from GPS World on-line
Perspectives — February 2007
With OPUS, one person can operate one L1/L2 receiver to bring NSRS-compatible control into a project with a time investment measured in terms of a few man-hours or less.
Feb 26, 2007
By:
Eric GakstatterOPUS-RS: Rapid Static Processing
This newsletter has a significant international distribution, and I know I just wrote about the US National Geodetic Survey (NGS) a few months ago, but the new OPUS-RS NGS program has international implications. It paints an interesting picture of where things are headed in the geodetic/survey community with respect to Continually Operating Reference Stations (CORS) — not just in the United States, but around the world.
OPUS-RS is a derivative of OPUS (Online Positioning User Service). After 15 months of testing, OPUS-RS was declared operational on January 31, 2007.
A little background: the original OPUS is a free, geodetic-quality, post-processing service for L1/L2 GPS data. It was introduced five years ago by the NGS and is operated/maintain by the NGS. In simple terms, here’s how it works:
• Collect at least 2 hours of L1/L2 GPS data.• Use the NGS OPUS web-based submission form to send your data to them in Receiver Independent Exchange Format (RINEX) format; some manufacturer formats are supported too.• Specify the antenna type.• Specify the antenna height.
Within minutes, the data is processed with respect to three CORS sites (based primarily on distance, site stability and number of observations) and the centimeter-level results are emailed to your inbox in a standard or extended report format. Results are provided in ITRF and NAD83/CORS96 epoch 2002.0 and also in geodetic, State Plan and UTM coordinates, all tied to the National Spatial Reference System (NSRS).
The value of the OPUS/OPUS-RS concept:
Simple to operate. Easy/quick access to the National Spatial Reference System (NSRS).
Low overhead. One person can operate one L1/L2 receiver to bring NSRS-compatible control into a project with a time investment measured in terms of a few man-hours or less.
Robust solution. Multiple baselines used.
The OPUS concept would be much more difficult to implement without the CORS system in place. Briefly, CORS is a network of just more than 1,000 GPS reference stations spread throughout North and Central America. NGS owns very few of these. They are owned/operated by government agencies and private companies who choose to participate in the NGS CORS program and follow the NGS specifications for operating/maintaining a CORS site.
For detailed information on the CORS program, you can visit:
http://www.ngs.noaa.gov/CORS/For more information on the OPUS program, you can visit:
http://www.ngs.noaa.gov/OPUS/That brings us to the newest flavor of OPUS called OPUS-RS (Rapid Static). Whereas OPUS is much like the hours-long static GPS work you may have done in the past, OPUS-RS resembles rapid-fast/static approach you may have used in the past, except it uses publicly available GPS reference stations.
If you recall from above, OPUS requires at least two hours of observation time. OPUS-RS can process data sets with as little as 15 minutes of observation time. Where it differs from the rapid/fast static processing you may have done in the past is that OPUS-RS uses reference stations as far as 200 kilometers away. Given the density of CORS sites in North America, this means you can use the service in most locations in Canada, the United States (including Alaska/Hawaii), Mexico, Central America, Virgin Islands, and several smaller regions.
Although the OPUS-RS user interface appears strikingly similar to OPUS, OPUS-RS uses an entirely new software engine developed by the NGS and contractor Ohio State University. OPUS-RS is the first product using the new software engine, and that is the cornerstone of future NGS products.
If you’ve ever been involved in a static GPS survey campaign and had to manage more than one GPS receiver, you’ll appreciate OPUS-RS as well as its predecessor, OPUS. Without OPUS/OPUS-RS, you would have to setup and operate your own reference station(s) and perform the post-processing yourself.
OPUS-RS uses up to six GPS reference stations up to 200 kilometers away to produce a centimeter-level solution. OPUS-RS is picky about site stability, distance, and number of observations, so it won’t always select the closest reference stations to you. If you choose to select the reference station manually, you can select up to six.
To give OPUS-RS (or OPUS) a run and get comfortable with its performance, you don’t have to own a GPS receiver or even leave your office for that matter. You can download data from the CORS site nearest you and then submit it to OPUS-RS as if you collected it yourself. You can push the OPUS-RS envelop by submitting very short datasets or datasets during various times of the day.
Here are the steps and links if you want to do this yourself.
To download data that you’ll submit to OPUS-RS:
Go to
http://www.ngs.noaa.gov/CORSClick on the CORS coverage map and find a station you want to use. Remember, you will use data from this station to simulate your own that you’ve collected. For this exercise, I downloaded data from the CORV site.
Once you’ve clicked on the name of the station you want to use, look on the left side of the screen and click on Custom Files (UFCORS).
Then select a day/time of the data you want. For Number of Hours of data you wish to receive, select “1”.
In next screen, choose 30 seconds between points and select YES for NGS data sheet.
Then click Submit.
Note that data downloaded from CORS is sent to you in minimum 1-hour blocks. If you want to challenge OPUS-RS with shorter datasets like 15 minutes, you’ll have to use a text editor like Windows Notepad to shorten the data file. The data file is in RINEX format, which is an ASCII format. If you’re in a patient mood, a very useful but non-user-friendly program exists called TEQC that does a good job of manipulating RINEX files a number of different ways. You can download TEQC free of charge at:
http://facility.unavco.org/software/teqc/ teqc.htmlI used TEQC to chop up the RINEX data into 15-minute sections that I was interested in processing.
