Today I occupied a point surrounded by tall trees for 30 minutes, reading 5-7 satellites. I ran out of time, shut off the receiver, fully expecting that the solution would be "ambiguous" ..as it was after processing, standard error 1.2 feet.

I know that if I had stayed on the point another 30 minutes, I would most likely have a good solution.

Question: If I go back to that point tomorrow, occupy for another 30 minutes in the same conditions, will the processor attempt to make use of the new data as though it was a 60 minute observation? That is, does it accumulate data from different observations to reach a solution?

Nearly,

Ya got me on that one. never tried it. Let us all know if it works OK.

Or NOT.

Jimbo

Nearly,

Sounds a lot like the experiment we tried a couple of weeks ago with the dual 5 minute static sessions. They were apparantly processed seperately and the second session was viewed as a repeat vector. That's the reason we've been tinkering with the multiple 5 minute sessions using the k-kit. It does seem to handle the data differently.

Looks like, to me anyway, the multipath monster gotcha. I'm hoping that with longer sessions in known multipath canopy, and the additional "data scrubbing" capabilities we may get in ver. 2.4, we'll be able to safely clean up some of those problems you (and I) seem to encounter. Not sure if anything will help your previous 30 minute session other than a repeat session and "gut instinct.

Just my babbling opinions.

J.D.

Nearly,

The processing software will not "accumulate" observations in the manner you described. The two separate observations will produce two separate vectors.

Bill

Almost Perpendicular,

As Dad pointed out, we can do this in the kinematic mode doing PK type observations. The biggest drawback is that the data must reside in the SAME kinematic session. So if one occupation busts then the whole thing does, unless you happen to have more than two occupations on the point. I would definitely like to see a feature that would allow for processing of multiple sessions in a single solution as selected by the user. The user could select between using an adjustment of two or more observations or complete processing of two or more observations. I don't think this currently exists (as it has been pointed out that the new ver 2.x does not process psuedo kinematic data as the processor views two kinematic occupations as separate observations. This, by the way is not bad because it should allow for the same warm and fuzzies you get when viewing repeat vector data in static work.) I like the change, but I think a blend of what you and I are wanting would be a nice addition to any future releases.

Regards,
Shawn Billings
SIT Texas

Yes, this second vector will be treated as a repeat vector if the 2 stations are named the same as they were with the previous vector. You can look at the differences between the 2 vectors in the Repeat Vector tab.


With Solutions 2.40 you will be able to view the raw data and clean the data up. You might then be able to get the 2 repeat vectors to agree reasonably. It depends on how much multipath and signal blockage is present.


When the Minimimally Constrained Network Adjustment is carried out with all vectors included, each repeat vector will be tested separately by the Tau test. The vector (repeated or not) with the largest Normalized length residual should be cleaned. Run the minimally constrained adjustment again with all vectors included.

Then, if there are still tau test failures it is best to first exclude any vector (repeat or not) that failed the Tau test and has the largest Normalized length residual, and perform the Minimally constrained adjustment again, continuing this process until you have no vectors that fail the Tau test.

Often in Least Squares adjustments of GPS data there is one vector with the largest normalized length residual which causes the vectors near it to also fail the Tau test. Once this vector is excluded, the other vectors often do not fail the tau test.


After excluding all vectors that failed the Tau test in the Minimally Constrained able to re-include some of the vectors that had failed the Tau test previously, as long as they do not fail the Tau test again, and there are at least 2 vectors coming into every station. I believe it is best if there are at least 2 independent vectors coming into each point. More independent vectors per point is definitely better for the Minimally Constrained adjustment.



Regards,
Richard

Magellan Technical Support

Richard,

Will we be able to clean up the raw data, save it then run it thru Processor 1.2 ????

Effectively regaining PK capability and utilizing the new Tools in Solutions !

Jimbo

Jimbo,

This is my most sincere hope as well Jimbo. Could we be so fortunate as to have this flexibility?

Richard,

Your advice is well taken and explains a great deal, but I think you may have over-looked a few questions. As Nearly Normal pointed out and as I attempted to say, there are occassions where processing multiple sessions together as opposed to adjusting them may yield better results (If not better, then at least more flexibility to the user). Psuedo kinematic depends on this feature. The main problem I see is that all data in a PK session must be in a single session file using V1.2. Therefore if a point is blown in yesterday's PK work I can not go to the point and observe for only 5 more minutes, but must observe for 5 minutes, wait an hour and observe again. Having simultaneous processing of data at the users discretion
translates to a potential hour of field time savings provided that my repeat today agrees well with one of my occupations yesterday. In Nearly's case he could save a potential of 30 min of reobservation time by using data from yesterdays blown point provided that the error was simply due to not having obtained enough data in the failed session. To be honest, I don't understand all of what I am talking about, except that as learned from experience, there is a substancial difference between processing two sessions of data together, and adjusting the results of two processed points. (I know that's nothing new to you, but felt it deserves some comment). If this feature could be added to subsequent versions of Solutions, I think it might be worth while. Would such a processing technique be viable???

Thanks in Advance
Shawn

Richard Phelan:

When you say that with v2.40 we will be able to "clean up data" do you mean that by removing the signal-to-noise "blinks" caused by tree trunks the solution can be improved on a given observation?

My understanding of the processing algorithm is intuitive only, but I assume that if a single satellite blinks out for 3 minutes then an extra 3 minutes will gather equivalent satellite exposure.

I am learning after 1.5 years of daily trial n error that if you have several tall conifers in your horizon, the solution is really hard to predict. For example, 5 tall trees within 100 feet of the receiver: with perfect sky this would be a 12-minute shot with 8 sats, and a 20-minute shot with 7, but with the trees I can gather 1 hour until I have a solid left light (base 1/2 mile away) and sometimes its a great solution and sometimes its useless. As I say, its unpredictable.

Would the new tools in v2.4 improve an observation like this?

I locate a lot of remote corners that take 3 hours to reach on foot one way, which amounts to one corner per day, so naturally I am interested in making the time on site as efficient as possible(and as "fixable" as possible in those cases that present ambiguities.)

Thanx for your attention to my earlier question.

I response to the question about whether or not "cleaned" GPS data from Solutions 2.40 could be saved, and then imported into LOCUS processor to be processed again:

Solutions 2.40 does not have a method of saving the "cleaned" GPS data so that the data can be imported into LOCUS processor as raw observations (B, E, and D-files).


Regards

Richard