Q. Does GPS replace traditional surveying methods? Why or why not?

A. William Martin, Ashtech Inc.: No. Two major reasons. One, GPS has a basic requirement of a clear path between the user and the satellites. Areas where a clear view of the sky is not available will cause degradation of accuracy best case, complete system failure worst case (not positioning available). Such areas include dense foliage, canyons (natural or urban), tunnels, etc. Two, there will always exist surveying applications where using traditional methods will result in better accuracy and/or higher productivity. This is especially true for applications where the area within which the survey is performed is small.

Eric Gakstatter, Corvallis Microtechnology Inc.: No. The reason is because GPS has a fundamental limitation in that it is a "line-of-sight" technology. In other words, the receiver antenna must have an unobstructed view of the satellites. What if you are close to a building mapping a manhole cover or underneath a bridge mapping a soil sample?
      However, when the conditions are right, GPS is a very efficient survey tool, so the answer is to use a mix of GPS and traditional survey methods to get the job done in an efficient manner. The problem is that most vendors don't take this into account.
      For example, if you need to survey a parcel of land to determine acreage but half way around the traverse, you are in an environment where you cannot receive GPS signals. If you are totally relying on GPS equipment, then you are stuck and may not be able to finish the job that day.
      With a properly designed GPS data collection system, you should be able to have a smooth transition into using traditional traversing techniques by entering azimuth, distance, slope manually (or with another instrument) from the last recorded GPS point.

Wendy Corcoran, NovAtel Communications: Although GPS is coming closer to achieving traditional surveying methods, limitations still exist. The biggest limitation is line of sight to the satellites. GPS signals get blocked or distorted when surveying in forests that are heavily treed. In most cadastre surveys performed, time is not always saved using GPS because the structure on the property blocks the GPS signal reception. Because most residential properties are under two acres, the conventional methods are as fast. GPS does not work for tunneling projects either because of the signal blockage.
      Levelling is not readily replaced by GPS at this time, however, in some situations GPS can be used, while in others, the two methods could be combined. GPS heights are in reference to a mathematical model called the ellipsoid whereas traditional surveying methods use a physical reference called the geoid. The difference between the two systems is sometimes given with monumented points but in most cases a model of the separation is used. Most manufacturers of GPS equipment can tell you which model their software or you can order the modeled data from National Geodetic Survey.
      The modeled separation between the two systems may not yield the accuracy required for some higher order leveling and most models only cover North America. Good geoid heights can be attained from GPS in localized areas where there may be several bench mark heights to work with. By observing GPS heights on traditional height monuments, a localized model can be generated. The user of GPS should be informed about what needs to be done to achieve the accuracies they require before proceeding into a project. The GPS manufacturer's customer support group should be able to assist you in this.

Jon Clark, SOKKIA Corp.: GPS does not replace traditional surveying methods; rather, it is an additional tool for measuring position. GPS may not be a practical surveying method for every type of job because of cost, local signal obstruction or accuracy requirements.

Paul Skoog, Trimble Navigation Ltd.: Trimble's experience in producing real-time and post-processed GPS solutions has shown that GPS can replace traditional surveying methods in many circumstances with increased productivity. The limitations are if the required accuracy is greater than 5 mm horizontal or 1 cm vertical, and enough satellites are visible to the GPS antenna. GPS can be used in any weather, but in heavy tree canopy or under a bridge for example, GPS surveying becomes difficult if not impossible.

Q. What are the restrictions on where GPS can and cannot be used?

A. Martin: Refer to previous answer.

Gakstatter: GPS has a fundamental limitation in that it is a "line-of-sight" technology. If the antenna doesn't have a clear path to the satellite, it cannot receive the signal and cannot be used. Operating underneath trees is a good test for receivers because this is where many of them fall apart. You will find that some cannot receive the signal at all, some will receive a signal but give you a lousy position and the good ones will receive the signal and give you a fairly good position. Regardless, you should realize that all of them do not perform as well as in a clear environment.
      Operation in deep canyons is difficult for GPS users. You should use a mission planning program so you can plan to do the field work when the most satellites are visible given the terrain constraints. The mission planning program should be powerful enough to allow you to "block out" parts of the sky so you can closely create the same environment in which you will be doing your GPS work. For example, you should be able to tell it that from 10 degrees to 85 degrees azimuth and up to 40 degrees above the horizon is a large hill. Given this information, the program will not take into account any satellites located behind the hill.

