GPS Q&A: Industry experts answer reader's GPS questions

Q. What is the impact on the accuracy of differential GPS when the base station and roving GPS receiver(s) are separated by large distances? D.S. Englewood, Colo. A.

A. John C. Bohlke, Sokkia Corp.: Accuracy is degraded when the roving GPS receiver is separated from the base station by large distances. Processing GPS data with differential corrections assumes that both the rover and base are under the same conditions. Since the atmospheric conditions vary over large distances, the differential corrections computed at the base station may not exactly pertain to the roving receiver. The decrease in accuracy is sometimes quantified by a parts per million value that, when added to an absolute value, represents an accuracy based on the distance between the base and the rover. However, this prediction is somewhat unreliable because atmospheric conditions continually change.

Naush Ladha, NovAtel GPS: For this discussion, we will assume that when we talk about large distance, we are referring to distances greater than 1,000 kilometers (600 miles). Typically, for this type of baseline length only code data is used in a differential GPS system. Carrier-phase data is typically used for distances much shorter than 1,000 kilometers. (The advantage of using carrier-phase data to produce centimeter-level accuracies is greatly reduced when large distances are involved.)
      GPS operates in a similar fashion as conventional surveying tools such as electronic distance measuring instruments (EDMs). This means that there is a constant and proportional error associated with computed positions. The proportional error will depend on the distance the base and rover receivers are apart. Therefore, the larger the distance, the lower the accuracy. We also have to take in account the quality of the data being received. Better receivers generally provide cleaner signals and thus better accuracy.
      When operating in differential GPS mode, you require at least four common satellites at the base and rover. The number of common satellites being tracked at large distances is less than at short distances. This is important because the accuracy of GPS and DGPS positions depend a great deal on how many satellites are being used in the solution (redundancy) and the geometry of the satellites being used (DOP). DOP stands for dilution of precision and refers to the geometry of the satellites. A good DOP occurs when the satellites being tracked and used are evenly distributed throughout the sky. A bad DOP occurs when the satellites being tracked and used are not evenly distributed throughout the sky, i.e., they are grouped together in one part of the sky.
      Also, the principal of DGPS positioning assumes that there are common errors at the base and rover stations. These errors include: atmospheric errors, satellite clock and ephemeris errors, and selective availability (SA). Typically, in a differential GPS survey, a receiver occupies a survey control marker at a known location referred to as the base stations. The base station collects GPS data and computes a position. This position is then compared against the published coordinates. The difference between these two positions in the way of range errors to the satellites are your differential corrections. Usually, these corrections are then passed to your rover unit(s) for use in computing the roverÕs differentially corrected positions. However, the further apart the base and rover receivers are, the less their errors are in common. Thus, the differential corrections computed at your base are less applicable at your roverÕs location at large distances.

William Martin / Frank van Diggelen, Ashtech: There are two primary errors sources which will degrade the accuracy of baseline measurements as the baseline length increases. These are orbital errors and errors induced by the effects of the atmosphere (primarily ionosphere) on the GPS signal. On average, they combine to degrade baseline accuracy by 1 to 2 ppm of baseline length.
      Orbital errors and atmospheric effects on the GPS signal are the reason why, on many GPS receiver specification sheets, the accuracy specification has a base value and a baseline length dependent value. For example, a common accuracy specification for a high end survey grade GPS receiver is 0.005m + 1 ppm. The 0.005 portion is the base error. The 1ppm portion is the baseline length dependent error caused by the effects of the atmosphere. To determine the expected accuracy, the user must compute the baseline length dependent portion of the error and add it to the base error. As an example, lets assume a user wishes to perform a GPS survey on a 10 kilometer baseline using equipment with an accuracy specification of 0.005m + 1ppm. In most cases, the user can expect to achieve an accuracy of 0.015m, thatÕs 0.005m + [10,000m (1/1,000,000)]. The same holds true for lower grade receivers, regardless of whether or not the accuracy specification includes a baseline length dependent component. For instance, if the GPS equipment used in the above baseline measurement has a specified accuracy of 1m, the expected accuracy on a 10km baseline would be 1.010m.

Sam Shaw, Trimble Navigation: One of the basic assumptions behind differential GPS is that the base and the rover are experiencing the same error conditions at the same time. Distance between the base and the rover can cause this assumption to be partially invalidated. This shows up as a position that has been successfully corrected, but is not of the accuracy that could have been achieved had the base and rover been closer together. The separation error is known as Ôspatial decorrelationÕ and is often expressed in Ôparts per millionÕ (ppm) of the distance between the base and the rover.
      The main problem caused by a large separation distance is that the GPS signal received at the rover has traveled through a different ionospheric and tropospheric path than the one received at the base. Since the correction is generated at the base, the correction is likely to be more and more incorrect for a given rover as the separation between base and rover increases.
      The magnitude of the spatial decorrelation can depend very much on the quality of the postprocessing software. An acceptable decorrelation should be 2 ppm or less. This means that the horizontal position accuracy degrades at a rate of just 2 millimeters per kilometer of distance.

