A 1.189Mb PDF of this article as it appeared in the magazine—complete with images—is available by clicking HERE
Altus Positioning Systems was founded in the waning days of the financial boon of the early 2000’s which was then profoundly punctuated with a doggedly persistent recession. For those of you who may have attempted to or actually did create a business in this time frame, you have a unique perspective of the tenacity required simply to survive such a vortex as Altus has. For this reason, and more, I have considerable respect for Altus and its ability to have not only survived these interim years but to also have thrived–as evidenced by their latest product offering, the Altus APS-3L.
I reviewed the first generation Altus APS-3 in 2009, and found the system to be very well conceived and executed, providing top tier results compared to receivers of the time. This led Dad to invest in a base/ rover system of his own some time later, which we still use today. Cosmetically, the APS-3 receivers are virtually identical. Forgoing the trend to put new technology in a new package, Altus has instead kept the shell unchanged. And, in this instance, good for them for doing so as the casement has proven to be thoroughly utilitarian.
Two battery doors access the two hot-swappable batteries that power the rover for almost seven hours, while a small door in back accesses an SD card slot which came preloaded with a 2 gigabyte card. Employing an SD card for data storage eliminates the need for cabling or download software at the computer, and replacements are easily obtainable.
Also within this small door is a sim card slot for the internal GSM/GPRS cell modem which I did not test in this review. A single button powers the receiver on and off and also allows for limited user interface (such as commencing recording of raw GNSS data or resetting the receiver without the need for an external controller). Found at the bottom of the receiver are three ports: external power (LEMO 4-pin), serial 1 (LEMO 5-pin) and serial 2 (LEMO 8-pin). Provided with the system were two external power adapters (one for each receiver) and two LEMO 5-pin to DB9 cables for hardwired serial connection to data collector or PC. Broadcasting RTK corrections from the base via the internal ArWest UHF 1 watt radio will diminish that not-quite-seven-hour battery life to about five hours. A small, inexpensive user-supplied 12 volt battery will power the base for days, and so the external power cables are quite handy. With the base powered by an external source, the two 5000 mAh lithium ion batteries (two are provided with each receiver) generally used with the base can be kept with the rover to double its endurance.
The bottom of the receiver also contains the TNC antenna connector for the UHF antenna and the stainless steel 5/8"x11 female lug for attaching to a pole or tribrach.
While there has been little change to the functional exterior of the APS-3L from its predecessor, internally the APS-3L is radically different. The APS-3L sports a 136 channel receiver using a Septentrio, AsteRx2eL receiver board, that tracks dual frequency (including L2C) GPS and GLONASS, as well as WAAS and EGNOS. Altus Positioning System’s use of the GLONASS signal in its newest receivers provides incredibly better fix capability over their first generation receivers.
First generation APS-3 receivers resolved integer ambiguities using only GPS satellites. Once the receiver "fixed", any available GLONASS signals could be incorporated into the solution in order to maintain it. While generally effective, this becomes problematic under weak GPS constellations or environments where some of the signals in an otherwise good GPS constellation are obstructed. In these situations, it is possible to track multiple satellites from GPS and GLONASS, but be unable to get a fixed solution. In fact, I found myself in just such a situation with our first generation APS-3 recently. The receiver was tracking eleven satellites (GPS and GLONASS) but simply would not fix because only five GPS satellites, with a poor geometry, were in view.
The latest generation of APS-3 receivers has no such limitation. The system can achieve a fixed solution using a mix of GPS and GLONASS satellites, and can do so with as few as two GPS satellites in view. This offers an incredible advantage on difficult work sites. I would, at this point, stress to the reader that no receivers are immune from the laws of physics. Fixing integer ambiguities (the process that provides precise centimeter-level positioning with GNSS) under dense canopy is unlikely, and, when it does happen, is unreliable. This is true regardless of manufacturer. Having made this point, I can state that the APS-3L fixed in some remarkably challenging locations and did so with excellent dependability. Under dense pine trees, it did not fix. Near trees, it fixed quickly. At the edge of a tree line, it fixed. Thirty feet north of my two story frame house, it fixed. Not only did it fix in these places, the resulting positions were highly repeatable. Fix times were also very fast–generally around six seconds.
I remain impressed with the precision of the first generation APS-3 and the newest generation seems to have only improved over the first. In a brief 6 hour, 45 minute test, I recorded an RTK shot every minute on a short, 70 foot baseline. The results were amazing. The standard deviation (68% confidence level) of the resulting 405 epochs was 0.017 foot horizontally and 0.027 foot vertically. This suggests a two sigma (95% confidence level) of about one centimeter (0.033 foot) and a vertical of less than two centimeters (0.067 foot). This is truly excellent epoch by epoch precision.
Along with RTK, the APS-3L is also designed to work with L band corrections (hence the "L" in the name) from TerraStar, a British company that provides Precise Point Positioning (PPP) services worldwide. The TerraStar network consists of 80 stations globally that provide corrections which are distributed through its own satellite based system. The corrections apply to both GPS and GLONASS L1 and L2 signals. TerraStar claims global accuracies in excess of 10 centimeters at the 95% confidence level. My own testing suggested this to be entirely realistic.