Corcoran: The restriction with GPS is that line of sight to the satellites has to be maintained. Under trees, GPS can still operate if the foliage is not too dense and the GPS antenna design is good. If the type of survey being done under trees requires consistent lock for good carrier phase data (L1), it is highly unlikely this will be the case because there will be too many cycle slips. If the data required is pseudoranges (C/A code), there will be a good chance that if the receiver displays lock to the satellites, the data will be processable. Again, buyer beware! Even if a manufacturer says their GPS receiver tracks under foliage, it does not necessarily mean you can process the data successfully. You should always test this signal reception and data quality for your application before purchasing.
      GPS can work in deep canyons but the window of four or more satellites will be very specific. Good preplanning with the appropriate elevation angle blockage should tell you the best time to survey. GPS does not work in tunnels and in most cases buildings because the signal is completely blocked. There are some cases where GPS tracks in buildings but it is usually near windows where the signal can penetrate.

Clark: Obstructions that block more than the minimum number of satellite signals restrict the usefulness of GPS. There are opportunities to collect reliable data in areas with minor obstructions, such as a tree or a building. In areas with many obstructions, such as canyons or forests, the accuracy will be unreliable and, in extreme cases, position observation is not possible.

Skoog: Where GPS surveying is limited depends on the view of the sky and the surrounding obstructions. GPS surveying is possible in some types and densities of tree canopy and not others. Obstructions such as buildings and canyon walls may or may not be a problem. It depends on whether the wall obstructs the sky or whether the wall reflects satellite signals (called multipath) to the antenna. In the event of obstructions or multipath, longer occupation times and antenna ground planes may be necessary to obtain good results.

Q. What are the pros and cons of real-time vs. post-processing and fixed integer vs. floating point solutions? Why is there not one combination of these features which gives me the best receiver?

A. Martin: Pros for real-time over post-processing: Real-time opens the door to more applications of GPS. Without real-time, any application requiring guidance of the user to a predetermined location is not possible. Examples include waypoint navigation, layout of predefined site designs, and slope staking (cut and fill). In applications where guidance is not a requirement, real-time gives the advantage of knowing that a valid solution, at the desired accuracy, has been obtained for each object surveyed.
     Cons for real-time over post-processing: Real-time GPS surveying adds the complexity and cost associated with a telemetry link between the base station (on known location) and the remote units. Added issues to deal with include possible licensing requirements, range of radios, obstructions to radio singles, etc.
      Pros of fixed integer vs. floating point solutions: Simply, a fixed integer solution is a more accurate solution than the floating point solution. The level of improved accuracy is dependent on a number of factors including occupation time and ionospheric conditions.
      Cons of fixed integer vs. floating point solutions: A fixed integer solution is more difficult to obtain requiring more stringent data collection procedures. It is also possible to incorrectly fix the integers causing errors that could exceed the accuracy of the float solution.
      The definition of the "best" receiver will differ between users and their applications. A real-time GPS receiver system capable of geodetic level accuracies will allow the user to also post-process the data and will give the user a float and/or a fixed integer solution. But the price of such a system may not be attractive for some applications, depending on desired level of accuracy and productivity.

Gakstatter: Real-time processing is quite convenient when it is available. Most of the time, in our applications (GIS), the user wants to navigate back to a point that was recorded before (i.e. soil sample, section corner). Without real-time, they can get within 50-100 meters. In some cases this is close enough, however if there are several points close together (50 feet from each other), then it is difficult without real-time processing.
      The major problem with real-time processing is that some kind of data link is necessary between the base station and the field unit. This data link can be narrow band radios, FM pagers, Coast Guard beacons, Wide-Area DGPS, cellular modems, etc.
      Each of these data link methods has its own set of limitations with the primary one being coverage. The coverage is limited because most of the data links are radio frequency (RF)-based. You have encountered the same problem when you are listening to the radio in your car. Sometimes it just cuts out because of interference from another source that you aren't aware of or just because you are out of range.
      I would check around the job site you plan to work in to see what kind of coverage is available (for example, I live 45 miles from a station, but because of the hilly terrain, I cannot receive the RTCM message). Even if you think you have good coverage at your job site, I would still recommend that you use a system that will automatically record raw data when the RTCM message is not available so you can correct the data in post-process mode and not be forced to wait in the field for the RTCM message to be received again.
      The reason there is not one receiver that can "do it all" is that there are so many applications for GPS and each one has its own set of requirements.
      For example, there are hand-held receivers made for private pilots that are primarily for casual navigating; there are hand-held receivers that are made for professional GIS users who want to associate field data with GPS coordinates; and then there are high precision GPS receivers that help surveyors layout a subdivision. Each class has its own price and performance qualities. For example, if I was looking for a GPS receiver that will allow me to map utility poles, I would not be going to the local sporting goods store looking for a GPS receiver because the one at the store is probably designed primarily for hikers who want a rough position for a cheap price. To them, 100 meters is great. However, for my application I want a pretty good position (a few meters) and I want to be able to record a lot of descriptive data (i.e., number of wires, number of arms) along with the GPS position.
     So that you can narrow your search to the receivers in the class that you need, it is important to define your needs and then start inquiring about receivers in the class that can meet your needs and offer a good value for your dollar.