Once you have the data ready to submit to OPUS-RS, you need one more piece of information: the type of antenna. In my case, the antenna used by CORV is an Ashtech 700936E_C. You can find this in the header section at the beginning of the RINEX file.
Now, go to the OPUS-RS web site at:
http://www.ngs.noaa.gov/OPUS/OPUS-RS.htmlYou are prompted for your email address, the name of the file you want to send to OPUS-RS, the antenna type, and the antenna height. I left the antenna height at 0.0 because I used the CORV data. If you collected your own data in the field, you’d need to enter the correct antenna height.
Once the form is completed, click on the Submit button and check your email in a few minutes. The email you receive from OPUS-RS will contains geodetic results in the NAD83(CORS96) datum as well as in ITRF00. Also, you’ll receive State Plane Coordinates and UTM Coordinates referenced to the ITRF00 datum. You are also provided the ellipsoid height and ortho height based on GEOID03 and the NAVD88 vertical datum.
I processed two 15-minute datasets per day over a 10-day period from February 1 to February 10, 2007. Based on mission planning software, I choose a 15-minute period in the morning (~9am local time) when the PDOP was low and one in the afternoon when it was higher (~4pm local time).
Here are results from some of the days I processed. The values are in State Plane, Oregon North, meters. These are 15-minute datasets.
CORV = N105971.566, E2277335.373, 107.514 HAE
DATE/START
N
E
Z
dn
de
dz
Feb. 1 9:10amFeb. 1 4:00pm
105971.548105971.546
2277335.3872277335.389
107.455107.498
.018.020
-.014-.016
.059.016
Feb. 2 9:05amFeb. 2 3:50pm
105971.547105971.545
2277335.3882277335.389
107.503107.491
.019.021
-.015-.016
.011.023
Feb. 3 9:00amFeb. 3 3:50pm
105971.546105971.544
2277335.3892277335.390
107.526107.495
.02.022
-.016-.017
-.012.019
Feb. 4 8:55amFeb. 4 3:45pm
105971.545105971.544
2277335.3892277335.389
107.506107.488
.021.022
-.016-.016
.008.026
Feb. 5 8:50amFeb. 5 3:40pm
105971.546105971.545
2277335.3882277335.390
107.529107.500
.02.021
-.015-.017
-.015.014
Feb. 6 8:50amFeb. 6 3:40pm
105971.547105971.542
2277335.3902277335.390
107.515107.493
.019.024
-.017-.017
-.001.021
Feb. 7 8:45amFeb. 7 3:35pm
105971.545105971.543
2277335.3692277335.382
107.545107.511
.021.023
.004-.009
-.031.003
Feb. 8 8:40amFeb. 8 3:30pm
No solution105971.567
—2277336.923
—108.759
—-.001
—-1.55
—-1.245
Feb. 9 8:35amFeb. 9 3:30pm
105971.548105971.537
2277335.3842277335.389
107.437107.5
.018.029
-.011-.016
.077.014
Feb. 10 8:30amFeb. 10 3:25pm
105971.550105971.540
2277335.3872277335.392
107.519107.498
.016.026
-.014-.019
-.005.016
In looking at the data, what’s interesting is the precision of the collected data is pretty tight (with the exception of the Feb. 8 data which is discussed below). At first, I thought I had made an error somewhere in calculating the CORV State Plane coordinates, antenna selection, and so on, because the offset from CORV of each of the solutions was pretty consistent. However, I had the NGS folks check my work and I had done things right. They attributed the offset to error in the published CORV coordinates, seasonal variation of coordinates, and movement not accounted for in the NGS velocity models as CORV (Corvallis, OR) is in a region of considerable tectonic stress.
You’ll notice the difficulty on the Feb. 8 datasets. No solution was possible for the morning dataset even when I extended the dataset to 60 minutes — too many cycle slips. Also, the Feb. 8 afternoon solution had significant errors. It’s important to note that the quality indicators in the OPUS-RS report raised a red flag on these two datasets and also on the Feb. 9 morning dataset where the N and E components were very reasonable but the Up component was a little sloppy.
One item I was paying attention to were instances where the quality indicators were positive, but a poor solution was provided. I didn’t see this so it speaks well, given a limited dataset, of the integrity of the solution. OPUS-RS provides two sets of quality indicators for you to evaluate the position solution.
OPUS-RS also reports the reference stations that were used to determine the solution. There were six with the closest being 42km and the furthest being 187km away. Who would have thought you could obtain cm-level positioning using 15 minutes of data with those sorts of baseline distances?
I could spend a lot more time writing about this because there’s much more, but the great part about it is that you can try OPUS-RS without even leaving your desk. I even went back and processed some datasets I had from a current project. Give it a shot.
About the Author
Eric Gakstatter
About Eric Gakstatteremail:
egakstatter@questex.comSee more articles by Eric Gakstatter