Q. As newer satellites are launched, will they improve the accuracy of GPS? L.H. Houston, Texas

A. Bohlke: New GPS satellites may improve GPS accuracy for a couple of reasons. If the Department of Defense chooses to add the new satellites to the existing constellation, the increased satellite availability and resulting satellite geometry will improve the accuracy. Secondly, it has been proposed that the new satellites transmit an additional frequency with a second C/A code and without the anti-spoofing encryption. An additional frequency would improve accuracy by enhancing the ability of the GPS processor to determine atmospheric delays and compute pseudoranges.

Dietsch: The two basic measures of accuracy; associated with GPS are the accuracy of the measurements and the strength of geometry. GPS measurement accuracy can be represented by errors in the ephemeris, clock and timing errors, propagation errors and receiver noise. The Dilution of Precision (DOP) is an indication of the strength of geometry of the constellation specifically related to the time and place in which the user is planning an occupation. The launching of more satellites into the constellation will not have much effect on the measurement accuracy, but could significantly affect the Dilution of Precision.
      As it stands today, there are still windows that have only four or five healthy satellites visible at one time. These windows usually cover a small time span but can affect productivity, especially since this usually results in very poor satellite geometry. If the user were to look at the Dilution of Precision factor in some GPS planning software during one of these periods of poor coverage, they would see a Ôspike.Õ This spike indicates that for that window in time the satellite geometry is extremely weak. With more satellites available, the possibility of always having six or more satellites is appealing because it will almost always provide a suitable geometry that is conducive to productive surveying and mapping operations.

Ladha: Assuming that newly launched satellites are similar (provide the same information) to the existing satellites, the accuracy of GPS will improve to a certain extent when newer satellites are launched. As newer satellites are launched, more satellites will be available and you will be able to track more of them. Thus, you will be able to obtain lower DOP values (see previous Q&A questionÕs response for more details). This DOP value is very important as it is directly related to the positioning accuracy.
      There have also been discussions about future GPS satellites having additional capabilities such as the ability to transmit data on a third frequency. Current GPS satellites transmit data on two frequencies, L1 and L2. These changes will benefit the GPS users through increasing the achievable accuracy of the system.

Martin / van Diggelen: If you are a military user: Yes. If you are a civilian user: No, not until SA is switched off. The new satellites will have several enhancements that will improve accuracy of a stand-alone system that is not degraded by SA.
      The new satellites will have systems to maintain more accurate orbits, but SA results in deliberately degraded transmission of the orbital data, so not benefit will be seen for civilian users until SA is turned off.
      For differential GPS the new satellites will have practically no effect on accuracy. There is discussion about a possible CA code being added to the L2 signal. This will make it easier to get high accuracy RTK position (such as is currently done with dual frequency P/Y-code receivers, such as the Z12 receiver), but the accuracy will be no better than current RTK accuracy, just cheaper.

About the participants:
John C. Bohlke is a GPS technical product manager at Sokkia Corp. in Overland Park, Kan. He may be reached at 913-492-4900 (phone), or 913-492-0188 (fax). Chris Dietsch is a product test engineer for Trimble Navigation in Sunnyvale, Calif. He may be reached at 408-481-8000 (phone), 408-481-8699 (fax), or e-mail: [email protected] Naush Ladha is an applications engineer in NovAtelÕs GPSÕs Sales and Marketing Division in Calgary, Alberta, Canada. Ladha may be reached at 403-295-4564 (phone), 403-295-4901 (fax), or e-mail: [email protected] William Martin is marketing manager, survey products at Ashtech in Sunnyvale, Calif. He may be reached at 408-524-1400 (phone), 408-524-1500 (fax), or e-mail: [email protected] Sam Shaw is a geographer at Trimble Navigation in Sunnyvale, Calif. He provides technical support for the development of GIS Data Capture applications. He may be reached at 408-481-8704 (phone) or e-mail: [email protected] Frank van Diggelen is marketing manager, OEM & Navigation at Ashtech in Sunnyvale, Calif. He may be reached at 408-524-1400 (phone), 408-524-1500 (fax), or e-mail: [email protected]

Back