[FOSS-GPS] High accuracy positioning with low cost GPS devices:a FOSS project

[Eugenio Realini]

Michael Tandy wrote:
>
> Something I've been wondering about is: the UBX-RXM-RAW message has a
> known jitter of a few milliseconds [1] and the time-tag is only
> precise to 1 millisecond, while some sources [2] suggest that time-tag
> errors between receivers need to be in the microsecond range. Was the
> imprecise time tag a problem for you, and if so how did you overcome
> it?
>   

We have read the links about the time tag problems.
Concerning goGPS, at the moment these problems are not taken into 
account. This means that we are considering master and rover perfectly 
synchronized: all the de-synchronization is thrown in the noise. This 
could be an explanation of the noise correlation.

A better solution would be achieved by changing the code and phase 
observation equations adding some model to describe the time-tag errors 
(maybe using other information coming from LEA-4T, such as UBX-TIM 
messages). This would be more important for phase observations than for 
code observations, because of their different accuracy. Let's consider 
for example two receivers kept still and let's compute their phase 
double difference over a certain period. In principle, phase double 
differences should be equal to a constant value (without cycle-slips), 
but actually, if at least one of the two receivers is an u-blox, it 
comes out that double differences have a significant trend. As you 
suggested, this could be attributed to time-tag errors; a time-tag error 
model could be calibrated to reduce (or eliminate!) this trend.

Do you have any idea about how to handle these time-tag issue?