then on the next time split, which are set at 2 min, it returned a fix with (5) which I though means that 5 of the 7 (2 min splits) matched within the tolerances. If that's how it works, why wasn't I seeing the other fixes during the first 14 minutes? If you have 5 RTPK splits that match, it seems based on old RTPK recommendations I've read on here that we can have confidence in the shot without taking the time get another shot. Am I looking at that right?
 

Michael Green

New Member
Please allow an rtpk solution. if I store the rtk now I have no rtpk to compare last or future shots with. White the previous version there was always a rtpk solution to evaluate. So 20 minutes into this observation I only have 7 epoch to use
1630338560117591055561459628121.jpg
 
Matt Johnson: may I suggest that we have an updated monthly video webinar on this topic? It seems very complicated and that there have been changes since the program that Shawn Billings gave.
 
Top