Notes on the GPS Positioning Exercise for the Geographer's Craft

Peter H. Dana

2/6/96

Introduction

The object of the exercise to attempt to measure the four corners of the Geography Building with GPS. The task seems simple enough and yet is quite difficult to accomplish. This is a very common problem with many positioning projects that have selected GPS as the source for position data.

The task is difficult because GPS depends the ability of the receiver antenna to "see" four or more satellites at the same time in order to provide a position solution at all. Without tracking four satellites the receiver cannot provide (without aiding) a three-dimension fix. The task of measuring the corners of the Geography Building is even more difficult because the ability to track four satellites is not enough to insure an accurate position solution. To achieve the specified accuracies of GPS, the receiver must track four satellites which are well distributed around the sky so that the satellites measurements provide good Geometric Dilution of Precision (GDOP). Essentially the satellites must be selected so that they provide range measurements with respect to east and west, north and south , and up and down all at the same time. They cannot do this at all if they are all in the same place in the sky, and can only do it well if they are "nicely" distributed around the sky with some low in the sky and some high, with some to the east and west and some to the north and south.

The Trimble GeoExplorer receiver provides the user with an indicator of the effect of GDOP on the position solution. The PDOP figure available on the user Satellite Tracking screen reached from the GPS Status screen under the Main Menu screen, gives a figure for the geometric dilution of precision for three dimensional position, PDOP, or Position DOP. This number can be used to estimate the effect geometry will have on the position solution. If the PDOP is more than around 6.0, the solution will be twice as bad as if the PDOP was 3.0 (a more normal value). Check the PDOP value before and after taking position data.

You will find that while attempting to use the GPS receiver right up against the corner of the building that much of the sky is blocked by other structures and by the Geography Building itself. The portion of the sky that can be seen by the antenna is very limited. While you will occasionally be able to see the required four satellites, the GDOP will usually be rather poor. One of the reasons for this exercise is to point out the problems in specifying GPS for positioning projects and a common problem is the inability of the receiver to do the job in the project environment. Whenever trees, terrain, buildings, or even heavy foliage cover block the GPS signals, the specified accuracies of GPS may not be available.

Because there are 24 SVs orbiting the earth, and because these SV are constantly changing their positions with respect to the receiver, you will probably be able to measure the signals from four satellites at each of the corners of the building. You may have to wait for the satellites to move into positions from which they can be seen. Because they orbit the earth every twelve hours and move across the sky above you in only a few hours, there positions do change fairly quickly and so you should not have to wait too long.

Consider other ways of using GPS position to measure the corners without taking measurements right at the corners of the building. It might be possible to take readings at points that are lined up with the sides of the building. It might require eight measurements rather than four, to develop a set of positions that would allow you to draw four straight lines that would intersect at the four corners of the Geography Building, but the difficulties of measuring the corners directly might make this a worth doing.

The Trimble Navigation GeoExplorer GPS Receiver

This is a six channel GPS receiver that can provide GPS position data in several modes of GPS operation.

Stand-alone C/A code GPS positioning

In this mode, the GeoExplorer tracks up to eight GPS space vehicles (SVs) and provides three-dimensional position fixes in a variety of coordinate systems and geodetic datums. As specified in the Federal Radionavigation Plan, this Standard Positioning Service (SPS) of GPS can provide horizontal position fixes to an accuracy of 100 meters (95% - meaning that a distribution of independent position fixes will provide position estimates such that 95% - or two standard deviations - of the radial distances from the true position to the estimates will be less than 100 meters). Vertical accuracies are specified as 156 meters (95%).

Real-Time Differential GPS (DGPS) Positioning

In this mode, the GeoExplorer can be interfaced to a communication link that provides differential corrections for each individual SV measurement. These corrections can result in positioning accuracies of one to five meters. A source of DGPS corrections must be available. Several public and private agencies provide DGPS real-time corrections in different areas.

Post-Processed DGPS Positioning

In this mode the GeoExplorer collects SV measurements that are later downloaded to a computer for DGPS correction. Correction is done with respect to data files collected at a reference station within a few hundred kilometers of the remote GeoExplorer. This mode can also result in one to five meter positioning accuracies, but without the need for a communication link to a real-time DGPS service.

Post-Processed Carrier Phase DGPS Positioning

The GeoExplorer can also be used in a specialized sub-set of survey-mode GPS carrier-phase positioning. With specialized software and data files, a sub-meter position accuracy can be obtained within 30 kilometers of a reference base station that supplies the specialized measurement data files that are required for this mode.

The GeoExplorer as a Base Station

The GeoExplorer can also act as a source for DGPS base station, or reference DGPS. The receiver has a limited storage capacity and so is not a good choice for a permanent base station, but two GeoExplorers, one at a precisely known location, and the other a roving remote receiver can act as a complete DGPS system.

In this exercise we will use the GeoExplorer in two modes, the stand-alone mode, and the post-processed DGPS mode.

Hints for GPS Measurements and DGPS Corrections

In addition to the receiver and a set of fresh batteries (four AAs), you will need the interface cable and a PC loaded with the GEO-PC software to accomplish the task.

