I am in need of information regarding RTK and collecting raw data while observing a point in RTK mode and how it relates to Kinematic surveying.

I recently upgraded to the latest and greatest Z-Extreme firmware and I use TDS for data collection.

Yesterday I surveyed around 30 highway right-of-way monuments in RTK mode while collecting raw data at the Rover and the RTK base; additionally I had a Promark2 unit set up on another control point several miles away. My general procedures are to double tie the monuments from two different RTK base stations and adjust the points using the observation files. I will also do some static observations on primary points (section corners) throughout the day.

I ran out of time and did not get a chance to shoot the right-of-way a second time, so I came back to the office, downloaded all the receivers and to my surprise, I had static vectors from both the RTK base and the Promark2 receiver to the highway right-of-way points. I post-processed the vectors and I had only 2 float solution out of the 60 vectors. The occupation times were around 1-2 minutes in addition to that; I also have the RTK solution (which may be the same solution as the static vector from the RTK base).

So the question is? When the RTK engine initializes at the rover (OTF), does this have the same result as initializing on a known point or using the bar with the Promark2 kinematic scheme. Is this how RTK/Kinematic works or did I get lucky?

Certainly something to ponder and surely experiment with.

JT Essex

I have seen this sort of thing myself without a complete understanding of why it seems to work so well. I'm guessing that you were not expecting solutions from the ProMark2 to the ZXtreme rover.

I saw this same sort of thing with the Z-12 Sensors and the Dimensions. My guess is that the ambiguity resolution provided by the two dual frequency receivers provided a narrow enough circle for the single frequency vector to come together. It's only a guess since I do not know how the processing engine makes it work.

It makes a little bit of sense when you consider that using two single frequency receivers as a base with one single frequency rover you also get two vectors, regardless of which base provided the initialization.

Here is something that seems more strange to me. A guy with one dual frequency receiver and three single frequency receivers uses the dual frequency receiver as a rover. He told me that the solutions are much more dependable when he does it that way than when he leaves the dual frequency receiver as a base when he wants to use OPUS. I do not know why this would make a difference. But I am content with whatever provides him with good solutions.

I would not exactly call the RTK and Stop&Go solutions identical because two different processing engines are used with slightly different data sets but it is essentially the same satellite geometry.

Now I have a question that may affect how short a data set will provide a solution. This is about that raw position that you get when you download the GPS data and look at it before you do any processing. Before SA was turned off most of the positions were within about 300 feet. After SA was turned off most of the positions were within 50 feet. But something said at the seminar in Riverside makes me wonder whether the circle is much smaller in the last year or two.

Some people mention very short data sets that work. These are data sets shorter than what I would use but when they are good and repeatable results there is no reason to disagree. As one instructor at the seminar in Riverside said, one of the most difficult questions to answer is...

What is the least amount of effort required to get the right answer?

My question is, what is the size of the circle these days? I'm curious about the horizontal and the vertical. When you edit those control site coordinates, how close are they really?

JT, what happens if you pull out the L1L2 base data and reprocess your rover shots against the PM2?

Also, how many birds were in the sky and where are you geographically?

Finally, how long was the rover receiver running and collecting raw data BEFORE you took your first stop and go shot? It's possible that you had enough data to fix ambiguities long before you ever took that first shot...

Look at your TDS raw file for dxdydz data and see how it compares to your vectors in Solutions....

inquiring minds want to know..

I would surmise that in general dual frequency observations can be of less time than single frequency. With the base sitting still, single frequency can get a long and happpy solution. When Solutions solves the vector between them, the dual frequncy says "I am good and fast" and the single frequency says "I ain't movin". VOILA! SOUP IN A MINUTE!

L1 gives very good short range vectors, so intuitively it now makes sense. I had thought that when I got 1 dual frequency unit I would use it as my base, but now I will consider the opposite.

How close are we?

Part of the new satellites going up is the abiltity of one SV to communicate with another SV. One of the most important things an SV receives is the satellite ephemeris. Currently it is once an orbit, and I believe it receives it while over the Pacific so that it has current information when it gets over us. There would be no good reason to update it more often unless a better algorythm for calculating and predicting orbit is available. I believe the better algorythm is already in place. Evidence to this is that post orbit data is now available sooner than in the past. We have more satellites, more CORS to check them against, and every year faster computers allow more information to be processed in less time.

10' circle.

I did not specifically pay attention to the raw position from this weeks dual frequency work, but I think it was in the decimal seconds. Putting in the OPUS numbers over the raw, all corrections were in the decimals only, with some in the tenths, most in the hundreths and beyond. Height was grosser.

