I am a student and very new/unfamiliar with RTK and surveying methods.
I have collected surveys of stream-channel cross-sections using the Emlid Reach RS2 GNSS receiver acting as a rover with an NTRIP connection accessed via hotspot. I exported the data as a CSV and plotted the data as a cross-section profile using a GIS workflow to compare to historical surveys.
I have some confusion upon looking at the data. Due to vegetation, several points were collected in FLOAT mode, since it appeared that the accuracy was good enough for the purposes of the survey. Upon looking at the data, many of these float points appear to make logical sense although the Elevation RMS may be in the range of 0.5 - 1.0. Some of these points however, jump 10 or more feet randomly and are definitely not representative of what I observed, however the RMS error is still less than one foot. This very confusing to me and I am also wondering what other ways to quantify error (PDOP?)
My main questions are:
What does RMS represent (for dummies)? I was under the impression it was showing the error, but I no longer believe this. Can FLOAT points be relied on at all? Could this be an NTRIP connection error (cell service was sometimes not ideal)? How come some of the FLOAT points appear reliable and some not, and how can I actually quantify the error associated with those points to determine what data is useable?
Attached is a screen shot of the raw data files. I apologize for all the questions and if folks can point me to additional readings or resources to better understand this method I would greatly appreciate it!
If you are trying to salvage the float positions that you have, I would first eliminate any high RMS points and high PDOP points. Float positions aren’t always poor but like Tim said, fix is really what you want 100% of the time. Remember that you can be logging while collecting on NTRIP. This gives you a back up plan by attempting to do static processing from a publicly available base with a reasonable baseline distance to your study area. if you did log, there’s a chance that you could get PPK fixed off the nearest reliable CORS.
Emlid has a vast resource on the operation procedures of their receivers.
Just using a simple search on Google with “Emlid docs” gives this link
Many professional knowledgeable people and Emlid employees will bend over backwards to help you, however I suggest you read up on the documentation first to a least familiarize yourself with the procedures and operation of their receivers.
Float solution status means the rover receives corrections from the base but can’t resolve all ambiguities. Its accuracy usually ranges from several decimeters to 1-2 meters. Fix means the receiver resolved ambiguities in the position calculation. It usually has centimeter-level accuracy.
While some FLOAT points may appear accurate, others can be significantly off, as you’ve observed. So, a Fix solution is preferred for high-accuracy applications.
RMS
It’s the average deviation of your survey points from the actual position. Lower RMS error means your data is generally more accurate—like scoring closer to the bullseye in a target game.
Satellite Geometry:
PDOP shows the quality of satellite geometry. High DOP means poor geometry when the satellites are too close to each other in the field of view. The lower the DOP, the more accurate your data is. Ideally, aim for a DOP value below 2.
Possible reasons for getting FLOAT points:
Environmental conditions: The GNSS receiver calculates a solution based on the signals it receives from the satellites. When they are blocked by obstacles nearby, such as trees, it becomes harder for the unit to resolve all the ambiguities in its position and provide you with an accurate solution. That’s why we recommend following recommendations from this guide for the receiver placement.
Corrections stream: You need to ensure a stable internet connection while using an NTRIP. You can find what else is worth checking from the NTRIP service’s side in our support tip.
RTK (Real-Time Kinematic) offers immediate centimeter-level accuracy using real-time corrections from a base station. PPK (Post-Processed Kinematic) involves recording raw data during the survey and processing it later in Emlid Studio. So, you can still get similar high accuracy without needing real-time corrections.
@timd1971, thanks for sharing the link! I believe our guides will provide further helpful details.
Thank you so much for this, this is extremely helpful. I think that next time I am going to try using a different device to use as a hotspot for the NTRIP connection, since the service was a bit weak.