You should turn the receiver on and delete all the files previously stored within it before starting your exercise. If you install new batteries reset the battery usage to zero in the Configuration menu. The nice thing to do is use the batteries that are in the receiver and install your new ones after you are done for use by the next group.

Because you will need a DGPS correction file from a Community Base Station to correct you measurements, and because these files cover one hour periods of time, you might want to schedule your project so that you will make all your measurements within a particular one-hour period. This will make post-processing much easier.

Find a place near the building that provides a good view of the sky. Turn the receiver on by pressing the bottom button. Hold the receiver so that the antenna under the Trimble logo (an engraved sextant) is pointing up toward the sky. The receiver is curved so that if you keep the antenna pointing up the keys and screen are still visible.

While the receiver begins to track satellites and gather the required Navigation Message from each available satellite, use the menu keys as instructed in the user guide to find the GPS Status menu and the Sat. Tracking screen. Watch as the receiver acquires SVs from the list of SVs that should be visible from receiver (using the last position solution) and when the receiver is tracking at least four switch to the position screen and wait until the receiver begins to update the position. Note that the position will change as the noise and bias errors move the apparent position of the receiver around. Note too that position movement is most apparent in the HAE (height above reference ellipsoid) figure for two reasons. The height is displayed in meters while latitude and longitude are displayed in decimal seconds (1.0 seconds is about 30 meters) and so the displayed numbers change less for latitude and longitude than they do for height. In addition the sensitivity of GPS to errors is more (156 meters Vs 100 meters) in the vertical direction (due in large part to the ionospheric delays).

When you are tracking enough satellite to get position fixes, move toward the spot you have selected for the first set of measurements. If, close to the building, you can no longer get position fixes, wait, or reconsider the approach you have taken to the task. When you are at a location you do wish to measure, check the PDOP value, locate the Data Capture Screen, and open a "rover" file as instructed in the user guide. Take at least 30 position readings before closing the file. Remember to open and close a new file at each measurement point. Note the file names so that you can recover the data later. When you close the file, note the PDOP value. If you think it was too high, you might want to try again or reconsider where you want to make your measurements. You may also wish to record one independent position reading at each point to compare un-corrected, stand-alone GPS with DGPS.

After collecting four (or more) data files, turn the receiver off and return to the lab.

Correcting GPS Data Files

Get the DGPS base station correction files from the Community Base Station for the data and time of your data file collection. The files are stored so that the file names indicate the data and time of applicability. For February 12, 1996, between 2:00 and 3:00 PM the file name would be T6021214.EXE. This file must be stored in the GEO-PC\DATA directory and then run by typing T6021214 at the DOS prompt. The self-extracting file needed for correction, T6021214.SSF will be placed in the GEO-PC\DATA directory.

Connect the interface cable to the back of the receiver, start the GEO-PC software and download your un-corrected files to the GEO-PC \DATA directory.

Following the instructions in the GEO-PC software guide, set up the reference location, the reference files and correct your data files. Be very careful to be sure you know exactly what data files are being used in this process. Use the graphic display features to look at your uncorrected and corrected data files. Use the statistics menu to compute mean and standard deviations for both corrected and uncorrected data files.

Using the Geo-PC software convert the positions in each of your files to ASCII GIS files in the Texas Central Zone State Place Coordinate (SPC) system in survey feet. Do this for each uncorrected and corrected file. Clever folks will find that the software can be used to take each corrected file and produce a single file with one line containing the mean position in SPC for each of the measurement points.

As part of the exercise you are to provide one independent un-corrected position for each corner of the building. Use the first position in the un-corrected files if you did not record a single position reading while making the measurements.

When you plot your final results in State Plane Coordinates you may find that your DGPS position estimates do not describe a rectangle and that there are errors of perhaps even tens of meters. This may well be the result of poor PDOP during data collection. If your errors are more than 20 meters, you may have to repeat the measurements at one or more points to properly complete the exercise.

Results

If you are lucky and were able to make measurements at a time when SV availability was good and the PDOP values during your data collection session were low it is possible to end up with building coordinates within one to two meters of the actual building corners. If your position plots do not agree with the maps consider whether it is more likely the maps or the GPS positions that are in error. If your State Plane Coordinate positions describe a rectangle the size of the Geography Building but the rectangle is shifted or rotated with respect to the SPEC coordinate system on the maps, remember that local plane surveying techniques (used to make the maps) are very accurate in a relative sense but that GPS might provide better accuracy in an absolute geodetic coordinate system. The maps may be better representations of local features, and GPS may be better able to describe positions with respect to a global coordinate system.

Note the extent to which your un-corrected, stand-alone GPS errors exceed the DGPS solutions. Expect even more than the specified 100 meter error if you have collected the points in poor PDOP situations.

No matter how well you did, expect the DGPS estimate accuracies to be the result of a process subject to noise, bias, and possibly blunder errors that at the best can approach one meter accuracies and in real situations, such as we have tried to show here, may be as much as tens of meters. Planning, care, and a good measurement environment are all necessary in order to achieve the specified (and advertised) accuracies possible with differential GPS.