Taking accurate measurements in a difficult environment for GPS

#1
I was recently working in the field on a topographic survey in a densely covered lot, from the building itself and tree cover. This lot was also very steep. We set up the Triumph 1-M base at the top of the lot on their driveway and it was fairly open overhead which led me to believe it was a good spot for the base. We set another point at the bottom of the lot. This point had about half the sky covered from the building. We let the Triumph LS log data for 2 or 3 hours and then shot it again for about an hour. These points were measured conventionally with a total station too. I have attached the GPS reports:
1537549974225.png
1537550000720.png

DPOS REPORT
JAVAD POST-PROCESSING SERVICE
=============================


USER: DATE: 14.09.2018
FILE: 5000 TIME: 21:56:09.171 GPS time


SOFTWARE: DPOS(5/29/2018 9:09:56 AM) START: 2018/09/13 05:29:24.000 PM
EPHEMERIS: brdc STOP: 2018/09/13 09:47:59.000 PM
OBS USED: 7463 / 9246 TIME SPAN: 259 min
ANT NAME: JAVTRIUMPH_1MR NONE Geometry factor: 1.2
ANT HEIGHT: Slant(SHMP) - 1.777 meters FIXED AMB: Yes
ANT DELTA N/E/H: 0.000/0.000/0.000 OVERALL RMS: 0.008 meters


REF FRAME: NAD83(2011)(EPOCH:2010.0000) IGS08 (EPOCH:2018.7000)

X: -2708515.323(m) 0.004(m) -2708516.298(m) 0.004(m)
Y: -4253129.588(m) 0.004(m) -4253128.099(m) 0.004(m)
Z: 3892852.679(m) 0.004(m) 3892852.777(m) 0.004(m)

LAT: N 37 51 20.34711 0.004(m) N 37 51 20.36418 0.004(m)
E LON:E 237 30 35.43359 0.004(m) E 237 30 35.36724 0.004(m)
W LON:W 122 29 24.56641 0.004(m) W 122 29 24.63276 0.004(m)
EL HGT: 75.891(m) 0.004(m) 75.373(m) 0.004(m)
ORTHO HGT: 108.358(m) 0.004(m) [NAVD88 (Computed using GEOID12B)]

UTM COORDINATES STATE PLANE COORDINATES
UTM (Zone 10) SPC (NAD83(2011) /0403 California zone 3)
Northing (Y) [meters] 4189921.964 652313.963
Easting (X) [meters] 10544850.569 1824882.822
Convergence [degrees] 0.31288250 -1.21843795
Point Scale 0.99962478 0.99993084
Combined Factor 0.99961287 0.99991893

US NATIONAL GRID DESIGNATOR: NONE


BASE STATIONS USED
PID DESIGNATION LATITUDE LONGITUDE DISTANCE(m)
NONE CAPO(CORS) N 37 42 51.020 W 122 13 19.445 28360.1
NONE OHLN(CORS) N 38 00 22.498 W 122 16 22.708 25376.9
NONE P181(CORS) N 37 54 52.349 W 122 22 36.267 11927.8
NONE P224(CORS) N 37 51 50.014 W 122 13 08.563 23878.5
NONE TIBB(CORS) N 37 53 27.139 W 122 26 51.317 5414.7

RESIDUALS OF POSITION
VECTOR dX,m dY,m dZ,m dN,m dE,m dU,m
CAPO - 5000 0.003 0.005 -0.008 -0.002 0.000 -0.009
OHLN - 5000 -0.025 -0.001 0.006 -0.004 -0.020 0.015
P181 - 5000 0.010 -0.009 0.015 0.011 0.013 0.011
P224 - 5000 -0.024 -0.010 0.018 0.001 -0.015 0.029
TIBB - 5000 -0.031 -0.008 -0.002 -0.015 -0.022 0.017

****Disregard point 5002****. For 5001 and 5003 very few epochs were gathered, over a short duration, which was likely due to the location of the setup near the tall building. Between the RTK shots (5001 & 5003) on the LS there was 4.3' of horizontal difference and 2.2' of vertical difference. I used the shot with the longer time set to cook which was point 5001.

The RTK data is obviously not accurate but the raw static data that was logging over the long period of time I would think would be more accurate. The horizontal difference between the conventional points and the GPS base and most accurate RTK point (5001) was 6.4' and the vertical difference between point 5001 was 1.1'. With results like this I don't believe I can trust the RTK shot but how do I know that I can trust the Base measurement? I am attempting to convert the total station measurements to NAVD88. Do I need to remeasure shots with GPS or can the base be confidently used?

