Static GPS

What is differential GPS? It is the method of GPS surveying that employs more than one GPS receiver to simultaneously collect data which will be processed together to remove common errors. The final result of such processing is the realization of a three-dimensional vector between receivers. This vector can then be "added" to the known coordinates of one of the occupied points to create a "relative" position for the other, hence the term "relative GPS."
Differential GPS has many forms. It is often associated with Static GPS surveying and what is commonly known as Resource Grade Differential. However, in addition to Static GPS, most GPS surveying techniques today use differential methods, such as Real-time Kinematic (RTK) and Real-time Differential. Since differential techniques are designed to mitigate errors and increase accuracy, they are the methods of choice for most surveying applications.

Static GPS has the capability to produce relative positions at the sub-centimeter level on relatively short distances (a few hundred kilometers) and at the centimeter level over long distances (up to thousands of kilometers). In this article, we will look at the Static GPS technique, discuss project planning, and examine the advantages and disadvantages of Static GPS.


Why Use the Static Technique?
The Static Technique is most often employed when conducting control surveys, where accuracy is of primary concern. These surveys can be conducted to establish an azimuth pair (two intervisible marks) from which to base a conventional survey, or many single marks over a large area for the purpose of network or photogrammetric control. In either case, accuracy and reliability are the key factors.

Another advantage to using the Static Technique is that if the survey is planned properly to include existing vertical control, reliable heights can be computed for new points in the survey, thus possibly eliminating the need to conduct conventional leveling to support the project.
So what is the downside to Static GPS? Static GPS observations require fairly extensive observation times as compared to some other methods. How long should one observe? This question is best answered by the individual who is planning the survey, based on his or her past experience and project specifics. Above is a table that I use to determine minimum observation times for dual frequency observations.

Local factors such as a poor GPS window, a high multipath environment, or site accessibility may dictate observation times that are longer than normal.

Many people may think that these observation times are longer than needed, and they may be correct. My philosophy is that it is better to have more data than is really needed rather than not enough. Also, I have built a cushion into these observation times to counteract any minor problems that could be encountered in the field that might reduce the amount of data collected.

The most important part of any GPS survey is proper planning. This is especially true for Static GPS. Logistics play a major role in the success of Static surveys. This is due to the fact that Static GPS usually involves at least two (usually three or more) receivers, collecting data simultaneously on multiple points that may be many kilometers apart, and requiring many observation sessions. The following real-life example is of a survey that was conducted in southern Vermont; the Project Scope is outlined below:

The first step in the planning process is to produce a rough map of the project area, showing where the control stations are needed. Using an atlas, USGS quad sheet, or a GIS, produce a map of the project area. It is important that this map show all roads and trails that could be used for project access (see Figure 1).

Next, add to this map any existing (primary) control in the area (horizontal and vertical). Search the National Geodetic Survey (NGS), or local control databases for existing stations that could be used to control the GPS project. Remember that the primary control stations need to be at least as accurate if not more accurate than the desired accuracy of the survey. In this case, First Order horizontal has been specified. That means that we need to find existing High Accuracy Reference Network (HARN) marks of A- and B-order and/or First Order marks to control our survey horizontally. The vertical accuracy requirement has been specified as five centimeters. This will allow us to use existing First or Second-order bench marks to control our survey vertically (see Figure 2).

As can be seen on the map, there are no horizontal control points in the local project area, however, there are a number of vertical control points in the project area that might be occupied with GPS. The NGS datasheet will indicated if the mark has been reported as suitable for GPS occupations, although a field visit will be necessary to verify that.

The number of control marks will vary depending on the size and specifics of the project. Generally speaking, a GPS project should be tied to a minimum of two (preferably three), well-distributed horizontal control stations, and three (preferably four) well-distributed vertical control stations (if heights are required).

Next, identify any existing marks (USGS, county, town, etc.) which could serve as project points. This is a step that may often be overlooked, but has the potential to greatly reduce project cost. If existing marks can be found and used as project points, regardless of their original purpose, then this may eliminate the need to set new marks, and will reduce crew hours and monumentation costs.

