The GPS Observer
Standards and specifications for horizontal controlThe last GPS Observer column was written by David Zilkoski, deputy director of the National Geodetic Survey. Dave was kind enough to put the final touches on the series of articles on standards and specifications for vertical control networks. We are going to ask Dave to write another article on this subject because it needs to be specified how the 2 cm accuracy GPS-derived orthometric heights are derived. I can tell you, without stealing his thunder, that it takes more than a few minutes of GPS observing time to accomplish those results.
In this column, I’m going to provide a brief review of the standards and specifications for horizontal control. The reference book I’m using should be on the shelf of every person who is even thinking about establishing geodetic control using GPS. The publication, Geometric Geodetic Accuracy Standards and Specifications for using GPS Relative Positioning Techniques, is available from the National Geodetic Survey. The document is dated May 11, 1988. (When I gave many GPS seminars in past years, this publication was included as an appendix in the seminar manual.)
You will notice that the words horizontal control were not mentioned in the title of this document. On the first page it states, “The GPS specifications are for control surveys performed by relative positioning techniques where two or more receivers are collecting carrier phase measurement data simultaneously. They are a guide for determining how to meet requirements for horizontal, vertical and azimuth accuracy standards.” The document was written to compliment the Federal Geodetic Control Committee (FGCC) Standards and Specifications for Geodetic Control Networks dated September 1984. That document was the reference used to write the first three GPS Observer articles on “Standards and Specifications for Vertical Control Networks.”
The 1984 standards for horizontal control are based on “distance accuracy standard,” which is the ratio of the relative positional error of a pair of control points to the horizontal separation of those points. As this ratio increases, the classification of the control survey degrades. As you will see later, a first order survey (order 1) is one part in 100,000 (1/100,000). If a relative positional error is constant, classification degrades as the minimum spacing between stations decreases.
The 1984 vertical control standards, which are based on elevation difference accuracies, are more stringent. (You saw that in the previous series of articles.)
GPS Classification StandardsFor GPS, six “orders” of geometric relative positioning standards are specified; these are summarized in Table 1. It should be noted that the accuracy standards are at the 95 percent confidence level (2), not the 68 percent value (1) that one normally gets from a network adjustment.
Most GPS surveyors are only concerned with establishing geodetic control of Order C. Order C is subdivided into Order 1; Order 2, class I and II; and Order 3. Most of us are more familiar with the expressions First-order, Second-order classes I and II, and Third-order. As stated earlier, the most stringent of the Order C surveys is Order 1, which requires a relative position accuracy of 10 parts per million (ppm), which is one part in 100,000 (1/100,000). And as most GPS surveyors know, a relative positioning error of 1/100,000 is not good because the system usually gives relative positioning error between 1/400,000 and 1/1,000,000 and sometimes even higher.
The remainder of this article will be a discussion of guidelines for GPS field survey procedures. This information is given in a long table, Table 4, in the 1988 document. I have reproduced only part of the first page and called it Table 2. I would like to expound on the guidelines important to the average GPS surveyor.
Before we begin the discussion, it needs to be said that long observation times are necessary to establish geodetic control. Techniques like rapid static, fast static, kinematic and real-time kinematic are not acceptable to establish control that meet the standards described in this document.
Looking at the top of the table, notice that the relative positioning standards are referred to as Group and Order. For AA, A and B surveys, it is Group and Order AA, Group and Order A, and Group and Order B. Group C was discussed earlier.
- The first item is two frequency observations required. Single frequency receivers (L1 only) are acceptable for all Group C surveys but not for the others. Believe me, L1-only technology is not dead.
- The recommended number of receivers observing simultaneously is three for any Group C survey; more are required for other surveys.
- The third guideline, satellite observations—is not a critical item. I don't know of any GPS manufacturer that uses RDOP (correct me if I'm wrong).
- The period of observing session—is critical. The statement “4 or more simultaneous satellite observations” goes without saying. The next statement, “Triple difference processing,” is probably not important today; most GPS software uses double difference processing.
Look at the next line, “other processing techniques.” The general requirement is 30 to 60 minutes of observing time for Group C surveys. There is also a time requirement of 20 to 30 minutes for continuous and simultaneous observations between all receivers.
- The data sampling rate, maximum time interval between observations, is 15 to 30 seconds.
- The minimum number of quadrants from which satellite signals are observed is three. The “or 2(k)” is explained later in the document. For two satellites, the satellites should pass through quadrants diagonally opposite of each other.
- The maximum angle above horizon for obstructions is 20 to 40 degrees.
All of these guidelines must be adhered to in order to meet the published FGCC guidelines (the name today is the Federal Geodetic Control Subcommittee).
That’s all we have room for in this article. Look for Dave Zilkoski’s article in March on how to achieve 2 cm accuracies for GPS-derived orthometric heights.