Corcoran: The need for real-time positions vs. post-processed positions depends first and foremost on your application. If you need to guide a vessel or aircraft, you need that information as it happens, not after the fact. If you have the options of either real-time or post-processing, the pros and cons are as follows:
      real-time
      Pro: The answer and quality of the data is immediately available which will save you time back in the office or time in having to revisit a site because the data was not good.
      Con: In order to perform real-time positioning, a differential link is required. There are many ways the differential corrections can be acquired but it will either cost you money to access them (i.e. DCI, J.E. Chance or Que Paging) or you will need a radio link and a base station. This will require licensing from the FCC which is getting harder to obtain. The exception to this is the Coast Guard Beacon system which will require the beacon receiver but no link or base station. The second con for real-time is that there is no compensation for problems in the data. Real-time positions are sometimes less accurate because of this, but again depending on the accuracy you require, it may be sufficient.
      POST-PROCESSED DATA
      Pro: Data can be examined more closely to isolate trends, smooth the data and the data can be processed forwards and backwards to obtain the best answer or bridge problems such as cycle slips. The history of a data file can be used to obtain accurate positions or to isolated and corrected problems through post-processing.
      Con: Post-processing takes time and trained personnel.
      Fixed integer solutions vs. a floating point solution is typically a matter of accuracy and robustness of solution. Fixed integer solutions require that the integers be resolved therefore it takes more time. For a real-time application, this could be up to 10 minutes before resolving the ambiguities, but once resolved the accuracies will be within a few centimeters if the remote is not more than 10-15 km from the base station. A floating point solution computes the ambiguity but leaves it as a whole number and does not try to isolate its integer value. That means less time and typically a 20 cm or less positioning accuracy. Since integers are not being resolved, they cannot be resolved incorrectly so there is no chance of drifts in the position from wrongly chosen integers. This translates to a more robust solution with less convergence time (1-2 minutes).
      There are GPS manufacturers whose receivers will operate in real-time or they will supply the software to post-process. Most GPS manufacturers can post-process data to a fixed integer solution using the carrier phase data but have also chosen the fixed integer solution as the real-time technique only. The float point solution is only offered by one or two manufacturers. The "best" receiver is the one that is going to perform to the accuracy, distance and occupation time you need for your application. If this receiver can perform all of the other functions as well, that's great but don't pay extra because of it.

Clark: Overall, a single combination of these features would not provide the best solution for all GPS users. Some features are more suited for certain needs or applications, and it is up to the user to determine what best suits the requirements of a particular job.
      Some advantages of real-time over post-processing include the ability to stake-out, navigate and obtain accurate positions while working in the field. However, real-time capability relies on tracking a correction signal that may not provide sufficient coverage. Real-time typically provides less accuracy than post-processed data.
       The primary advantage of a fixed integer solution over a floating point solution is its ability to identify major errors. However, it takes less time to differentially process data using a floating point solution.

Skoog: The advantage of real-time surveying is productivity. It is possible to perform stake-out, topographic, and control surveys real-time with centimeter level accuracy without a line of site requirement associated with conventional methods. This level of productivity and confidence in the results is not possible with post-processed techniques. The advantage of post-processed techniques is higher accuracy over baselines longer than 10 to 15 km. The advantage of a fixed integer solution is high accuracy vs. the float solution. The most accurate solution is always a fixed integer solution. In some cases this may not be possible so a float solution is the next best.

About the Participants:
William Martin serves as marketing manager, survey products at Ashtech Inc. in Sunnyvale, Calif. He can be reached at 408-524-1400.
Eric Gakstatter is vice president of marketing at Corvallis Microtechnology Inc. in Corvallis, Ore. He can be reached at 503-752-5456.
Wendy Corcoran serves as manager, survey products at NovAtel Communications Ltd. in Calgary, Alberta, Canada. She can be reached at 403-295-4900.
Jon Clark is a systems marketing manager at SOKKIA Corp. in Overland Park, Kan. He can be reached at 913-492-4900 or 800-4-SOKKIA in the U.S.
Paul Skoog is a product manager for land survey receivers and antennas at Trimble Navigation Ltd. in Sunnyvale, Calif. He can be reached at 408-481-8000.

Back