I had the understanding that if a measurement was taken over a long period of time it would achieve better results but perhaps this is not always the case. I need to know the limitations of the equipment that I am using in order to confidently accept the measurements.

I am still new to using the equipment so any advice on what to do when you are in an environment that is not friendly with GPS surveying is appreciated. Thank you all for taking the time to further my understanding of GPS and its applications using the Javad products.

Tyler Brown
 
#3
From what I can see, your rtk observations are too short. Do a search for "under canopy" under my name.
It did not collect enough data but the LS was running for a couple of hours on point 5001. It only logged those 4 epochs and the furthest time between the two shots was only 34 seconds.. forgive me if I interpret that incorrectly. Does the LS not collect static information as well? I assumed it may still have accurate static since it had been out there for so long.
 
#4
From what I can see, your rtk observations are too short. Do a search for "under canopy" under my name.
I see in your other post that I should have waited until it at least logged another point to satisfy the time requirement. What if this does not happen? What should be the protocol for a stubborn measurement such as what I was measuring.
 

Sean Joyce

Active Member
#5
It did not collect enough data but the LS was running for a couple of hours on point 5001. It only logged those 4 epochs and the furthest time between the two shots was only 34 seconds.. forgive me if I interpret that incorrectly. Does the LS not collect static information as well? I assumed it may still have accurate static since it had been out there for so long.
Buildings and a base that is not in a really good open are are killers from my experience.
 

Shawn Billings

Shawn Billings
5PLS
#6
It appears that the LS did collect raw data, but I can't see where the base/rover vector was processed. 5001, 5002, and 5003 are all using RTK solutions with only a few epochs. Did the vectors from base and rover process also?

Sean raises a good point. You mention the base was in a mostly open place. That's usually not good enough. You can push the rover into some miserable places, but the base really needs very open sky.

I think you misread what I said regarding another point. That's not what I'm suggesting. What profile are you using? Generally if you are working in canopy, you need to use the boundary profile. It requires a minimum of three minutes and often longer in canopy.

https://support.javad.com/index.php...inute-observations-in-canopy.2545/#post-22893
 
#7
I have attached an image of the base location, it was open other than a bit of the garage and maybe some trees behind me in the photo.
20180913_102523.jpg


I am using the boundary profile which is set to 200 epochs and 180 seconds but I stopped the rover after a couple hours even though it did not reach its validation stage or even get close. I realize this is why I have bad RTK measurements on the point. Is there a way to get an accurate position of the point using the raw static data when the rover was attempting to collect for a few hours? If the base is in a bad location does this mean we should not be using dpos? I am looking for a way to gain a potentially accurate measurement by processing through the CORS stations separately.
 

Shawn Billings

Shawn Billings
5PLS
#9
this video is a little dated now, but it covers the general work flow of processing files on the Triumph-LS. Use the second option in the settings screen of DPOS (Base-Rover processing + Base CORS processing).
 

Matthew D. Sibole

Well-Known Member
5PLS
#10
While this equipment is fantastic it is just another tool. Pick the best tool for the job.

With that said there are ways to get shots under or near heavy canopy and obstructions. What you are looking for is a few epochs agree with each other that have at least 240 seconds between them. You may have let the LS sit there for an extended period of time but on point 5001 you only have 4 epochs with 34 seconds between them. Not nearly enough time to make sure it is a good shot.
 
#11
Matthew Sibole above is seriously right.
It does not matter how long you were on that point. What matters is TIME between 1st, and last epoch.
I've sat in impossibleville, for 15 min. Got 10 groups, that were all no good. Finally, a good group passed the confidence test, and went all the way through phase 3, which took 3 to 5 minutes. For a total time on that point, of 18 to 20 mins. Now, ppk was on, so it did get the full 18 to 20 min. In ppk, but you won't see that til back at the office.
For Rtk, time between epochs, in the SAME group is what matters.
N
 
#12
You might get a good position with 5001 on post process. 5003 I highly doubt. Look at the GDOP (above 5). You're only getting a solution based on one direction. It's like trying to cross tape something in only knowing one distance and no direction. Unlike canopy (in most cases), buildings can completely destroy the needed geometry. You need to get far enough away from the building that the receiver can see something in that direction (other than directly above), or raise the receiver high enough. You might try the intersection routine. But probably quicker to use total station.