Float mode with accuracy below 5 cm

Q&A forumFloat mode with accuracy below 5 cm
tseco asked 4 months ago

We are working with a base station and with moving base/rover configuration.
In some areas the accuracy provided by the GPS is around 5cm but with float mode. In this mode, the postion suffers from some jumps and the gps data in not valid for localization purposes. 
We would like to know if it is possible to get that accuracy in float mode and what are the generic causes of this float mode. In https://www.ardusimple.com/question/3d-dgnss-float-fixed-what-meaning-exactly/ there are some relations between the accuracy and fixed and float mode.
Thanks in advance,

Ardusimple Staff replied 4 months ago

FLOAT is a way to tell the user that the confidence level of the RTK calculation is low, because one mathematical equation could not be resolved. It’s still better than NO RTK, but position might jump as you have experienced. If you use the accuracy estimate value, we recommend to multiply it by 2 or 2.5 to get an accuracy level value with high confidence.

How to increase the RTK FIXED time? Better signals in the rover and making sure electronics are inside a box and not with direct airflow or sunlight.

alpmustafa93 replied 4 months ago

does it really affect if the electronics(rover module) get sunlight/airflow?? it is the first time I heard this.

Ardusimple Staff replied 4 months ago

yes, it affect performance

clive1 replied 4 months ago

If the local clock source (TCXO) in the ZED module takes a frequency excursion it will look like a massive acceleration event, and the correlators will lose tracking across all satellites/channels. An open module would get this with a sudden changes of speed/direction, where the air flow changes dramatically/quickly.

You could no doubt demonstrate this with a can of freeze mist.

When testing TCXO in a lab, you want to cover them, as people walking past will disturb the air/temperature in a room.

1 Answers
clive1 answered 4 months ago

We would like to know if it is possible to get that accuracy in float mode and what are the generic causes of this float mode.

Sure, it means it can’t close the deal on 2 cm because there is too much measurement noise, things like indirect signals reflecting off close surfaces, or being blocked on one side, like in the shadow of a house or fence. A less than ideal base station sky view will also diminish the ability of the rover to resolve its position.
Use ground planes under the antennas, use the uCenter SkyMap to see what blind spots you have, mount the antenna so it doesn’t oscillate/vibrate, and do a proper survey of its location so the measurements/observations aren’t distorted

clive1 replied 4 months ago

Perhaps you have access to a roof mounted choke ring antenna, with a well established location? Check you are getting both bands for each constellation you are using. Check you are not getting RTCM3 packet loss in your inter-receiver connections. Use the highest baud rates that are permissible, and check that there isn’t significant latency in the links.

Record and analyze UBX-NAV-SIG data to understand the quality and availability of signals, and whether you have pairs of signals for each constellation, and complementary data from the base.