I have processed (a lot) of airborne GPS data, but have not done any PP Kinematic on the ground in the data collector.?ÿ
I recently had a client send me some data that was collected in PPK mode. It was stop and go, they were occupying photo control as observed control. Worked well. So I looked into it and saw that on the data collector it wants a certain amount of time to initialize, the defaults basically same as fast static (8 minutes for 6 or more SV's).?ÿ
But, if doing a real time kinematic survey the initialization time is typically a few seconds. So I wonder what is the difference? If both the rover and the base are collecting at 1s epochs, why does it take so long to assure a successful init? I realize that in real time the receiver is continuously tracking the signal, and storing a reading every epoch, but if there are no cycle slips why can't the post processed data init as quickly??ÿ
I would be interested to hear from anyone who does this style regularly to see what the real init times are.?ÿ
I can't say what the real time is going to be. I have mine set at 3 minutes, the 8 minutes is a default, it's a worst case setting and it was?ÿthat way?ÿ20 years ago, so it seems to be needing an update.
If you start your PPK session and continue it for?ÿ3 minutes and it initializes then all the points collected before the initialization will also be good. At least that's my experience. That can be problematic when you loose lock during those 3 minutes. Then the points between start-up and?ÿinitialization will not be good.
I've tried it at 30 seconds and it worked well, I haven't pushed it more than that. Really it's mostly about what you are comfortable with, your experience, site conditions, ect. I would suggest some trial and error if you really want to find out your limits with PPK, clearly an R10 will preform better than a 4000 series.
?ÿ
?ÿ
My use would be in an area with no cell coverage. In locations with no VRS we setup our own base and run an NTRIP caster in the office. We then survey discrete points (photo/Lidar ground control points). So it is typically just pull up to a location, then initialize and survey the point as Observed Control a couple of times with the corrections coming in over cell. When there is no cell, we fall back to doing a static occupation for 15 minutes. PPK would be an alternative for locations with no cell service. So I don't usually have any time before getting to the point for initialization. If I have to wait 8 minutes I might as well do static as I have been doing for many years.?ÿ
?ÿ
I have always wondered why real time can init in seconds but yet post processed static needs minutes.?ÿ?ÿ
I'm anxious to listen in on this too.?ÿ Were always reverting to static due to the 8 minute "default" setting.?ÿ Like to better understand this as well.?ÿ Fewer and fewer people I guess are losing lock, or out of cell these days.?ÿ Very common here.
?ÿ
Joel
There is a setting in survey styles, I've turned down the 8 minute setting to 3, you can turn it down to whatever you would like. I have tried it at 30 seconds a few times and it worked, but there are no guarantees the 8 minutes is probably fool-proof considering clear skies and the regular issues for good GPS collection.
I can't stand waiting 8 minutes, my work flow isn't impacted with 3 cause I'm usually collecting lath, pulling out tools, marking a cap, maybe searching for a point. I've never had a session fail using 3 minutes.
I probably wouldn't use 30 seconds, you just don't know if you are really fixed till you download the PPK session. I never collect them in the receiver, I have had problems with doing that trying to mesh with the base receiver in TBC.
John is correct with his thinking, there is no reason why the PPK can't fix as fast as RTK, they are using the same info. However, we have all stood in the field waiting on a fix that is taking longer, the setting in PPK is set that way to avoid those sessions. 8 minutes is very conservative.
The RTK initialization time is a function of number of satellites, distance w.r.t the base station, and the local conditions (for example, afternoon strong ionospheric condition or not, high multipath or not). In the real time system, the GNSS receiver RTK engine can give the best (possible correct with confidence 99.9%) positioning solution based on some complicate algorithms using its own tracking satellites and the satellites from the base station. For a modern GNSS receivers tracking at least GPS + GLO, the initialization time normally varies from a few seconds for a short baselines (normally < 5 km) to a few minutes for long baselines (10-30 km). For the PPK (Post-Processing Kinematic) case, the rover receiver does not have any information about the base station, so it is normally to set a relatively longer period to log the data in case that you have a base station far away from the job site.?ÿ