When processing PPK, I have one processing instruction that says to process it once, and then look at a graph of the Standard Deviation and of the RMS, and then reprocess it with a new value for the SD that is down near to the level of the RMS. I have a couple of other instructions that make no mention of this step. Does performing this step actually help anything?
Stephen
I may have misunderstood the question but I think reprocessing the GPS data does not alter the positional accuracy of each PPK point. If the RMS is 1 reprocessing it using another accuracy filter will still produce the same results.
What I do for PPK processing is to process it at default values & select or delete the points that I think are outside of my accuracy requirements.
Thanks for responding Agrimensor. I probably should have worded the question with more detail and specificity...
The software we are using is GrafNav and we are using it for mobile mapping truck-mounted scanner trajectories. We received these instructions from a UK tech who was training us in this particular system.
You add your rover and base files and then let it process once. Then you look at two charts: "RMS - C/A Code" and the "RMS - Doppler" which show a Standard Deviation line and an RMS line. For both charts the goal is to bring the SD line down to close proximity with the RMS line. Reprocess it, but this time go to Advanced Settings and in the Measurements tab, under Measurement Standard Deviations, change the Code and Doppler setttings to lower numbers, what you judge to be very close to the RMS numbers from the charts you looked at. Further we were told to not make the Doppler setting any lower than 0.1m/s. Then you re-process it and then the charts show the lines very close together, which again, is what we were trying to achieve.
This week, we are with a senior tech who has been processing helicopter trajectories a long time, and this was new to him and he doubts that it has any real effect on the results. Which may be true, and also is probably easily tested, but I was just wondering if others have encountered this.
Stephen
Howdy,
It seems to me that what is being recommended is "merely" a rescaling of the weight matrix to reflect the reality that the computed weights were too optimistic.
If weights were correctly determined the variance of unit weight would be one (hence the term "unit weight"). The rescaling will "force" the adjustment to better approach one. Statistics and variances will be impacted generally not derived positions.
HTH,
DMM
I agree. That is correct.
Not a GPS guru
I thought PPK stood for Punt, Pass and Kick.........the nationally sponsored competition for kids held each year.
