Well we just completed the processing phase of our psuedo kinematic test using Locus receivers and processor 1.2. Unfortunately the test would appear to be a failure. The test consisted of 5 min obs of 9 known points of various baseline lengths (all under 6 miles). After the circuit was tied we made an additional series of obs 5 min in length on the points. The second round was about 2 hrs after the first. Only one point resulted in a pass of the nine and was about 2500' from the base station. We had good s/v coverage (8-9 and occasionally only 7) with low PDOP. I am uncertain at this time if the poor results were in the processor, the receivers or if this method is just not as robust as we had hoped. Most points were around 1-2' off and the verticals about the same. Baseline distances didn't seem to have an effect as some of the further points observed measured closer than the medium range points. Perhaps it is best to let the dead (psuedo kinematic) rest.

Shawn

Shawn,

Sounds like a valiant attempt to increase the usefulness of the Locus units.

If Bill was serious about adding the capability back into the software he might want to look at your data to see why there are so many failed points !
Or is this the best, sub-meter, that PK can do ?

PK might be dead but still a great effort.

Thanks for sharing,

Jimbo

Hey, it wasn't a failure. It was a successful elimination of a non-viable procedure. I've learned more from so-called "failed" experiments in my lifetime than any successes.

Still wonder if this procedure would work on those points within about 2000 feet or so where we cannot maintain sv lock for standard kinematic? Maybe another weekend experiment in the works..... We'll see.

J.D.

P.S. Thanks Brian, Bill, Linda and all the other Ashtech folks for putting up with our B.S. I guess we think we're pseudo-scientists. (or is that sumo-schitchofrenickz)

Good attempt, Shawn.

You never know until you try. Thanks you for sharing the results.

-Greg

New test results soon. Different approach to same ends. I'm sure Shawn will post some details soon (since he'll be in the office most of the week).

J.D.

Shawn,

I'm curious. If you send me the data, I'll take a look to see what happened. My guess is that the failure is due to the processor used, or a problem with field techique. Contact Tech Support for instructions on ftp'ing the data, and ask them to pass it on to me, or e-mail it to bewing@magellangps.com.

J.D. raises a good point; the purpose of such an experiment is to determine whether a given procedure is viable. If it is not, the experiment isn't a failure, as you have eliminated one possibility. .

Regards,
Brian
Modified By Brian D. Ewing, PLS on 7/16/2001 at 9:41 AM

We had a successful run yesterday on three points (all that were tested) using psuedo kinematic. As Brian said, field technique vs. processor needs caused the failure in the first test, and as Dad said, we know what you can not do. After yesterday's test though, we can say we now know what you can do and it seems very viable to the surveyor in the weeds (Greg, I thought you might take an interest in this). The procedure is identical to standard kinematic except for no need to initialize or maintain lock. Syncronize as usual to the base and then to the rover and collect data. Occupy each point at least twice using 5 min observations with about one hour between the 1st and 2nd. NOTE and this is very important-Occupations 1 and 2 must be in the same kinematic session file to be properly processed. This also means (as our first test indicates) that 5 min static observations will not process properly. Float solutions will occur (1-2' error). Now it would seem possible that for redundancy 3 occupations could be made in a session file, or 2 occupations made in the first session file and 2 made in a subsequent session file. The great part of this method is that it does not replace kinematic, but compliments it. Both procedures can be performed seemlessly in the same session with the only change being an initialization for kinematic, and changing the occupation timer and no change made in processing.

Shawn

This only represents three points tied but the data was extremely good so I thought I might post the results. "POST" is our base station and minimally constrained control for our solutions shown. The points we observed are part of the City of Kilgore GPS control network that was established using L1 and L1/L2 Static GPS.

POST E 3100685.732
N 6835851.505
ELE 397.444

0028 E 3102498.424 0.021
MEAS N 6832456.472 0.033
ELE 396.868 0.083

0028 E 3102498.393
CALL N 6832456.463
ELE 396.89
-------------------------------

0023 E 3100823.205 0.021
MEAS N 6838743.981 0.028
ELE 396.195 0.082

0023 E 3100823.204
CALL N 6838743.975
ELE 396.22
-------------------------------

0019 E 3101717.091 0.030
MEAS N 6848711.057 0.043
ELE 370.031 0.159

0019 E 3101716.998
CALL N 6848711.084
ELE 369.94
-------------------------------

The column on the right are the 95% confidence errors.

Shawn

Shawn,

That is good news for all of us. I think this will help all that are using the Locus in K mode.

Techinque and processor needs were definitly the key. I have a feeling this is a GOOD TECHNIQUE. Hopefully the capability will be added back into the V2.0 software, possibly with a switch if nothing else.

I might have some free time late this afternoon to run my own testing.

Thanks,

Jimbo

Congratulations Shawn! Experimentation is the only way to see if something works.

By the way your previous posts on the main board about wanting to better understand least squares were right on the money too. The more one understands the tools they use the more likely they are to use them right. Given your natural curiosity I hope you will be able to take college courses to build up your mathematics background so that you can delve deeper into subjects such as least squares.

At your age you have a long way to go and by ten years from now coordinate measurement technology will be unrecognizable relative to that today and will definitely favor those with a fair amount of mathematical capability.

Shawn, JD, Brian and others,

Ran a good mix today of K and Pk, Pk being to prior control set many moons ago, so as to have a means of x-checking.

I'll start a new thread with these results, hopefully tomorrow.

About ½ the area (18 acres) with K also topo'd with Total Station. Wanted to have a good x-check on the final surface here. Didn't topo the same spots intentionally, but the contours should end up relatively close.

Jimbo

Good luck, Jimbo. Let us know what happens.