Localizations with DPOS/VB-RTK/J-Field

[Shawn Billings]

In the latest release version of J-Field, some exciting improvements have been made to the use of DPOS/VB-RTK and localizations.

In review, localizations in J-Field perform one of three operations:

1. Create Custom Projection to satisfy a set of grid coordinates to surveyed geodetic positions. An example of this might include using the Triumph-LS to survey several well-spaced points from a conventional survey in order to relate the RTK survey to the conventional survey or vice versa. In this context the geodetic relationship of the total station survey is Unknown. The rotation to geodetic North is unknown and the relationship of the grid origin (ie 5000,5000) to the Earth (ie Lat/Long) is unknown. The localization determines this relationship.
2. Determine a transformation from surveyed points to a known geodetic system. An example of this might be that you set your base up with an autonomous position. You then proceed to tie into points with known coordinates in a known coordinate system, such as State Plane. You tie into one or more points and perform a localization that determines the rotation, scale and translation (N, E, U) from your autonomous base to the surveyed control. If your surveyed coordinate system matches your target coordinate system, there should generally be no rotation or scale (outside of very minor observational errors in the survey or in the control), only translation.
3. Similar to No. 2 above, determine base translation for restarting the base. In this case all points using the autonomous base are adjusted by the translation parameters and the base point is translated as well. It should be noted, this requires the user to return to the base and restart it with the newly determined base position, so that the broadcast corrections being transmitted match the newly determined base position.

DPOS/VB-RTK works similarly to No. 3 above. Begin the survey with an autonomous position, while conducting the RTK survey, the base logs raw data. The raw data is downloaded to the LS at the end of the survey and sent to the Javad servers to process the data to reference stations around the world. In the United States, the NGS CORS are used. The process is similar to OPUS. Once the data is processed, the results are returned to the LS and the base position and all rover positions determined in that session are translated according to the DPOS solution. Note that this adjusts the positions, but this presents a problem for localizations. Localizations are built from the geodetic positions (lat/long) underlying the grid positions. So a problem occurs when the geodetic positions of the surveyed points used to determine the localization change, then the localization is no longer valid and the grid values of the surveyed points will no longer match the target coordinates of the localization. The development team responded to this predicament and the solution is now available in the current release of J-Field.

Now, a user can perform a localization in the field based on an autonomous base position, later perform a DPOS/VB-RTK adjustment, and then select the Localize button on the Home screen and the localization is automatically recomputed from the new DPOS solution.

 

[Shawn Billings]

Initial Field Localization:

Here is a localization that I did this weekend. Point 1 is fixed in 3D, point 2 is fixed in 2D (NE Horiz.)

The results of this localization produced the following translation, rotation and scale. This baseline is 570 feet in length, so 3 arc seconds is pretty small, 0.008', and 22.8ppm is 0.013 at that distance. The coordinates of 1 and 2 (the points I was localizing to) were HARN derived from a static survey we did a couple of years ago.

I forced these to zero by checking the box beside Rotate and Scale Difference.

For fun, I changed point 2 to "Check" which meant I was only using Point 1.

This resulted in my final Localization (using an autonomous base) being:

From this localization, we can surmise that the offset from my autonomous base position to the HARN derived control was about N -6.08 +3.16 -9.69.

 

[Shawn Billings]

DPOS Results and Final Localization:

After the fieldwork was complete, I used the stop base procedure in Base Rover setup to download the raw data file from the base receiver. Generally, it is necessary to wait until after GPS midnight for the data to become available to process in DPOS, so by the next morning, I was able to send the file to DPOS:

With this done, I then let J-Field adjust all of the rover points determined from this base position by this shift.

With this done, I navigated to Localize from the Home Screen and was notified:

The localization values were recomputed:

This reveals a shift from HARN to CORS of N -0.044, +0.137 and -0.297. Horizontally this is consistent with other HARN to CORS observations I have made in the area (See http://www.javad.com/jgnss/javad/news/pr20150312.html ). Based on other testing on this site, I suspect the original work had an elevation error, which I'm investigating now. This error isn't critical to showing the new localization feature.

So what's the upshot of this?
I exported the points from my survey expressed in the localization, before and after DPOS. Because the localization was automatically recomputed after DPOS, the grid coordinates did not change even though the underlying Geodetic positions did. This is fantastic news for anyone doing a localization from an autonomous base. You can now work your localization, stake points, adjust the base and surveyed points by DPOS, and your localized grid coordinates will not change. This means you can then toggle the coordinates to Latitude and Longitude or State Plane and have a high degree of geodetic accuracy (from DPOS) and toggle back to the localization to view your work in relation to your local system without issue.

 

[Shawn Billings]

Here are some of the coordinates as the process progressed:
VB101_1 and VB102_1 are surveyed points from a base on site. Thirty second occupations
with fixed solutions between base and rover.

To give some idea of the geodetic side these are the State Plane Coordinates of the
points as surveyed.
VB101_1: (State Plane - Autonomous Base)
N 6837701.276 E 3050573.924 U 400.660
VB102_1: (State Plane - Autonomous Base)
N 6837135.485 E 3050644.025 U 399.297

After localization, these surveyed coordinates match my local system very well.
VB101_1: (Local System - Autonomous Base)
N 197260.867 E 10565.917 U 390.951
VB102_1: (Local System - Autonomous Base)
N 196692.948 E 10616.429 U 389.587

As expected and desired, after DPOS, the geodetic positions of the surveyed
points change, translating with the previously autonomous base.
VB101_1: (State Plane - After DPOS)
N 6837695.269 E 3050577.435 U 390.668
VB102_1: (State Plane - After DPOS)
N 6837129.478 E 3050647.536 U 389.305

An unintended consequence of this change in geodetic
positions is the change in surveyed positions within the local system.
These surveyed points no longer match my local system since they
have been translated too.
VB101_1: (Local System - After DPOS)
N 197254.743 E 10569.219 U 380.959
VB102_1: (Local System - After DPOS)
N 196686.823 E 10619.731 U 379.596

Selecting Localize from the Home screen however, automatically
recomputes the local system and now my surveyed positions again
match my local system AND my geodetic positions are also accurate.
VB101_1: (Local System - After DPOS, After Selecting Localize)
N 197260.867 E 10565.917 U 390.951
VB102_1: (Local System - After DPOS, After Selecting Localize)
N 196692.947 E 10616.429 U 389.588

 

[Nate The Surveyor]

I've been at it for a while.
Localization is out of order.
It works fine, until I dpos it. Then, it all fouls up.
I keep it away from DPOS, and it's fine.
It now refuses to localize, from the updated (dposed) points.
I'm one of these idiots, that will sometimes stay up all night, solving thing, but this is out of my league.
I cannot write code. If I could....
I'd......
Dunno!

N

 

[Nate The Surveyor]

Moderator, please move this to a better place? if you want...
However, it now won't localize with the original base point.
N