With a 4 hour and 25 minute test and a 24 hour test on the Stumpwater R&D station "POST" (which has been precisely located with multiple long observations processed through OPUS) I recorded one minute epochs resulting in 265 shots and 1425 shots respectively. With PPP technology, a convergence period is required–somewhere on the order of 20 minutes. The first few epochs recorded were several feet from the true position of the POST. The position continued to converge though, and within about twenty minutes the solution became remarkably stable. Excepting the convergence epochs, the remaining epochs resulted in a standard deviation of 0.16 foot horizontally in the first test and 0.11 foot horizontally in the second which would equate to about 10 centimeters at a 95% confidence level. Vertically I observed a standard deviation of 0.14 foot in the first test and 0.21 foot in the second (approximately 10 centimeters at the 95% confidence level).
Because the corrections are satellite based, there is no concern with radio range from a base or cell phone coverage for a RTN. The sacrifice is in somewhat less precision. But in the sphere of land surveying not all points require centimeter level precision. Several scenarios come to mind that might benefit from such precision, but as with all GNSS surveying, better efficiency and capability necessitates better knowledge and understanding of the nuts and bolts of the equipment’s requirements for producing
quality results and what the results actually represent. TerraStar is no exception.
In making the comparison between TerraStar positions and my very accurate position of station POST, one thing becomes evident very quickly–TerraStar is not providing corrections based on NAD83–nor should it. "What difference does it make?" you may be wondering. The components will vary based on your location on the planet, but here in Stumpwater the difference is about 2 feet north by 3 feet east and about 4 feet in elevation, leading to a 3D difference of about 5½ feet (or two meters). This is because TerraStar is based on ITRF2008 which is very near the actual coordinates of the GPS constellation. Don’t let this intimidate you, but do make the effort to understand that there is a difference and that if you want the most from TerraStar, you’ll make the effort to transform properly from ITRF to NAD83, should you be working in an NAD83 based coordinate system (such as State Plane). For those of you using OPUS, you are likely familiar with the differences already (or should be) as both ITRF2008 coordinates and NAD83 coordinates are provided on OPUS solutions.
Altus has looked to cooperation with data collection software providers such as MicroSurvey and Carlson to make this step easier. With the supplied data collector, a Getac Nautiz X7 running Carlson SurvCE 3.0, I noticed that there were several techniques used to accommodate the offset. In one method, the user can just enter XYZ (ECEF) offsets between the two, in another more sophisticated solution, the software computes a localization between the TerraStar position and RTK positions, so that should the user lose RTK, the TerraStar solution can fill in the void with positions that are compatible with the RTK positions. Nice touch. I would however like to see the NGS 7 parameter transformation formulae for ITRF to NAD83 transformations supported in the background as an option.
With all of this in mind, what could a surveyor do with better-than-decimeter positioning? Let’s dream together a moment.
A few months ago we did a survey of a large tract of land. The tract was so large that we exceeded our radio range. Rather than move the base or worry with extending our radio range, we set the base and rover to record raw data. We used the (much less precise) WAAS corrected positions to navigate around the tract, finding monuments. (By the way WAAS suffers the same ITRF-NAD83 issues as TerraStar). Once the monuments were recovered, we’d collect a short session on the point for post-processing. The results were excellent and offered a great deal of flexibility in the field. Had we been able to use the TerraStar service, we would have been able to navigate much more accurately and locate less critical points with excellent precision, all without any need for radio or cell contact.
For staking, which obviously requires real-time positioning, TerraStar could be amazing on projects that don’t necessarily require centimeter accuracy. Pipelines, waterlines, sewer pressure mains, buried cables and ditch-to-drain roads could all be done with one rover in the field and no need for cell service or a base station.
Feasibility studies to support engineering proposals could also benefit from this level of precision as well as proposed radio/cell tower locations, well locations and hydrographic surveying. In fact, in extremely rural areas of the country, I would consider sub-decimeter positioning for boundary determination, provided it met applicable regulatory requirements and client needs.
Altus Positioning Systems made a splash with their first generation receiver, the APS-3 by providing a carefully planned, well-constructed receiver that also performed well. In spite of challenging economic circumstances since its founding, Altus has continued to improve the APS-3 both in terms of performance and capability. The APS-3L should be able to tackle most any field task suitable for GNSS. Even more, the objective evidence from my own evaluation leads me to conclude it would do so with extreme precision and reliability.
Shawn Billings is a licensed land surveyor in East Texas and works for Billings Surveying and Mapping Company, which was established in 1983 by his father, J. D. Billings. Together they perform surveys for boundary retracement, sewer and water infrastructure routes, and land development.
A 1.189Mb PDF of this article as it appeared in the magazine—complete with images—is available by clicking HERE