Paul in PA
Modified By Lawrence Paul Lopresti on 11/14/2004 at 9:07 AM

Ellipsoid : GRS80 / WGS84 (NAD83)
Equatorial axis, a = 6378137.0000
Polar axis, b = 6356752.3141
Inverse flattening, 1/f = 298.25722210088

First Station : test
----------------
LAT = 42 51 14.34926 North
LON = 72 12 50.77849 West

Second Station : aldn
----------------
LAT = 42 51 14.21960 North
LON = 72 12 50.63740 West

Forward azimuth FAZ = 141 19 8.7820 From North
Back azimuth BAZ = 321 19 8.8779 From North
Ellipsoidal distance S = 5.1254 m

Good Morning

Thanks for your replies, I did not expect to get any vectors from the RTK base or the L1 base. I only downloaded the RTK raw data file because the older firmware version for the Z-Extremes had some issues with file storage. But the new firmware for the Z-Extremes is a marked difference and works very well. Since I upgraded the Z-Extreme firmware this is the first time I have seen the vectors come out in Solutions. Most of the time the RTK raw data file will be full of nonsensical point numbers and question marks.

For all you Z-Extreme users, I would highly recommend upgrading to the new firmware. Doesn’t cost anything and it’s easy to do.

I used the RTK observation file and the vectors from the L1/L2 base and the L1 base and adjusted the network. I then check the results against the RTK data alone. I found a couple of hundredths difference among most inverses.

Robert
I am working around Latitude 46-30 and Longitude 108-03. As per your question regarding how long the rover was collecting data before a collected a point, most of the time I had the RTK rover inside the pick-up, drove up to the corner, recovered the monument, took the rover out of the truck and went over and shot the point when it initialized. So not long at all, which has me really wondering? I did this all day long, but generally I have very little tree interference and around 6-8 satellites on all shots.

Here are a couple of differences for the lat/lon before and after adjustment. The raw data is just an autonomous down loaded position and the adj value is two 8 hour sessions post processed with 5 CORS stations at each end of the project.

Decimal seconds

Lat
Raw .30064
Adj .24151
Lon
Raw .59367
Adj .57439


Lat
Raw .28461
Adj .11386
Lon
Raw .68054
Adj .69248


Lat
Raw .33443
Adj .45939
Lon
Raw .09524
Adj .07510

Later on today I may make a point by point comparison of the RTK coordinates against the adjusted Coordinates.

I will certainly take a closer look at the data, if I had known about this earlier I would have extended my occupation times. I will do that in the future though.

Thank to all who have responded and have a nice weekend.

JT

Point by point comparison.
It was easier than I thought, below are the differences between the RTK data and the post processed and adjusted data and the repective inverse distances
(adjusted - RTK)


Delta Delta
PT N E Inverse
23 0.028 -0.033 0.043
24 0.028 -0.033 0.043
25 0.024 -0.028 0.036
26 0.022 -0.031 0.038
27 0.023 -0.030 0.038
28 0.023 -0.033 0.040
30 0.014 -0.021 0.025
31 0.017 -0.024 0.029
32 0.018 -0.021 0.027
33 0.021 -0.025 0.033
34 0.021 -0.030 0.037
35 0.020 -0.024 0.031
36 0.017 -0.023 0.028
37 0.017 -0.025 0.030
38 0.018 -0.025 0.030
39 0.019 -0.017 0.025
40 0.015 -0.019 0.025
41 0.017 -0.014 0.022
42 0.016 -0.014 0.022
43 0.007 -0.018 0.019
44 0.013 -0.016 0.020
45 0.018 -0.014 0.023
46 0.018 -0.017 0.024
47 0.008 -0.014 0.016
48 0.009 -0.013 0.015
49 0.004 -0.023 0.023
50 0.010 -0.016 0.019
51 0.011 -0.016 0.020
52 0.011 -0.011 0.016
53 0.006 -0.011 0.013
54 0.001 -0.006 0.006
55 0.002 -0.009 0.009
56 0.005 -0.011 0.012
57 0.003 -0.010 0.011
58 0.006 -0.011 0.012
59 0.001 -0.007 0.007
60 0.004 -0.007 0.008
61 -0.009 -0.004 0.009
62 -0.001 -0.010 0.010
63 -0.018 -0.007 0.020
64 -0.016 -0.016 0.022
65 -0.006 -0.009 0.011
66 -0.010 -0.018 0.020
67 -0.011 -0.020 0.023
68 -0.018 -0.013 0.022

JT