Reconnaissance and Mark Setting
Next, take the project maps and station descriptions to the field and recover all existing marks that are to be used for the project. It is important to update the station descriptions if necessary, as the crew member finding the mark may not be the one observing it. Use your GPS equipment to navigate to the location where new marks are required. It is inevitable that at least some of the proposed locations will be in areas that are not "GPS friendly." When this is the case, find a location as near as possible to the proposed location that has an unobstructed view of the sky. This is also a good time to work with clients on their choice of locations. In some cases, the proposed locations are flexible. In this example, we were able to move four of the proposed locations to existing NGS bench marks. This saved monumentation costs, and reduced the number of observations needed to tie in the project vertically. Set new marks where needed and write station descriptions. A good station description is another detail that is sometimes overlooked but cannot be overemphasized. Make sure that the "To Reach" description can get observers to the station no matter where they are coming from. Use major road intersections that can be located on an atlas or road map for the point of beginning. This way, the observer can locate and navigate to the station. Don't forget to include local ties and a site sketch—you might have to find that mark again next winter (under three feet of snow).

Plan the Observations
I take a two-step approach to this process. First, on the project map, draw all vectors (network diagram) that need to be observed. At this point, there is no consideration of the number of receivers or session design, just an ideal picture of all necessary connections to horizontal and vertical control, and project points (including any repeat vectors). Next, using the known number of receivers, design sessions that efficiently produce the desired vectors (see Figures 3 and 4).

A Non-Trivial Detail Worth Discussing
There are two schools of thought regarding the number of vectors that should be produced from a GPS observation session. Some say that if there are four receivers, then six vectors should be produced (three independent, and three trivial). Others contend that if four receivers are used in a session, then only the "non-trivial" or "independent" vectors are of statistical value (one less than the number of receivers). I have always leaned toward the non-trivial side of the debate. I equate the scenario with locating points 2 and 3 of a triangle with an EDM from point 1, and then letting the data collector compute the inverse between points 2 and 3 to close the figure. In theory this works, unless there is an error with one or both of the true measurements. If there is an error, the inverse will close the figure as if no error existed. I prefer to have an independent measurement to close the figure, as it is a true check on the other two measurements. Having said this, we will continue by using the independent vectors only. So for this example, only three vectors will be produced from each observation session.

Back to The Plan
For this project, as you have no doubt noticed, I have opted to repeat every vector. This was done to aid in the derivation of heights, which is best discussed in a later article. The important aspect of this exercise is to see how the observations are planned. To properly plan these observations, efficiency must be balanced with proper field procedures. The tendency is to leave one observer on a station and work around him. To get truly independent measurements, we need to have each station occupied by at least two different observers. This will help to identify problems like a bad instrument height, or an observer setting up on the wrong mark. (Yes, it really does happen). Try to move the observers as efficiently as possible by keeping them going in the same direction, and by keeping as short a distance as possible.

Use your mission planning software to make sure that there is adequate satellite coverage during your proposed observation. Be sure to check availability for individual marks that may have excessive obstructions.

Draw up a final schedule and assign times to the sessions. Remember to include enough "move time" between sessions to accommodate the observer with the longest move. One note on this issue: we outfit our observers with cellular phones. This allows them to make changes to the observation schedule "real-time" if required. Sometimes due to unforeseen circumstances an observer will arrive to a station late. Without communications, the session will probably be a failure. But if the observer calls the crew chief and indicates that there has been a problem, then the crew chief can extend the session, and adjust the rest of the schedule by simply calling the rest of the observers. The final schedule for the sample project follows on page 56.
Have another member of the team check the schedule for omissions or errors. I know others would never do this, but I have been known to assign two observers to the same mark during the same session Make up observer packages, one for each observer, that contain a project map, schedule, observer checklist, station descriptions, and field observation logs (yes, I still use observation logs). I have found that the observation logs are a necessary evil. Having a comprehensive observation log is sometimes the only way to determine what actually took place in the field. It will also keep your office staff from pulling out their hair. Also, have a member of the team check each of the packages for missing information (this will keep the team from pulling out your hair).

In the next article, we will look briefly at field procedures and delve into data processing and analysis procedures.


Dan Martin is the Geodetic Program Supervisor for the Technical Services Division of the Vermont Agency of Transportation and a Contributing Writer for the magazine.

» Back to our December 2001 Issue

Website design and hosting provided by 270net Technologies in Frederick, Maryland.